AI-driven Music Therapy

Mental disorders show a rapid increase and cause considerable harm to individuals as well as the society in recent decade. Hence, mental disorders have become a serious public health challenge in nowadays society. Timely treatment of mental disorders plays a critical role for reducing the harm of mental illness to individuals and society. Music therapy is a type of non-pharmaceutical method in treating such mental disorders. However, conventional music therapy suffers from a number of issues resulting in a lack of popularity. Thanks to the rapid development of Artificial Intelligence (AI), especially the AI Generated Content (AIGC), it provides a chance to address these issues. Nevertheless, to the best of our knowledge, there is no work investigating music therapy from AIGC and closed-loop perspective. In this paper, we summarise some universal music therapy methods and discuss their shortages. Then, we indicate some AIGC techniques, especially the music generation, for their application in music therapy. Moreover, we present a closed-loop music therapy system and introduce its implementation details. Finally, we discuss some challenges in AIGC-based music therapy with proposing further research direction, and we suggest the potential of this system to become a consumer-grade product for treating mental disorders.

1578 Background: Studies on the role of music therapy for improving the quality of life of cancer patients are scarce. Artificial intelligence (AI) is being increasingly utilised in healthcare universally. Echo Care ©DigiNxtHlt Solutions leverages AI technology multi-class neural networks to make personalised recommendations by learning continuously from patients' usage and interactions. We aim to study impact of this AI based music therapy in health-related quality of life (HR-QOL) in cancer patients undergoing chemotherapy. Methods: A pilot interventional study for 50 consecutive cancer patients being administered chemotherapy at daycare facility was conducted. After informed consent, each patient underwent assessments during chemotherapy on week 1, week 3 and week 6. Vitals assessment, FACT G7 and HADS assessment was done at baseline, 3 rd week and 6 th week.30 minutes experience of sounds rendered by “echo Care” AI based application ©DigiNxtHlt Solutions was provided at each session of chemotherapy at 1 st week, 3 rd week and 6 th week. Vitals were checked on admission to daycare unit followed by HR-QOL assessments which were done on digital platform by patients using unique ID and password. After completing the FACT G7 and HADS assessment, option was given to select from multiple music modules generated by AI based on data entered by patients. Once they completed the 30min session, post session vitals were measured. Qualitative feedback also was collected. The primary objective was quality of life as defined by FACT G7/ HADS scores and secondary endpoint was change in physiological vital parameters at baseline, 3 weeks, 6 weeks. Statistical analysis done by SPSS version 22.0. Results: A total of 50 patients were enrolled for the study during study period from March 2023 to Dec 2024. Mean HADS score for depression was 10.2 +/-4 and for anxiety was 7.5 +/- 3. 40% patients had high HADS score for anxiety. 18% had high HADS score for depression.50% of patients had improvement in HADS score for anxiety and depression whereas 50% did not have much change in their scores. Mean FACT-G7 score is 13.5+/-2 with majority having value less than 16 (70%). From qualitative feedback, 25% reported “very satisfied”, whereas 75% reported to be “satisfied” with music intervention. 54% patients felt less anxious, 50% had calming effect and masking of pain, 37.5% felt mood elevation, positivity and closer to their family. 30% patients felt muscle relaxing effect. 37.5% felt sleepy during music session with 25% not feeling any change in them during the sessions. Conclusions: Artificial intelligence-based music therapy intervention in cancer patient undergoing chemotherapy shows promising results in terms of health-related quality of life, satisfaction and experience and can be offered to cancer patients as a non-pharmacological intervention to improve their quality of life during their treatment. Clinical trial information: CTRI/2023/03/050509 .

Abstract Artificial Intelligence (AI) has quickly permeated various professional fields, including creative and therapeutic disciplines. In music therapy, AI provides tools that can augment traditional practices through enhanced workflows, creativity, personalized interventions, and data-driven insights. This article explores the potential of AI in music therapy, detailing its applications across planning, implementation, and evaluation phases. Trending AI technologies for generating text, images, music, and videos, that enrich therapy by tailoring interventions to diverse populations and individual client needs, are introduced. Selection criteria are outlined to guide music therapists in choosing appropriate tools for their practice. As an example, the e-book Pippin the Piano demonstrates how AI can generate valuable clinical resources. A field test highlights how the AI-generated materials complement telepractice through dynamic storytelling and interactive engagement while maintaining the therapist's central role in guiding therapy and building meaningful connections. Ultimately, AI holds significant promise for innovative therapeutic practices, but its integration requires digital skills, ethical awareness, and must adhered to professional standards. Therefore, an initial framework of core competencies is presented to guide music therapists in navigating these challenges. In conclusion, thoughtfully incorporating AI enables therapists to expand accessibility, inclusivity, and impact, while remaining competitive in an AI-driven job market. Keywords: artificial intelligence, music therapy practice, personalized care Resumen La Inteligencia Artificial (IA) ha irrumpido rápidamente en diversos campos profesionales, incluyendo disciplinas creativas y terapéuticas. En la musicoterapia, la IA proporciona herramientas que complementan las prácticas tradicionales mediante la optimización de flujos de trabajo, el fomento de la creatividad, la personalización de intervenciones y el análisis basado en datos. Este artículo explora el potencial de la IA en la musicoterapia, detallando sus aplicaciones en las fases de planificación, implementación y evaluación. Se presentan tecnologías de IA emergentes para la generación de texto, imágenes, música y videos, las cuales enriquecen la terapia al adaptar las intervenciones a diversas poblaciones y necesidades individuales. Además, se describen criterios de selección para guiar a los musicoterapeutas en la elección de herramientas adecuadas para su práctica. Como ejemplo, el libro electrónico Pippin the Piano demuestra cómo la IA puede generar valiosos recursos clínicos. Un estudio de campo destaca cómo los materiales generados por IA complementan la telepráctica a través de narrativas dinámicas e interacciones participativas, preservando el rol central del terapeuta en la guía del proceso terapéutico y la construcción de conexiones significativas. En última instancia, la IA ofrece un gran potencial para prácticas terapéuticas innovadoras, pero su integración exige competencias digitales, conciencia ética y apego a estándares profesionales. Por ello, se presenta un marco inicial de competencias clave para orientar a los musicoterapeutas en estos desafíos. En conclusión, la integración reflexiva de la IA permite a los terapeutas ampliar la accesibilidad, la inclusividad y el impacto, al tiempo que se mantienen competitivos en un mercado laboral impulsado por la IA Palabras clave: inteligencia artificial, práctica de la musicoterapia, atención personalizada.

This study explores the theoretical framework for constructing personalized pathways in receptive music therapy by integrating the Traditional Chinese Medicine (TCM) principle of “Three Categories of Etiological Factors System” with artificial intelligence (AI) technology. The three dimensions—temporal appropriateness, geographical appropriateness, and individual appropriateness - offers a holistic approach to personalized healthcare that aligns with modern precision medicine. Through systematic analysis of how these three dimensions apply to music therapy practice, we propose a comprehensive AI-based multidimensional assessment model and dynamic adjustment mechanism. The research establishes a theoretical system encompassing three core modules: individual characteristic identification, environmental factor analysis, and temporal pattern recognition. This framework enables the intelligent selection and dynamic optimization of therapeutic music based on patients' physiological states, psychological profiles, cultural backgrounds, circadian rhythms, seasonal variations, and environmental contexts. By leveraging machine learning algorithms and deep learning technologies, the system can process multimodal data to generate highly personalized treatment protocols and adjust them in real-time according to therapeutic responses. This integration of ancient medical wisdom with cutting-edge AI technology provides novel theoretical perspectives and practical pathways for achieving precision-oriented and intelligentized music therapy, potentially enhancing treatment efficacy and accessibility across diverse clinical settings including mental health, neurological rehabilitation, pain management, and pediatric developmental disorders.

Due to its non-pharmacological nature and low cost, music therapy has gained increasing attention in the fields of medicine and mental health and has been applied in the treatment of various diseases. However, traditional music therapy is limited by the lack of standardized treatment duration and frequency, as well as its inability to meet individual patient needs, hindering further development and broader application. This paper reviews relevant studies, highlighting the therapeutic effects of current music therapy methods on various diseases, and explores the application and efficacy of artificial intelligence (AI) technologies such as Random Forest (RF), Long Short-Term Memory networks (LSTM), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN) in supporting music therapy. The results show that music therapy can help improve patients negative emotions and alleviate symptoms, while AI technologies can enhance and optimize traditional music therapy in areas such as music selection and treatment plan formulation, leading to more personalized and effective outcomes. In conclusion, music therapy demonstrates potential in clinical applications, and integrating AI technologies may offer more individualized and effective treatment plans. Future research and practice in this area should be further strengthened.

In recent years, music has been regarded as a promising non-pharmacological intervention for a number of physical and mental conditions. Five-elements music therapy—based on the five-element theory—is a unique non-pharmacological therapy of East Asian traditional medicine. It has the potential to effectively provide individualized music therapy to individuals with illness. However, one limitation of this music therapy is that the classification of the five elements and its application is mainly based on subjective judgment. The development of artificial intelligence (AI) has enabled the acoustic analysis of multi-factor sound sources. This can develop five-element music therapy. Here, we discussed the challenges proposed by the future combination of five-element music therapy and AI. Further, we hypothesized that AI may promote its use in the medical field.

Introduction This study explores the effectiveness of music therapy in repairing psychological trauma in bereaved families, aiming to provide a comprehensive understanding of its potential therapeutic impact. It begins with an analysis of the current situation faced by bereaved families, identifying the psychological challenges they experience. Methods The research design included the recruitment of participants from bereaved families, who were then divided into an experimental group and a control group. An optimized Long Short-Term Memory (LSTM) network model was constructed to analyze music therapy data, tailored specifically to capture the nuances of this therapeutic process. The experimental procedure detailed the specific operations involved in the music therapy sessions and established a clear comparison framework between the two groups. Results The performance of the proposed LSTM model demonstrated high accuracy (0.85), precision (0.86), recall (0.84), and F1-score (0.85), indicating its effectiveness in predicting treatment outcomes. When compared to the Self-Reporting Inventory-90 (SCL-90) scale, the model captured the trend of treatment effects with a high level of accuracy and reliability. Notably, participants numbered 2, 6, and 8 in the experimental group showed substantial improvement rates of 67.21%, 71.45%, and 75.67%, respectively, in their psychological health issues. Discussion The comparative analysis between the experimental and control groups confirmed that the music therapy approach, as guided by the proposed LSTM model, led to a more significant improvement in psychological health issues for bereaved families. This suggests that the model offers a promising avenue for enhancing the effectiveness of music therapy in this context.

This study aims to explore the theoretical underpinnings and experimental frameworks of music therapy, particularly when augmented by artificial intelligence (AI) technology. The increasing integration of AI in various scientific domains and daily life prompts its consideration in medical and musical applications. The primary objective is to assess the efficacy of AI-generated music in music therapy settings. To achieve this, a multi-faceted approach combining synthesis, analysis, induction, deduction, and descriptive methods is employed. Additionally, the historical-genetic method is utilized to trace the genesis and evolution of the music therapy-AI intersection. The practical implications of this research are profound, as the findings can inform the development of music therapy practices that incorporate AI. Initial findings suggest that AI enables a more individualized approach in music therapy, which is advantageous. Furthermore, AI-assisted melody creation for therapeutic purposes is significantly faster than traditional methods relying solely on human composition. Consequently, this study underscores the novelty and significance of AI in music therapy research.

Against the backdrop of refined healthcare management, traditional Chinese Medicine (TCM) Five-Tone Therapy faces limitations like insufficient personalization. This study aims to construct a personalized clinical therapeutic music management system integrating TCM Five-Tone Therapy with generative artificial intelligence technology and explore its application value and implementation pathways in enhancing patient satisfaction, alleviating anxiety, and improving pain management. Centered on the TCM theory linking the Five Tones ("Gong, Shang, Jue, Zhi, Yu") to the Five Zang Organs and Five Emotions, the system adopts a closed-loop architecture with multimodal data collection, Five-Tone feature mapping, personalized music generation, and feedback optimization, tailored to patients’ clinical conditions, real-time emotional states, and musical preferences. It converts relevant indicators into quantifiable musical parameters aligned with Five-Tone Therapy principles, generating music precisely matching patients’ physical and psychological needs. This system provides an innovative pathway for clinical emotional regulation, enhances hospital service quality, and shows significant potential in anxiety and pain management, deserving promotion in healthcare institutions.

Abstract In order to improve the accuracy and reliability of EEG emotion recognition and avoid the problems of poor decomposition effect and long time consumption caused by manual parameter selection, this paper constructs an EEG emotion recognition model based on optimized variational modal decomposition. Aiming at the modal aliasing problem existing in traditional decomposition methods, the KH algorithm is used to search for the optimal penalty factor and the number of decomposition layers of the VMD, and KH-VMD decomposition is performed on the EEG signals in the DEAP dataset. The time-domain, frequency-domain, and nonlinear features of IMFs under different time windows are extracted, respectively, and the Catboost classifier completes the construction of the EEG emotion recognition model and emotion classification. Considering the two conditions of the complexity of the network structure of the KH-VMD model and the average classification accuracy of different brain regions in different music environments, the WEE features of the target EEG can constitute the optimal classification network by taking the WEE features of the target EEG as the input of the KH-VMD classification model. At this time, the average classification accuracy that can be obtained with differentiated brain regions and differentiated music environments is 0.8314 and 0.8204. After 8 weeks of music therapy, the experimental group’s low anxiety scores of pleasure and arousal on the Negative Picture SAM scale were 3.11 and 3.2, which were significantly lower than those of the control group’s low-anxiety subjects. The experimental group with high anxiety had anxiety scores and sleep quality scores that were 5.23 and 3.01 points lower than before the intervention. Therefore, music therapy can effectively alleviate psychological anxiety and enhance sleep quality.

This study proposes an artificial intelligence-based individualized music therapy system targeting the emotional and memory functions of the elderly with Mild Cognitive Impairment (MCI). The system can realtime identify users' emotional states and then dynamically adjusts music intervention content accordingly. Thus, it achieves emotional regulation and auxiliary improvement of memory functions. The study innovatively proposes an emotion recognition method based on prefrontal Electroencephalogram (EEG) and Heart Rate Variability (HRV). The system adopts a feature fusion strategy of 6-lead prefrontal EEG and HRV, and uses the Support Vector Machine (SVM) algorithm to achieve emotion classification. In addition, a closed-loop system of "emotion recognition-music intervention" is constructed, enabling the system to instantly adjust the type and parameters of played music according to the emotional changes of the elderly. In the experiment involving 60 elderly MCI patients, the system can complete high-precision emotion recognition with only 6 leads, which greatly simplifies the equipment complexity. Results show that the average accuracy of emotion recognition by the proposed method reaches 89.7%, which is significantly higher than 83.3% of the single EEG method and 76.6% of the single HRV method. This study verifies the core role of the prefrontal lobe region in emotional processing and provides an engineering basis for simplifying the emotion recognition system.

Based on public datasets, this study designs a simulation experiment to explore the potential impact of an integrated treatment framework incorporating artificial intelligence (AI) on the long-term mental health of patients with depression. Different from previous studies that focused on a single intervention method, this paper proposes combining AI technology with music therapy and herbal therapy to construct an individualized intervention optimization and feedback system. During the sixmonth simulated intervention, the system evaluates the effects of different treatment pathways on psychological assessment indicators such as Patient Health Questionnaire-9 (PHQ-9) and Hamilton Depression Rating Scale (HAMD). The results show that the integrated treatment group is significantly superior to the control group in the main scoring indicators. This study verifies the feasibility of applying AI to the integration of multimodal treatment strategies under non-clinical conditions and provides new ideas for personalized intervention in depression.

Music‐based interventions may offer accessible adjunctive support for several mental conditions, but average treatment effects are heterogeneous and objective outcomes are often inconsistent. Artificial intelligence (AI) can select or generate music at scale, yet tailoring the musical stimulus alone does not establish a biologically individualized treatment. This conceptual perspective distinguishes three levels of individualization: the patient′s clinical and cultural context, the neural or psychophysiological target, and the musical stimulus and dose. Group‐level disorder‐network findings should serve only as provisional priors. Before an AI system adapts music to presumed target, the candidate target should demonstrate within‐person reliability, relevance to the intended symptom or function, directional interpretability, and modifiability by music relative to an active control. We refine two development pathways: theory‐driven AI selection and data‐driven AI composition. Both require patient‐specific calibration, prespecified falsification criteria, held‐out and external validation, subgroup fairness analyses, clinician oversight, and confirmation that changes in a neural surrogate correspond to clinically meaningful outcomes. Electroencephalography is currently the most feasible primary signal for a real‐time feedback loop; peripheral physiology and functional near‐infrared spectroscopy may provide complementary information, whereas functional magnetic resonance imaging is more practical for offline target discovery and calibration. Individualized music therapy should not be defined as AI‐generated music optimized toward a group‐average disorder signature. A defensible model must individualize both the therapeutic target and the musical intervention, place patient goals and culture within the optimization process, and remain falsifiable, equitable, and clinician supervised.

The reason why music can affect people's emotional experience is that the stimulation can be transmitted to the brain through hearing, such as the thalamus and lenticular nucleus. Music therapy has a positive auxiliary treatment effect on mental health. Therefore, an evaluation model of the auxiliary effect of music therapy on mental health based on artificial intelligence technology is proposed. We construct the constraint index parameters for the evaluation of music therapy's auxiliary effect on mental health, take parent pressure, self-pressure, teacher pressure, and social pressure as the questionnaire object parameters, take class type as the independent variable, carry out an independent sample t-test, and construct an adaptive information extraction model for the evaluation of music therapy's auxiliary effect on mental health. Paired sample t-test is used to analyze whether there is a difference between the experimental group and the control group on the learning stress scale. According to the analysis of the difference between the experimental group and the control group, combined with the difference test analysis of the data of the stress release of music therapy on mental health, the quantitative evaluation of the auxiliary effect of music therapy on mental health is realized through artificial intelligence optimization control. The experimental results show that the accuracy and reliability of this method to analyze the auxiliary effect of music therapy on mental health are high. There are obvious changes in the data of students' self-pressure, and the difference between the average value and standard deviation of the data before and after the course is obvious. From the perspective of the effectiveness of the course, the students in the class who implement the four relaxation experience courses in the course are under the pressure of parents, self-pressure, teacher pressure, and social pressure. There are obvious changes in the five aspects of learning pressure compared with that before the implementation of the course. After the course experience, the pressure value of most students decreases, and the course intervention effect is obvious.

: As the academic burden grows, students' mental health problems are becoming more apparent. Music is a natural, safe and accepted art form, and is widely used in psychotherapy. The traditional assessment frameworks are however not suitable to measure the impact of music therapy (MT) on the mental health of students, accurately and efficiently. To compensate for this, an evaluation system of music therapy effectiveness based on support vector machine algorithms has been built. This system will automatically track and evaluate students psychological reactions to MT and adjust the MT therapy programme in real-time to provide optimal therapeutic outcomes. As experimental subjects, the students who participated in this study were students who took MT for the first time in the third year of junior high school, third year of senior high school and fourth year of college. MT sessions were held and students' psychological response data was regularly collected and compared with data from traditional evaluation methods. The final experimental outcomes showed that the AI-based MT evaluation system can attain an average evaluation accuracy of 91.25%, evaluation time of 9 minutes and 58 seconds, and cost of 42.46 USD. Overall, these findings are better than the traditional methods of evaluation for their accuracy, efficiency and cost-effectiveness.

This paper aims to study the importance of artificial intelligence in the music therapy. Artificial intelligence (IA) in the music therapy use many different approaches to meet your needs. In general, the types of experiences you might have fall into two broad categories: Active interventions and receptive interventions. In active interventions, the people can play instruments. But in receptive interventions, persons can spend some time together talking about music as a way to process their thoughts and feelings. However, the artificial intelligence will guide people in creating and/or listening to music during the session. Moreover, people can do one or more of the following activities: Create music, sing music, improvise, play on instrument, move to music and listen to music.

With the transformation of medical model and the promotion of the construction of healthy China, continuing to follow the traditional development model has been unable to meet the health needs of the new era. As early as the 1970s, traditional Chinese medicine tried to connect with artificial intelligence technology to seek a breakthrough in the development model of traditional Chinese medicine. In the new round of scientific and technological revolution and industrial reform, artificial intelligence, as an important driving force, presents new characteristics such as deep learning, cross-border integration and man-machine cooperation, which is more conducive to the organic combination with traditional Chinese medicine. As a traditional Chinese medicine therapy, the five elements music therapy, with its unique pertinence, can make the five tones and the five internal organs reach a harmonious resonance state when the patients listen, and make the body resonate, so as to smooth the spirit and express their feelings. Five elements music therapy has achieved great results in clinical psychotherapy, and through the combination with artificial intelligence, five elements music therapy can be presented to patients in a closer way. This paper aims to explore the development of this innovative way.

Now, with the development of the times, people’s pace of life is faster and faster, they are under more and more pressure in life and workplace, and the phenomenon of mental sub-health is becoming more and more common. How to relieve people’s pressure and help people establish a positive and healthy psychology has become an important problem. To solve this problem, music therapy is an effective and feasible method. The idea of music therapy has existed since ancient times. With the development of the times, the discipline of music therapy has developed in China for nearly 40 years. With the rapid development of science and technology, music therapy should also innovate music therapy with the development of the times and use computer technology. In this paper, a new music therapy system based on artificial intelligence is proposed to evaluate the efficacy of music therapy, monitor the physical condition of subjects in real time, and adjust the type of treatment music according to these information.

Background Given the high prevalence and cost of Alzheimer disease (AD), it is crucial to develop equitable interventions to address lifestyle factors associated with AD incidence (eg, depression). While lifestyle interventions show promise for reducing cognitive decline, culturally sensitive interventions are needed to ensure acceptability and engagement. Given the increased risk for AD and health care barriers among rural-residing older adults, tailoring interventions to align with rural culture and distinct needs is important to improve accessibility and adherence. Objective This protocol aims to develop an intelligent recommendation system capable of identifying the optimal therapeutic music components to elicit engagement and resonate with diverse rural-residing older adults at risk for AD. Aim 1 is to develop culturally inclusive user personas for rural-residing older adults to understand their goals and challenges for music-based digital health intervention. Aim 2 is to develop knowledge embedding–based machine learning (ML) models that use music metadata and survey response data to identify optimal therapeutic music components for enhancing engagement and emotional resonance for depression among rural-residing older adults at risk for AD. Aim 3 is to assess acceptability for personalized therapeutic music sessions and ML-based music recommendations with a separate sample. Methods Participants (N=1200) will be aged 55 years or older and residing in the United States. In phase 1, participants (n=1000) will receive 5 randomized songs and complete a survey to understand the sentiment, cultural relevance, and perceived benefit for each song. Brief, researcher-created Likert surveys will be used. In phase 2, survey data will be used to develop ML algorithms in collaboration with the University of Massachusetts Amherst Center for Data Science and Artificial Intelligence. These ML models will be integrated into the digital music intervention and tested with a separate sample of 200 participants. Similar to phase 1, participants will be provided with sets of songs generated by the recommendation system based on the target goal (ie, to reduce depression). The recommendation accuracy of the ML algorithm will be assessed using multiple performance metrics, including root-mean-square error and normalized discounted cumulative gain as well as the mean acceptability score with a goal of 85% user acceptability. Results Participant recruitment is complete for phases 1 and 2 as of June 2025. Data analysis for the results of aims 1, 2, and 3 are underway and results are expected to be published in the fall of 2025. Conclusions This protocol seeks to use ML to improve the equitability and accessibility of a digital lifestyle intervention for AD. International Registered Report Identifier (IRRID) DERR1-10.2196/73711

Music is widely being recognized as a non-pharmacological intervention for anxiety, yet its physiological mechanisms remain unexplored. This study examines the effect of Kaapi raga (a south Indian classical music) and weightless ambient music on cardiovascular regulation and anxiety reduction using heart rate variability (HRV) and mean arterial pressure (MAP). Twenty participants with high anxiety (HAM-A >25) completed a protocol involving rest, music listening, and attention tasks while ECG and MAP were continuously recorded.HRV features were extracted across time, frequency, and nonlinear domains and analysed using machine learning classifiers. Random Forest and XGBoost achieved the highest performance (70.8% accuracy; AUC up to 0.89), outperforming linear models. Feature ablation revealed non-linear SD2, pNN50, HR mean, and high-frequency power percentage as key predictors of music-induced physiological changes.The findings here demonstrate that ensemble machine learning methods can robustly model music’s impact on anxiety, while domain-specific HRV features provide actionable biomarkers for developing personalized, evidence-based music therapy interventions.

Music therapy is the treatment of college students' psychology through various techniques and methods of music, and this paper focuses on researching and analyzing the improvement effect of music therapy on college students' mental health in the context of cultural education. Students' physiological data are collected and denoised, and machine learning models are used to realize the multimodal fusion of all kinds of physiological signal features to obtain the objective psychological state assessment values of college students. The subjective assessment results of the mental health assessment scales were then combined to analyze the improvement effect of mental health of college students in the music therapy intervention. The analysis of the psychological health status of the students before and after the intervention experiment revealed that the objective assessment values of the psychological state of the college students in the intervention group gradually tended to be positive with the music therapy, and the subjective assessment results of the psychological health scales of the students in the intervention group were signi(cid:976)icantly better than those of the non-intervention group after the experiment (P<0.05). Music therapy has a signi(cid:976)icant role in intervening in the mental health of college students and resolving their psychological malaise, which is of great practical and guiding signi(cid:976)icance in improving the psychological tolerance and health of college students.

With the development of the entertainment media industry, music has become an essential part of people's lives, but in the face of massive music data, how to recommend appropriate music according to different environments as well as different moods of users has become a hot research issue in recent years. In this context, mood-based music classification task has received more and more attention. Music is a kind of art that expresses the emotions of human life, so the emotions contained in music can be identified by extracting music features through computer technology. In this paper, a music emotion classification algorithm is proposed based on deep belief network, combined with convolutional neural network. Experiments have proved that the method has better results in music emotion classification task than both traditional machine learning methods and original deep belief networks to improve. Music is a special world language and the language closest to the human heart. Music is a product of human emotional communication. It can infect the listener through the unique musical emotions, the emotions expressed by different music are obviously different, but the emotions expressed by the same music are also different from person to person; the emotional changes produced by the listener due to the music are also similar to the emotional changes in the listener's heart in a certain way, so it is said that the music is also the closest language to the human mind. In this paper, based on the big data algorithm foundation of deep learning, we have carried out a certain research on music intervention in college students' emotion regulation. There has been some new groundbreaking research on computer fusion music for mood regulation.

The harmful impacts of excessive stress on people’s health have been widely acknowledged, necessitating effective methods for its identification. Recognizing the importance of early stress detection and intervention, this research aims to contribute to the field of healthcare. To achieve this objective, this study classifies electrocardiogram (ECG) signals by assessing physio-psychological states, specifically stress and examines the role of music therapy in alleviating stress. ECG signals, recorded both before and after a music therapy session, were collected. Using signal processing techniques, essential features were extracted from these ECG signals, resulting in a more accurate identification of stress. Additionally, through experimentation and model evaluation, k-nearest Neighbors (KNN) and Classification and Regression Trees (CART) were determined to be the most effective models for this classification. Both models consistently yielded 90% accuracy. These identified extracted features and models are vital to effectively recognizing stress in ECG signals, offering valuable insights for future studies and clinical applications. This research contributes not only to the development of tools for stress detection but also to the understanding of the therapeutic impact of music.

Music education is crucial to cognitive and emotional well-being. Nevertheless, the traditional approaches are not personalized and are unable to dynamically respond to the affective states of the learners, restricting their influence on the emotional interest. As a solution to this void, we will design an Adaptive Emotion-Aware Music Education System (AEMES) which can use the techniques of machine learning to monitor, model and respond to the real-time emotions of learners. The originality of this framework is that it is based on an integrated combination of emotion modeling, adaptive strategy selection, and personalized intervention mechanisms that allow the system to achieve continuous engagement of learners and promote emotional wellbeing. The suggested approach is strictly tested on a variety of datasets that include multimodal responses of learners, which include physiological data, facial expressions and behavioral interactions. In comparison with state-of-the-art methods that include CNN-LSTM, Audio-Transformer, and Visual-Transformer models, it is shown that AEMES performs much better than the existing methods in standard performance measures. The proposed framework has a quantitative performance of an accuracy of 87.2%, F1-score of 85.7% and AUROC of 0.91 compared to baseline models, which have a margin of over 4–8% across measures. The role of emotion modeling and adaptive strategies to the overall performance of the system is also confirmed by ablation studies. Emotional pathways (both in terms of time trends and multi-emotional interactions) visualization illustrates the possibility of the framework to dynamically control emotion of learners and ensure that it does not go negative during learning sessions. The overall results support the claim that AEMES is effective in providing personalized, emotion-sensitive music education, which would prove to be a powerful method of improving the learning process and emotional wellbeing.

No abstract available

Chronic pain is a prevalent concern for older individuals, often leading to a decline in mental well-being, especially through conditions like depression. This study explores the potential effectiveness of Music Intervention (MI) as a non-pharmacological approach to alleviate depressive symptoms in those experiencing chronic pain. Existing methodologies lack predictive accuracy, prompting the introduction of the Predicting Chronic Pain-based Music Intervention (PCP-MI) model. Utilizing machine learning, the PCP-MI model customizes music treatments based on individual characteristics and preferences, showcasing promising results across various metrics related to pain, anxiety, heart and respiratory rates, pain tolerance, emotional well-being, quality of life, and depression severity. The PCP-MI method demonstrated a mean performance across multiple metrics, encompassing pain intensity (17.75%), anxiety level (27.79%), heart rate (78.30 bpm), respiratory rate (15.90 bpm), pain tolerance threshold (59.37 seconds), emotional well-being (75.56%), quality of life (74.81%), and depression severity (65.27%). This research suggests a promising avenue for enhancing the psychological well-being of a vulnerable group, representing a significant advancement in comprehensive pain treatment approaches.

Background Prolonged Disorders of Consciousness (PDOC) resulting from severe acquired brain injury can lead to complex disabilities that make diagnosis challenging. The role of machine learning (ML) in diagnosing PDOC states and identifying intervention strategies is relatively under-explored, having focused on predicting mortality and poor outcome. This study aims to: (a) apply ML techniques to predict PDOC diagnostic states from variables obtained from two non-invasive neurobehavior assessment tools; and (b) apply network analysis for guiding possible intervention strategies. Methods The Coma Recovery Scale-Revised (CRS-R) is a well-established tool for assessing patients with PDOC. More recently, music has been found to be a useful medium for assessment of coma patients, leading to the standardization of a music-based assessment of awareness: Music Therapy Assessment Tool for Awareness in Disorders of Consciousness (MATADOC). CRS-R and MATADOC data were collected from 74 PDOC patients aged 16–70 years at three specialist centers in the USA, UK and Ireland. The data were analyzed by three ML techniques (neural networks, decision trees and cluster analysis) as well as modelled through system-level network analysis. Results PDOC diagnostic state can be predicted to a relatively high level of accuracy that sets a benchmark for future ML analysis using neurobehavioral data only. The outcomes of this study may also have implications for understanding the role of music therapy in interdisciplinary rehabilitation to help patients move from one coma state to another. Conclusions This study has shown how ML can derive rules for diagnosis of PDOC with data from two neurobehavioral tools without the need to harvest large clinical and imaging datasets. Network analysis using the measures obtained from these two non-invasive tools provides novel, system-level ways of interpreting possible transitions between PDOC states, leading to possible use in novel, next-generation decision-support systems for PDOC.

Wearable technology is very helpful in addressing the major health concern of the public, stress detection and its treatment. Personalized intervention recommendations are still not widely used, despite the fact that machine learning algorithms have demonstrated great accuracy in stress classification. This work uses simulated approach to compare pre vs post-intervention data statistically and combines machine learning for recommending the right intervention to the stressed individual at right time. Paired t-tests, repeated measures ANOVA, and post-hoc analysis are used to assess four non-pharmacological interventions: breathing exercises, music therapy, mindfulness, and sleep optimization. The findings support the need for personalized recommendations, as all therapies significantly improved physiological parameters (p < 0.001). Breathing intervention has the biggest impact on lowering heart rate (Cohen’s d = 0.70), but mindfulness has a more powerful effect for reducing sympathetic arousal, and individual variability for interventions is also found. Using Random Forest, machine learning models trained on simulated data achieved up to 94.8% test accuracy, demonstrating the potential for data-driven intervention recommendations. This concept offers a clear, empirically supported method for managing stress with wearable technology.

Pain is an unpleasant experience the neurobiology of which is influenced by psychosocial factors including negative affect. Music is a ubiquitous experience that can improve affect, potentially decreasing anxiety and catastrophizing, both of which are associated with greater pain severity. We hypothesized that a machine-learning generative music intervention in the form of a smartphone web app (Unwind) could be used to modulate the experience of pain. In this pilot study, we recruited 15 individuals with acute pain who were admitted to an observation unit in the emergency department, and were being treated with opioids. Participants used the music intervention (Unwind) during this brief hospitalization, after which we assessed their response to its use through a semi-structured qualitative interview. Overall, participants responded positively to Unwind. While some reported some technical challenges, participants were willing to continue using it at home. In particular, participants reported using Unwind to address their anxiety, and many used it to facilitate sleep in the presence of pain. This study demonstrates that individuals with acute pain will accept and use a smartphone-based music protocol.

BackgroundMusic-based interventions (MBIs) are gaining recognition as non-pharmacological strategies to support individuals with dementia. Rubato is an AI-driven music platform that uses machine learning and deep neural networks to deliver personalized music interventions aimed at reducing stress and anxiety through real-time heart rate variability monitoring.ObjectiveTo evaluate a personalized music-based intervention for individuals with mild dementia.MethodsIn this randomized controlled trial, twenty participants with mild dementia, including Alzheimer's disease, were enrolled, with ten randomly assigned to each of the intervention and control groups. Intervention participants used the Rubato app for 30 minutes daily over 12 weeks, while control participants received standard care. Pre- and post-intervention questionnaires assessed cognition (Mini-Mental State Examination), anxiety (State-Trait Anxiety Inventory), quality of life (Dementia Quality of Life Questionnaire), depression (Geriatric Depression Scale), and memory (Immediate Prose Memory Test).ResultsAnxiety and depression scores decreased significantly following the MBI in the intervention group compared to controls (p = 0.037 and p = 0.028, respectively). Overall, 80% of the intervention participants showed reduced anxiety compared to 30% of controls, and 70% showed reduced depression compared to 20% of controls. No statistically significant differences were found in cognition, quality of life, or memory, although trends toward improvement were noted in the intervention group.ConclusionsThese preliminary findings suggest that Rubato's MBI may reduce anxiety and depression in individuals with dementia, with clinically meaningful benefits despite limited statistical power. Larger studies with longer follow-up are needed to validate these findings and further examine effects on cognition and quality of life.

This work hereby proposes a new framework in the classification of stages of sleep, followed by personalized music therapy, in treating sleep disorders related to insomnia, sleep apnea, and narcolepsy. The proposed system is empowered with state-of-the-art machine learning algorithms applied to PSG data to enhance the detection of a variety of stages of sleep using real-time physiological signals such as EEG, EOG, and EMG. This system, Dream Tunes, provides personalized music therapy adapted to the detected sleep stages and thus improves quality sleep and, in general, therapeutic outcomes. The use of machine learning for treatment with music is an effective non-invasive intervention that can enhance sleep health. Testing confirms the viability of the approach proposed by higher accuracy in sleep stage detection and personalized music recommendation. Thus, the given study represented a huge development in the field of somnology with the use of technology and therapeutic ways to support better management of sleep.

Mental health issues often stem from subtle mood fluctuations that, if left unaddressed, can escalate into chronic psychological conditions. This paper introduces MindTune, an AI-driven mood recognition system that integrates personalised music therapy to regulate emotions and enhance mental well-being. The system uses deep learning and natural language processing (NLP) to detect emotional shifts through textual and physiological data, triggering the automatic selection of therapeutic music tailored to counteract negative moods. Unlike existing mental health detection models that focus on post-diagnosis treatment, MindTune aims at early-stage intervention, preventing mental health deterioration by addressing minor mood disturbances in real- time. The proposed system leverages machine learning algorithms to continuously refine its mood classification and song recommendation model, offering an adaptive, user-centric approach to emotional regulation. Experimental results demonstrate the system's effectiveness in enhancing mood stability and reducing stress indicators, highlighting its potential as an AI-powered tool for proactive mental health management.

Music therapy is acknowledged as one of the effective non-pharmacological interventions to regulate neuronal function enhancing mental health. This study examines them with biosensing technology, including EEG, HRV, GSR, and fNIRS to record participants’ neurophysiological responses in real time, with a particular focus on biomechanical aspects. Hence, participants listened to controlled music intervention consisting of classical, ambient and binaural beats while biosensors recorded not only neural oscillations but also muscle tension and movement patterns. Since there are close links between music stimuli and neuronal regulation, data analysis through wavelet transformation and machine learning models enabled the discovery of significant patterns. The research revealed that alpha wave synchronization which occurred in the frequency range of 8–13 Hz and theta/delta binaural beats substantially facilitated the levels of relaxation, mood and cognition. HRV analysis revealed lowering of sympathetic values in the experimental group, thus proving the stress-relieving impact of music therapy, which may also lead to improved muscle relaxation and reduced physical tension.

To address the long-standing challenges children with autism face in social skills and emotional regulation, this study introduces Emotion-based Music Intelligent Network (EmoMusik-Net)—a deep learning model designed for intelligent music therapy. The model focuses on emotional impairments exhibited during social interactions, integrating Transformer-based temporal modeling with a Transfer Learning-based Graph Convolutional Network (TL-GCN). This combination enables high-precision recognition of facial expression sequences and supports a dynamically adaptive, closed-loop mechanism for personalized music recommendation. EmoMusik-Net was trained and optimized using three publicly available emotional video datasets. A pre- and post-intervention study, conducted in collaboration with the families of 182 children with autism, employed questionnaire-based assessments to systematically evaluate the model’s real-world feasibility and effectiveness. Experimental results demonstrated that EmoMusik-Net achieved an emotion recognition accuracy above 0.970, an F1-score consistently over 0.960, and an Area Under the Curve (AUC) of 0.978. The model also showed outstanding robustness on large-scale datasets, with a stability score of 0.994, indicating strong classification performance and generalizability. In terms of intervention outcomes, boys aged 1–6 showed a marked increase in social interest scores, rising from 1.280 to 2.540—a 98.44% improvement. Girls aged 7–12 exhibited significant gains in emotional response scores, from 1.670 to 3.120—an 86.77% increase. Further statistical analysis using the Mixed-effects Model for Repeated Measures (MMRM) and bootstrap confidence interval estimation confirmed the intervention’s significance both statistically and clinically, with particularly strong effects observed in younger participants. Expert blind evaluations further validated the system’s effectiveness, showing high consistency in rhythm and emotion matching. The Intraclass Correlation Coefficient (ICC) ranged from 0.75 to 0.91, with matching accuracy surpassing 94% in certain subgroups. EmoMusik-Net not only addresses the current research gap in integrating intelligent emotion recognition with music-based interventions but also offers a responsive, technology-driven support tool for parents, educators, and clinicians. This approach holds strong potential to advance autism spectrum disorder interventions toward personalized, data-driven methodologies.

In a time when technology and therapeutic approaches are increasingly interacting, this study aims to investigate the game-changing combination of facial expression detection and music therapy. Based on the knowledge that emotions are observable, we provide a new method: continuously monitoring facial expressions using a camera in conjunction with real-time algorithms for recognition. Finding times of despair, tension, and disdain is the main goal. and deftly recommending musical therapies to relieve these emotional conditions. This study is motivated by the understanding of music’s significant influence on mental health and the unrealized promise of technology to improve therapy. The proposed method uses facial expression signals to build a dynamic and responsive technology that adjusts music suggestions in real time to meet each person’s unique emotional demands. The approach incorporates state-of-the-art facial expression recognition technology. Technology can capture the subtleties of emotional states. Participants take part in a setting where their emotions serve as the map for a well-chosen musical voyage. The insights garnered from this study may pave the way for a more nuanced and tailored approach to music therapy, harnessing the power of technology to enhance emotional resilience and foster a harmonious coexistence between mind and melody.

Stress is a common issue that affects decisionmaking, productivity, and overall health. Conventional stress assessment methods often rely on lengthy surveys or specialized medical devices, which may not be practical for everyday use. In this work, we propose a real-time stress detection system based on facial emotion recognition using a live camera feed. The system classifies stress levels as high, moderate, or low and provides instant feedback to the user. Along with recognition, the system includes music-based interventions to reduce stress, combining detection with relief. The model is designed to be accessible, relying on regular webcams, freely available frameworks, and efficient algorithms so that it can be applied in everyday scenarios without specialized equipment. Possible applications include workplace monitoring, educational support, and personal wellness systems. To make it practical, the system is designed to work with easily available webcams, open-source frameworks, and efficient algorithms. This setup allows it to run smoothly in everyday situations. This approach aims to make stress monitoring less intrusive and more usable for nonclinical applications. The framework was trained and validated using established facial expression datasets such as FER2013 and Affect Net. Evaluation results indicate that the integrated system achieves an accuracy of over 82 % in identifying stress-related emotional states. in distinguishing stress-related emotional states. By merging recognition and intervention in one framework, the system aims to promote healthier coping habits in daily life.

We, the Editor-in-Chief and Publisher of Disability and Rehabilitation: Assistive Technology, have retracted the following articles, which were published as part of a planned Special Issue titled “AI empowered Assistive Technology” and edited by Uzair Aslam Bhatti: Liu, Y., & Long, T. (2025). How does social media shape the duration of formal care for disabled elderly? Moderating effects based on digital competence and psychological expectancy. Disability and Rehabilitation: Assistive Technology, 20(7), 2146–2159. https://doi.org/10.1080/17483107.2025.2525539 Lu, X., & Guo, W. (2025). Utilizing artificial intelligence to enhance social connections – the alleviating effect of emotionally intelligent chatbots on loneliness. Disability and Rehabilitation: Assistive Technology, 1–11. https://doi.org/10.1080/17483107.2025.2540494 Xuanfeng, D. (2025). Gesture recognition and response system for special education using computer vision and human–computer interaction technology. Disability and Rehabilitation: Assistive Technology, 1–18. https://doi.org/10.1080/17483107.2025.2527226 Ren, H., Lv, J., Ren, L., & Guo, R. (2025). Optimizing the spatial allocation of pension resources in downtown Shanghai: a dynamically updated genetic algorithm approach. Disability and Rehabilitation: Assistive Technology, 1–14. https://doi.org/10.1080/17483107.2025.2536173 Kadiyala, B., Nippatla, R. P., Boyapati, S., Vasamsetty, C., Alavilli, S. K., & M, T. (2025). Prevention and health care intervention of common injuries in long-distance running for college teachers. Disability and Rehabilitation: Assistive Technology, 1–20. https://doi.org/10.1080/17483107.2025.2564371 Zhang, L., Wang, G., & Wang, M. (2025). Impacts of virtual reality-aided physical training on the balance ability and energy expenditure of children with moderate intellectual disability. Disability and Rehabilitation: Assistive Technology, 1–16. https://doi.org/10.1080/17483107.2025.2546026 Zhang, S., & Meng, Q. (2025). Intelligent sports rehabilitation: integrating deep learning and real-time monitoring to achieve personalized rehabilitation. Disability and Rehabilitation: Assistive Technology, 1–15. https://doi.org/10.1080/17483107.2025.2559187 Gao, P., Wei, D., Lu, W., & Zhai, Y. (2025). Unveiling the mechanisms influencing service efficiency in Chinese hospital portal sites: evidence from DEA and fsQCA analysis. Disability and Rehabilitation: Assistive Technology, 1–16. https://doi.org/10.1080/17483107.2025.2554956 Chen, J., Chen, G., & Wang, X. (2025). Configuration analysis of mobile phone addiction among college students: findings from fsQCA method. Disability and Rehabilitation: Assistive Technology, 1–14. https://doi.org/10.1080/17483107.2025.2548858 Ye, L., Du, J., & He, Z. (2025). Willingness to communicate with AI chatbots in English: the role of personality and trust among Chinese college students. Disability and Rehabilitation: Assistive Technology, 1–16. https://doi.org/10.1080/17483107.2025.2572532 Wan, L., Gao, W., Xi, P., Xu, K., Li, T., Wu, J., & Wu, D. (2025). AI-Empowered assistive technology for optimizing specimen submission in obstetrics and gynecology: integrating DeepSeek with the ADDIE model. Disability and Rehabilitation: Assistive Technology, 1–11. https://doi.org/10.1080/17483107.2025.2561248 Liao, Z. (2025). Artificial intelligence-enabled integration of “Fu” culture into the moral education system for special education across K-12 and higher education: a soft computing approach. Disability and Rehabilitation: Assistive Technology, 1–13. https://doi.org/10.1080/17483107.2025.2555541 Cen, X. (2025). Multimodal deep learning methods for speech and language rehabilitation: a cross-sectional observational study. Disability and Rehabilitation: Assistive Technology, 1–13. https://doi.org/10.1080/17483107.2025.2551708 Wang, Y., Li, H., Du, Y., Zhang, P., Shi, S., Ma, Y., & Wang, S. (2025). The relationship between job competence, demographic characteristics and professional misconduct among medical staff in the AI era. Disability and Rehabilitation: Assistive Technology, 1–12. https://doi.org/10.1080/17483107.2025.2561247 Xiao, B., Sun, H., Lin, S., Li, J., & Yu, Y. (2025). Virtual classrooms, real resilience: how AI tutors enhance cognitive rehabilitation in pediatric cancer survivors. Disability and Rehabilitation: Assistive Technology, 1–14. https://doi.org/10.1080/17483107.2025.2549898 Li, M. (2025). Evaluating the role of social support systems in enhancing the well-being of elderly individuals with disabilities. Disability and Rehabilitation: Assistive Technology, 1–21. https://doi.org/10.1080/17483107.2025.2568945 Xiao, X. (2025). Choral harmony: the role of collective singing in ritual, cultural identity and cognitive-affective synchronisation in the age of AI. Disability and Rehabilitation: Assistive Technology, 1–17. https://doi.org/10.1080/17483107.2025.2556025 Maashi, M., Alhefdhi, A., Alanazi, F., & Rizwanullah, M. (2025). Enhancing communication for people with hearing disabilities through robust sign language recognition using deep learning and the internet of things. Disability and Rehabilitation: Assistive Technology, 1–19. https://doi.org/10.1080/17483107.2025.2562454 Cheng, W., Guo, J., & Lu, D. (2025). CHATWELL: an AI-enabled adaptive tutoring system for improving mandarin composition skills in L2 students with learning difficulties. Disability and Rehabilitation: Assistive Technology, 20(7), 2358–2374. https://doi.org/10.1080/17483107.2025.2554955 Zhang, R. (2025). Personalised sports rehabilitation analysis using a fitness enhanced model based on big data and deep learning. Disability and Rehabilitation: Assistive Technology, 1–14. https://doi.org/10.1080/17483107.2025.2561926 Xue, Q., Yang, J., & Liu, X. (2025). An AI-integrated spatial auxiliary decision-support model for historic and cultural blocks based on the spatial triad. Disability and Rehabilitation: Assistive Technology, 1–17. https://doi.org/10.1080/17483107.2025.2554331 Huang, R. (2025). Between intimacy and surveillance: elderly users’ adaptation and privacy negotiation in human-robot interaction. Disability and Rehabilitation: Assistive Technology, 1–20. https://doi.org/10.1080/17483107.2025.2566379 Wang, P., Mi, B., Lu, J., & Zheng, F. (2025). Evaluating AI tutor feedback in medical education: a case study of computer basics and applications course for undergraduates. Disability and Rehabilitation: Assistive Technology, 1–15. https://doi.org/10.1080/17483107.2025.2560673 Liu, J., Liu, L., Wu, Y., Wang, Z., & Li, X. (2025). MRI feature engineering and SVM framework for schizophrenia recognition. Disability and Rehabilitation: Assistive Technology, 1–25. https://doi.org/10.1080/17483107.2025.2569801 Zhu, D., Xu, Q., & Jiang, C. (2025). Good intentions may be wrong: evidence from anonymous donation behaviour on Chinese charitable crowdfunding platform. Disability and Rehabilitation: Assistive Technology, 1–14. https://doi.org/10.1080/17483107.2025.2565405 Xiao, B., Lin, S., Yu, Y., Li, J., & Li, X. (2025). AI-Enhanced assistive interventions for adolescent cyberbullying: a gender-sensitive moderated mediation approach. Disability and Rehabilitation: Assistive Technology, 1–19. https://doi.org/10.1080/17483107.2025.2570890 Xiong, P., & Zhang, Y. (2025). Artificial intelligence as an assistive technology in language education: investigating TPACK, self-efficacy, and organisational support. Disability and Rehabilitation: Assistive Technology, 1–19. https://doi.org/10.1080/17483107.2025.2561927 Yu, Z., Xu, Z., & Qi, J. (2025). Disability-oriented data protection in AI-enabled assistive technologies: bridging gaps in China’s legal framework. Disability and Rehabilitation: Assistive Technology, 1–25. https://doi.org/10.1080/17483107.2025.2568940 Xiao, B., Xing, Q., Ye, J., & Nie, Y. (2025). The impact of social media on adolescent conformity: mechanisms, individual moderators, and implications for AI-Empowered assistive technologies. Disability and Rehabilitation: Assistive Technology, 1–19. https://doi.org/10.1080/17483107.2025.2570889 Yan, X., & Chen, H. (2025). From cognitive alignment to technological adaptation: understanding rural digital governance through AI-augmented human–data collaboration. Disability and Rehabilitation: Assistive Technology, 1–24. https://doi.org/10.1080/17483107.2025.2564370 Sun, H., Tang, X., & Jiang, Y. (2025). Dual-channel conduction modulated by digital literacy: examining technology shock awareness and informal digital learning through an AI-enabled human-computer interaction lens. Disability and Rehabilitation: Assistive Technology, 1–17. https://doi.org/10.1080/17483107.2025.2573221 Liu, Z. (2025). Hierarchical attention mechanism in deep learning improving music therapy rehabilitation through context aware emotion mapping. Disability and Rehabilitation: Assistive Technology, 1–15. https://doi.org/10.1080/17483107.2025.2568164 Following publication, the Publisher identified concerns regarding the peer review process. An investigation by the Taylor & Francis Publishing Ethics & Integrity team in full cooperation with the Editor-in-Chief concluded that the articles included in this Special Issue were not peer-reviewed appropriately in line with the Journal’s peer review standards and policy. As the stringency of the peer review process is core to the integrity of the publication process, the Editor-in-Chief and Publisher have decided to retract all of the articles within the above-named Special Issue. The Editor-in-Chief and the Publisher have not confirmed if the authors were aware of this compromised peer review process. The authors have been informed of this decision. We have been informed in our decision-making by our editorial policies and the COPE guidelines. The retracted articles will remain online to maintain the scholarly record, but they will be digitally watermarked on each page as ‘Retracted’.

Music regulates emotions, alleviating negativity and promoting positive mental states. Music therapy is widely used, notably for autistic patients and college students' mental well-being. College students face high stress, poor sleep, and reluctance to seek support, making music therapy particularly suitable. This paper present EMO-Music, which uses a smartwatch to intervene positively by recognizing users' emotions and playing music accordingly. It aims to transform music apps into functional tools for enhanced therapy. User experiments using the SUS scale are conducted to evaluate the feasibility of our system.

This research paper focuses on creating an advanced Music Recommendation System through a five-phase approach. Beginning with Data Collection and Extraction, the project utilizes the FER dataset for face images and the Spotify Web API to gather a diverse collection of musical tracks. Subsequent Data Cleaning and Preprocessing procedures refine the datasets, including streamlining the FER dataset for essential emotions and organizing music data in a database for accessibility. The research's core lies in Model Building, employing a custom Convolutional Neural Network (CNN) model named FERNET for precise emotion classification. Rigorous Model Optimization ensures an impressive 96% accuracy rate in emotion classification. The project concludes with the development of the Mood-Music App, a user-friendly application integrating the FER module, Emotion Recognition module, and Songs Recommendation module. This app utilizes real-time emotion analysis to provide a personalized and immersive music discovery experience, revolutionizing user engagement with music. The research showcases a comprehensive methodology, from data collection to the deployment of a sophisticated app, contributing significantly to the field of Music Recommendation Systems.

Aiming at the problems of weak emotional expression range, large modal differences and lack of dynamic adaptation of music intervention in the elderly, this paper proposes a deep learning based emotion recognition and personalized music therapy system for the elderly. This method fuses speech, face, physiological and interaction information, constructs a convolutional coding, bidirectional temporal modeling and cross-modal attention coordination framework, and realizes the stable discrimination of emotional states in the elderly. On this basis, a recommendation mechanism combining user preference, historical feedback and music content characteristics is introduced to form a closed loop of "identification-recommendation-update". The experimental results show that the Accuracy, Macro-F1 and AUC of the proposed model reach 93.84%, 92.47%and 95.12%respectively. After 4 weeks of intervention, the emotional improvement rate of the experimental group increases to 27.6%, and the average inference delay of the system is 23.6 ms. The research shows that this method has good application potential in intelligent elderly care and digital music therapy scenarios.

Music therapy is also an efficient non-pharmacological therapeutic approach to emotional regulation and mental health rehabilitation, and yet the clinical practice is still based on subjective judgments of therapists. Objective, automated emotion recognition could enhance the accuracy, flexibility, and scalability of therapy sessions by supplementing clinical judgment with data-driven affective feedback. This paper presents a multimodal deep learning framework for music-induced emotion recognition that integrates music-induced affect modeling with physiological affect representation learning in a staged, cross-dataset design. The structure includes two branches: an audio emotion encoder, trained on the DEAM and PMEmo 2018 datasets to continuously predict valence–arousal from musical stimuli, and a physiological affect encoder, trained on the DEAP and AMIGOS datasets using TorchEEG to classify emotions from EEG and peripheral signals. A cross-modal attention fusion module aligns and combines the learned representations from both branches into a shared embedding space for joint emotion inference. To be methodologically transparent, the framework employs a two-stage experimental approach: Stage 1 focuses on learning music emotion representations from audio datasets, and Stage 2 focuses on physiological affect learning and cross-dataset fusion validation. Importantly, since there is no publicly available synchronized clinical music therapy dataset, the two branches, audio and physiological, are trained on separate public proxy datasets and do not use audio from real clinical music therapy sessions. Therefore, the reported cross-dataset transfer validation should be considered to be a cross-platform level validation of multimodal integration in proxy conditions rather than direct transfer from proxy datasets to real clinical music therapy conditions. Experimental findings indicate that the proposed attention-based intermediate fusion method outperforms unimodal baselines and early- and late-fusion approaches in terms of accuracy, F1-score, and concordance correlation coefficient. Ablation and cross-dataset studies also testify to the strength of the suggested multimodal design.

<jats:p />

With the advancement of modern medical concepts, the beneficial effects of music on human health have gradually become accepted, and the corresponding music therapy has gradually become a new research direction that has received much attention in recent years. However, folk music has certain peculiarities that lead to the fact that there is no efficient way of selecting repertoire that can be carried out directly throughout the repertoire selection. This paper combines deep learning theory with ethnomusic therapy based on previous research and proposes a deep learning-based approach to ethnomusic therapy song selection. Since the feature extraction process in the traditional sense has insufficient information on each frame, excessive redundancy, inability to process multiple frames of continuous music signals containing relevant music features and weak noise immunity, it increases the computational effort and reduces the efficiency of the system. To address the above shortcomings, this paper introduces deep learning methods into the feature extraction process, combining the feature extraction process of the Deep Auto-encoder (DAE) with the music classification process of Gaussian mixture model, which forms a new DAE-GMM music classification model. Finally, in terms of music therapy selection, this paper compares the music selection method based on co-matrix and physiological signal with the one in this paper. From the theoretical and simulation plots, it can be seen that the method proposed in this paper can achieve both good music classifications from a large number of music and further optimize the process of music therapy song selection from both subjective and objective aspects by considering the therapeutic effect of music on patients. Through this article research results found that the depth of optimization feature vector to construct double the accuracy of the classifier is higher, in addition, compared with the characteristics of the original optimization classification model, using the gaussian mixture model can more accurately classify music, the original landscape “hometown” score of 0.9487, is preferred, insomnia patients mainly ceramic flute style soft tone, without excitant, low depression, have composed of nourishing the heart function.

Facial emotion recognition has been recognised as an important aspect of human–computer interaction, enabling computers to understand users' emotional states and interact intelligently with them. This work presents a real-time facial emotion recognition framework coupled with an automated music retrieval system. The system utilises a pre-trained deep residual network based on the ResNet50 model to recognise seven basic facial emotions: anger, contempt, disgust, fear, happiness, sadness, and surprise. Face regions are detected using the Haar Cascade classifier applied to webcam input. The emotional state is used as input to formulate a query to retrieve the appropriate music playlist over the internet. The system has been designed with a cooldown feature that improves the overall user experience by controlling the frequency of music playlist updates. The experimental results show a validation accuracy of 94.46%, thereby demonstrating the system's classification capability

No abstract available

In the treatment of children with autistic spectrum disorder (ASD) through music perception, the perception effect and the development of the disease are mainly reflected in the fluctuations of the electroencephalogram (EEG), which is clinically effective on the brain. There is an inaccuracy problem in electrogram judgment, and deep learning has great advantages in signal feature extraction and classification. Based on the theoretical basis of Deep Belief Network (DBN) in deep learning, this paper proposes a method that combines the optimized Restricted Boltzmann machine (RBM) feature extraction model with the softmax classification algorithm. Brain wave tracking analysis is performed on children with autism who have received different music perception treatments to improve classification accuracy and achieve the purpose of accurately judging the condition. Through continuous adjustment and optimization of the weight matrix in the model, a stable recognition model is obtained. The simulation results show that this optimization algorithm can effectively improve the recognition performance of DBN, with an accuracy of 94% in a certain environment, and has a better classification effect than other traditional classification methods.

No abstract available

Music can express and influence a wide range of emotional states and feelings in humans. The development of a system for recognizing emotions based on music analysis has generated significant interest among academic and industrial communities due to its applications in various fields such as human-machine interaction, music recommendation systems, music therapy, and so on. Music emotion recognition (MER) is the process of analysing and classifying the affective states conveyed by a piece of music. A survey of existing work on emotional music processing is indeed very helpful for carrying out further research in the field of music emotion recognition. Due to the importance of emotion recognition in Music Information Retrieval (MIR) research, a comprehensive survey is provided in this paper with a detailed study of emotion models, features, and various music databases. This paper emphasizes the machine learning and deep learning approaches used for MER to extract emotions from music. Finally, the paper is summarized with some possible future research directions.

This paper presents a novel emotion-based music recommendation system that integrates real-time facial expression analysis with personalized music selection using deep learning techniques. The proposed system employs Convolutional Neural Networks (CNNs) for accurate emotion recognition from facial expressions and automatically suggests music based on the detected emotional state. The system is trained using the FER2013 dataset and utilizes OpenCV for efficient face detection, implemented through a user-friendly graphical interface. Experimental results demonstrate 87.3% accuracy in emotion classification across five fundamental emotions (angry, happy, neutral, sad, surprise) with effective real-time music recommendation capabilities. The system represents a significant advancement in personalized multimedia experiences by seamlessly integrating computer vision, deep learning, and music therapy principles to enhance user emotional well-being through adaptive music selection.

The use of music therapy has shown results in improving both physiological and psychological health. In this research, we applied deep learning methods to forecast how calming music affects individuals. Our study involved creating a model using a dense neural network (DNN) to analyze data collected from a Muse2 device, which tracks metrics like heart rate, mind calm, and stillness before and after listening to music. By utilizing the TensorFlow and Keras frameworks, we processed time series data to explore the impact of music therapy on these responses. Our discoveries demonstrate that soothing music significantly reduces heart rate and improves calmness and stillness scores. We developed a structured data processing and analysis pipeline, including careful data preprocessing, dense neural network (DNN) model design, and rigorous performance evaluation. We established an approach for predicting shifts in these Muse2-tracked metrics following music therapy sessions. This study contributes to crafting settings that can enhance overall well-being with potential applications in fields such as psychology, medicine, and consumer electronics. By bridging theory with application, our research lays the foundation for targeted and efficient use of music-based interventions aimed at promoting relaxation and better health through specific physiological and mental indicators.

This paper proposes an interactive system based on spectrogram analysis, deep neural network and wavelet analysis aiming at the complexity and subjectivity of music emotion recognition. The system first uses a spectrogram to capture the time-frequency characteristics of music signals and then automatically extracts the deep emotion-related features through a convolutional neural network (CNN). This paper introduces the Mallat algorithm for wavelet decomposition to enhance the local details of audio signals to improve the accuracy of feature extraction. The experimental results show that the system performs well in recognizing music emotions, and the accuracy is significantly improved compared with the traditional method. In addition, the system supports real-time interaction, allowing users to personalize music experience by adjusting emotional labels, thus showing broad application prospects in music therapy, game entertainment and other fields. This study promotes the development of music emotion recognition technology and provides a new perspective for further exploration of deep learning in interdisciplinary applications.

Music plays a crucial role in shaping emotions and experiences, making its classification an important area of research with applications in therapy, recommendation systems, and affective computing. This study develops a deep learning-based system to classify music into three emotional categories: "Angry," "Happy," and "Sad." The dataset, consisting of 22 audio files collected from YouTube, was manually labelled, segmented into 30-second clips, and augmented using pitch shifting and time stretching to enhance diversity. Features were extracted using Mel-Frequency Cepstral Coefficients (MFCC) and spectral contrast to analyse the harmonic and timbral characteristics of the audio. Three deep learning models, CNN, CNN-LSTM, and CNN-GRU, were evaluated. CNN-GRU achieved the highest weighted accuracy of 99.10%, demonstrating superior performance. Future work includes adding more emotion categories, diversifying the dataset, exploring advanced architectures like transformers, optimising hyperparameters, implementing real-time applications, and conducting user studies to assess effectiveness. This research successfully developed and evaluated a music emotion classification system, contributing to advancements in the field.

This study introduces an emotion-aware music recommendation system that adapts song suggestions in real time based on a user’s facial expressions. The proposed framework integrates deep learning techniques for emotion recognition with machine learning algorithms for personalized music selection. Facial emotions are identified using a lightweight MobileNetV2- based convolutional neural network (CNN), ensuring fast and ac- curate detection. The recommendation process employs content- based filtering to align users’ emotional states with suitable music tracks. To maintain user privacy, the system implements secure data handling techniques, including encryption and authentication. Overall, the platform delivers responsive, mood-driven music recommendations with high accuracy and minimal latency, making it suitable for entertainment, emotional therapy, and adaptive music applications.

This paper proposes a novel framework for a real-time music visualization system designed for the hearing-impaired, utilizing AI and serverless computing. The system converts audio signals into visual representations that capture both the physical and emotional aspects of music. A neural network-based Music Emotion Recognition (MER) model extracts emotional cues, which are integrated into the visualizations. The serverless computing ensures accessibility, while an account management system and comment collection system enable customization and regular retraining of the model for better accuracy. Results demonstrate the framework's effectiveness, highlighting the scalability and cost-efficiency of serverless computing. This work significantly advances music accessibility for the hearing-impaired, enhancing sensory experiences and promoting mental well- being. The MER model shows superior performance, with a 46.8% lower Root Mean Squared Error (RMSE) compared to other works targeting the same 10-second audio fragment length and a 13.5% higher Pearson's correlation coefficient (PCC) for 30-second fragments.

No abstract available

Current methods for Music Emotion Recognition (MER) face challenges in effectively extracting features sensitive to emotions, especially those rich in temporal detail. Moreover, the narrow scope of music-related modalities impedes data integration from multiple sources, while including multiple modalities often leads to redundant information, which can degrade performance. To address these issues, we propose a lightweight framework for music emotion recognition that improves the extraction of features that are both sensitive to emotions and rich in temporal information and that integrates data from both audio and MIDI modalities while minimizing redundancy. Our approach involves developing two innovative unimodal encoders to learn embeddings from audio and MIDI-like features. Additionally, we introduce a Bi-modal Fusion Attention Model (BFAM) that integrates features from low-level to high-level semantic information across different modalities. Experimental evaluations on the EMOPIA and VGMIDI datasets show that our unimodal networks achieve accuracies that are 6.1% and 4.4% higher than baseline algorithms for MIDI and audio on the EMOPIA dataset, respectively. Furthermore, our BFAM achieves a 15.2% improvement in accuracy over the baseline, reaching 82.2%, which underscores its effectiveness for bi-modal MER applications.

Classical machine learning techniques have dominated Music Emotion Recognition. However, improvements have slowed down due to the complex and time-consuming task of handcrafting new emotionally relevant audio features. Deep learning methods have recently gained popularity in the field because of their ability to automatically learn relevant features from spectral representations of songs, eliminating such necessity. Nonetheless, there are limitations, such as the need for large amounts of quality labeled data, a common problem in MER research. To understand the effectiveness of these techniques, a comparison study using various classical machine learning and deep learning methods was conducted. The results showed that using an ensemble of a Dense Neural Network and a Convolutional Neural Network architecture resulted in a state-of-the-art 80.20% F1 score, an improvement of around 5% considering the best baseline results, concluding that future research should take advantage of both paradigms, that is, combining handcrafted features with feature learning.

Music is a universal language that stimulates a wide range of emotions in people. This emotional effect results from music's ability to activate the brain waves related to emotion and memory. This subject is crucial because it has the potential to significantly improve in many fields and advanced AI applications. Early methods used hand-crafted features in typical machine learning models, but more recent developments have made use of deep learning methods. These techniques, however, frequently failed to reflect the complex and long-range dependencies present in music. These challenges led to the use of transformer models, has proven more effective at capturing the complexity of musical emotions. Hence, we propose EmotionWaveNet, a modified Wav2Vec 2.0 model for MER in Turkish music. In contrast to earlier methods, our model utilizes self-supervised learning to derive significant representations from raw audio, while integrating maqam-aware feature engineering to improve the identification of emotion-specific patterns. We train and evaluate EmotionWaveNet using the Turkish Music Emotion Dataset, showcasing its exceptional ability to classify musical emotions. The model achieves an accuracy of 89.75%, greatly exceeding traditional classifiers such as SVM (77.46%) and ANN (77.37%), showcasing the effectiveness of transformer-based architectures for capturing the emotional essence of music. Beyond improving MER accuracy, this research contributes to bridging the cultural gap in computational music analysis by adapting state-of-the-art deep learning models to a low-resource, Eastern musical domain. To address the challenge of a limited dataset, we employ transfer learning, initially pretraining the model on a large Western music dataset before fine-tuning it on Turkish music. This approach allows the model to learn fundamental musical representations from a rich data source while adapting to the distinct characteristics of maqam-based Turkish music. The findings emphasize the potential of transformers for cross-cultural music emotion recognition and provide a foundation for future studies exploring diverse musical traditions.

The design of meaningful audio features is a key need to advance the state-of-the-art in music emotion recognition (MER). This article presents a survey on the existing emotionally-relevant computational audio features, supported by the music psychology literature on the relations between eight musical dimensions (melody, harmony, rhythm, dynamics, tone color, expressivity, texture and form) and specific emotions. Based on this review, current gaps and needs are identified and strategies for future research on feature engineering for MER are proposed, namely ideas for computational audio features that capture elements of musical form, texture and expressivity that should be further researched. Previous MER surveys offered broad reviews, covering topics such as emotion paradigms, approaches for the collection of ground-truth data, types of MER problems and overviewing different MER systems. On the contrary, our approach is to offer a deep and specific review on one key MER problem: the design of emotionally-relevant audio features.

With the rapid advancement of artificial intelligence technologies, music emotion analysis has gradually become an important research topic in the fields of music information retrieval and intelligent music applications. However, existing studies have mainly focused on the automatic recognition of music emotions, often neglecting the aesthetic understanding and emotional experience involved in the process of music appreciation. As a result, they struggle to fully capture the complexity and richness of musical emotional expression. To address this issue, this paper proposes an AI music–empowered deep learning method for music emotion appreciation classification, aiming to enhance the representation and classification of musical emotion features by introducing AI-based music generation and augmentation mechanisms.First, grounded in theories of music emotion appreciation, an emotion classification framework oriented toward musical aesthetic understanding is established. Next, deep learning techniques are employed to perform multi-level feature modeling on music audio signals, while AI music empowerment strategies are used to enrich training samples and strengthen emotion feature representations. On this basis, a music emotion appreciation classification model is designed to effectively distinguish musical works with different emotional characteristics. Experimental results demonstrate that the proposed method outperforms traditional music emotion classification approaches in terms of both classification accuracy and stability, and is better able to reflect the emotional connotations and aesthetic characteristics of musical works. The findings of this study provide a new technical perspective for the integration of music emotion computing and intelligent music appreciation, and also offer valuable references for practical applications such as music education and digital music services.

No abstract available

Music emotion recognition (MER) is a critical task in the field of music information retrieval. However, most MER research relies solely on single-scale music spectrograms and fails to consider the complementary effects of spectrograms at different scales. Meanwhile, fully extracting emotion-related information from spectrograms remains a major challenge in MER. In this paper, we propose a hybrid attention model based on multi-resolution spectrograms, named MSMHA. The MSMHA model takes multi-scale Mel-spectrograms as inputs, and each input is fed into a well-designed hybrid attention network. The designed attention network successively includes a low-level feature extraction module, a local feature extraction module based on window attention, a channel attention-based long skip connection module, a high-level feature extraction module, and a branch classifier. After being processed by the hybrid attention network, each branch can fully extract emotion-related semantic features from a spectrogram of the specific resolution and output an emotion-classification probability. Finally, a decision-level weighted fusion strategy is applied to the multi-branch outputs to generate the final classification results. The experimental results on the PMEmo dataset demonstrate that our model is both promising and effective, achieving classification accuracies of 90.9%, 86.36%, and 79.87% on the binary-arousal, binary-valence, and four-quadrant dimensions, respectively. Ablation studies further confirm the effectiveness of both the multi-resolution spectrogram inputs and each module of the hybrid attention network.

No abstract available

We listen to music every day and feel various emotions from different music, it brings us to the right mood and bonds us to other people and creates a shared experience. Depending on the situation and mental state different people may get affected diversely. However, we can extract common music emotional information even from people with diverse backgrounds and cultures. In this research paper, we develop a system that uses music to retrieve human emotions using the k-Nearest Neighbors (kNN) machine learning algorithm and compare the classifier with others to find the most accurate. How do we define music as happy or sad? Low-level features such as beat, pitch, rhythm, valence and tempo which made up the music, are used in the classification of music emotion. For example, high pitched music or music with a high frequency is usually associated with a happy and energizing state, and the opposite for slow music, which makes us feel sad and calm. The proposed system can be very useful in music-related systems and applications such as music retrieval, recommendation and understanding. Also, various disciplines such as musicology, physiology, psychology and cognitive sciences refer to music emotion recognition.

The paper under investigation explores the problem of emotion recognition in AI-generated music by applying the concepts of cognitive psychology, affective neuroscience, and computational creativity. It presents a valence-arousal modeling, deep neural architecture and listener-based feedback loop approach to the emotional interpretation, generation, and critical analysis of emotional signals in music through the use of AI. The system uses hybrid CNN -RNN and Transformer models to derive the temporal-spectral features and project them onto the affective dimensions including valence, arousal, and tension. Managerial-level evaluation A multi-stage test that integrates perceptual ratings, physiological sensing and behavioral analysis proves the emotional fidelity of the system, with the system able to yield high classification accuracy of 92.4% and a r correlation coefficient of r = 0.89 between predicted and perceived emotional levels in humans. Findings are valid that although AI is efficient in reproducing the syntax of emotion, its perception is representational and not experiential. Ethical issues related to authenticity, authorship and privacy are considered with a focus on transparent and culture-inclusive model design. Altogether, the study contributes to computational affective musicology which places AI as a creative partner that recreates and enhances emotional expression, but not eliminates it.

No abstract available

We present a platform and a dataset to help research on Music Emotion Recognition (MER). We developed the Music Enthusiasts platform aiming to improve the gathering and analysis of the so-called “ground truth” needed as input to MER systems. Firstly, our platform involves engaging participants using citizen science strategies and generate music emotion annotations – the platform presents didactic information and musical recommendations as incentivization, and collects data regarding demographics, mood, and language from each participant. Participants annotated each music excerpt with single free-text emotion words (in native language), distinct forced-choice emotion categories, preference, and familiarity. Additionally, participants stated the reasons for each annotation – including those distinctive of emotion perception and emotion induction. Secondly, our dataset was created for personalized MER and contains information from 181 participants, 4721 annotations, and 1161 music excerpts. To showcase the use of the dataset, we present a methodology for personalization of MER models based on active learning. The experiments show evidence that using the judgment of the crowd as prior knowledge for active learning allows for more effective personalization of MER systems for this particular dataset. Our dataset is publicly available and we invite researchers to use it for testing MER systems.

No abstract available

No abstract available

No abstract available

: This paper discusses the fusion and application of music emotion recognition technology and artificial intelligence (AI) in music education. With the rapid development of artificial intelligence technology, its application is increasingly widely used in the field of education, especially in music education. AI not only improves teaching efficiency, but also provides students with a more personalized and efficient learning experience. As an important branch of AI, music emotion recognition technology can accurately identify and interpret the emotion and artistic conception expressed by analyzing the melody, rhythm and harmony elements in music works, which is of great significance for students to deeply understand the connotation and essence of music works in the process of music appreciation and learning. This paper analyzes the current situation, advantages and challenges of the integration of music emotion recognition and AI in music education, and puts forward corresponding strategies and suggestions, aiming to provide theoretical reference and practical guidance for innovative practice in the field of music education.

Xinjiang ethnic music, as an integral part of Chinese culture, embodies rich emotional expression and cultural connotation. With the rapid development of artificial intelligence, how to effectively use intelligent methods to preserve and transmit ethnic music has become a research hotspot. This paper proposes a deep learning-based emotion recognition and generative modeling approach for Xinjiang ethnic music. A multimodal feature fusion network is constructed to accurately identify musical emotions, and a generative adversarial network is employed to model intelligent music generation. Experimental results show that the proposed method achieves 85.7% accuracy in emotion recognition. The generated music maintains traditional characteristics while expressing emotions effectively. A survey of 500 respondents shows that 55.6% hold a positive view toward AI-based emotion recognition, and 65.4% express interest in AI-generated music, validating the feasibility and application potential of intelligent music inheritance methods.

No abstract available

Existing studies indicate that music therapy has a positive effect on insomnia. However, the effectiveness of traditional music therapy is often limited by subjective evaluation and practical barriers. Therefore, it is necessary to provide a more reproducible, accessible home-based conceptual method to relieve insomnia. The well-established machine learning model will become the core of it. Instead of using therapists' subjective evaluation, objective data such as electroencephalogram (EEG), Electrocardiogram (ECG) and Photoplethysmography (PPG) are suitable to provide more consistent judgements. Additionally, the recent more reliable wearable EEG devices lay a favorable foundation for this concept. Methods of quantification of arousal (both sleep-related and emotional arousal) have been introduced, so that the time-series signals, after noises being filtered or avoided, are input to Long Short-Term Memory (LSTM) network continuously. After the initialization is finished with rule-based program, it will collect users' emotional feedback and predict users' reaction to music based on previous record of their feedback to the recommended music. It aims at improving sleep quality consequently. Reinforcement Learning (RL) will also be integrated to enable the algorithm adapt to users' preference repetitively after investigating the feedbacks.

No abstract available

The migrants are especially susceptible to mental health disturbances because of the stressors associated with displacement like the language obstacles, social isolation, discrimination, and access to culturally sensitive care. These treatment gaps can be bridged using digital therapeutics, but most of the currently existing digital resources lack the alignment to cultural identities of migrants, have inconsistent connectivity, and they have not met the requirements of multilingualism. The manuscript presents the Sonic Sanctuary, a personalized adaptive AI-based music therapy platform that can provide culturally sensitive and customized listening experiences and facilitated self-regulation therapies. The system is based on (i) Multilingual onboarding and screening, (ii) music preference and cultural context modeling, (iii) affective inference through lightweight selfreport, optional physiological response, and closed-loop recommendation engine, which learns over time to mitigate distress and promote positive affect by attentively to music features. The Sonic Sanctuary design philosophy focuses on a privacy-by-design design, subtle personalization plans, and a powerful system of evaluation specific to community health contexts. Through the implementation of the short, lowbandwidth sessions, the platform will supplement the current services and empower stepped-care paths of migrant populations. Existing literature indicates that musicinterventional approaches are promising in enhancing coping and wellness in refugees and displaced individuals, and digital mental-health interventions are potentially helpful, however, further, stronger, and more effective clinical studies are still required to support these claims.

Mental health disorders and suicidality are rising among adolescents and young adults (A-YA) while rates of treatment engagement continue to remain low. Though the effectiveness of music therapy to enhance mental health is yet unknown for A-YA, the power of music to evoke emotions, facilitate communication, and promote relaxation suggests its promise. In this research study we discuss a multimodal music recommendation system which is capable of taking inputs in the form of a BFI-10 questionnaire, a series of self-view questions in the form of a survey, and an image of the user’s face in order to determine music preferences based on existing correlations between personality traits and taste in music, as an effort to make advancements in the field of music therapy.

Abstract: Mental health has emerged as a prime area of concern over the last few years, particularly among the youth and corporate professionals. The current research formulates a framework that applies machine learning to label stress levels and then suggests music according to the age group of the user. Based on the analysis of physiological and survey-based inputs, the model classifies stress as Low, Moderate, and High and maps results against age for recommending personalized music. Methods like SMOTE-Tomek resampling, PCA, and Random Forest Classifier were applied for accuracy improvement. Results indicate high accuracy in stress classification and validate age-based music therapy as a complementary method for stress alleviation

No abstract available

This study proposes a data-driven personalized family music therapy recommendation framework to address the limitations of conventional recommendation systems in modeling emotional interactions among multiple family members. The framework employs a reinforcement learning approach tailored to family environments, where emotional states evolve through continuous interpersonal interactions and feedback. A relationship-aware family state representation is developed by integrating behavioral observations and physiological responses, enabling the modeling of both individual emotional states and emotional contagion among family members. To support adaptive intervention, the action space incorporates music selection, spatial delivery strategies, and intentional silence decisions, allowing the system to balance active intervention with natural emotional self-regulation. In addition, a multi-channel reward mechanism is designed by combining physiological recovery, behavioral relaxation, and environmental stability indicators, while a temporal credit assignment strategy addresses delayed therapeutic effects. Experimental evaluation on a multimodal dataset covering four representative family stress scenarios demonstrates that the proposed framework achieves 89.7% accuracy in family emotion representation, 86.3% consistency with expert intervention judgments, and an 82.3% positive emotion conversion rate after continuous intervention. These findings suggest that reinforcement learning can provide an effective foundation for transforming music recommendation systems from individual entertainment tools into proactive family emotional support systems.

Due to numerous variables, including the covid-19 pandemic, economic unpredictability and climate change, nowadays a significant amount of stress and depression is experienced by individuals. Chronic stress can put your health at higher risk of health concerns. This serious health problem can be addressed by analysing electroencephalogram (EEG) signals and detecting the stress levels. This paper proposes Stress Detection and Music Therapy Application in which Power Spectral Density of each frequency band is extracted from brain wave signals and theta to beta ratio, which highly correlates with stress level, is evaluated. This approach makes use of K-Means Clustering followed by Support Vector Machine (SVM) to classify the stress into 3 levels i.e. stressed, intermediate and not stressed. The relevant music is recommended depending stress level of the individual. The proposed methodology aims to facilitate relaxation for individuals dealing with panic and high-stress scenarios by leveraging the therapeutic benefits of music.

As a particularly prominent application of recommender systems on automated personalized service, the music recommendation has been widely used in various music network platforms, music education and music therapy. Importantly, the individual music preference for a certain moment is closely related to personal experience of the music and music literacy, as well as temporal scenario without any interruption. Therefore, this paper proposes a novel policy for music recommendation NRRS (Nonintrusive-Sensing and Reinforcement-Learning based Recommender Systems) by integrating prior research streams. Specifically, we develop a novel recommendation framework for sensing, learning and adaptation to user's current preference based on wireless sensing and reinforcement learning in real time during a listening session. The established music recommendation prototype monitors individual vital signals for listening music, and captures song characters, individual dynamic preferences, and that it can yield better listening experience for users.

In recent years, individuals have been experiencing increasing levels of emotional stress, and emotion-related disorders have gradually become a major factor affecting mental health. Music, due to its natural capacity for emotional regulation, has demonstrated unique advantages in assisting psychological intervention. However, most existing music therapy systems remain at the stage of static recommendation and lack the ability to perceive users' real-time emotional states and respond dynamically.To address this limitation, we design and implement a Micro-Expression-Based Closed-Loop Music Therapy System. This system consists of two key modules: (1) a micro-expression emotion recognition module that utilizes deep neural networks to perform real-time emotion classification based on users' facial micro-expressions during music listening; and (2) a music generation module, which incorporates an alignment network (Align-Net) to align the micro-expression image embeddings with the CLAP-encoded personalized music descriptions, and generates corresponding music using a Latent Diffusion Model (LDM).To evaluate the feasibility and effectiveness of the proposed system, we conducted user studies assessing both emotion recognition accuracy and the quality of generated music. Experimental results demonstrate that the system enables emotion-driven dynamic music generation and provides a practical foundation for building more intelligent and responsive music therapy technologies.

Abstract In recent years, music therapy has gradually become an adjunct to conventional medicine and drug therapy. In this paper, genetic algorithms are used in conventional music therapy, and they are also used to optimize the combination of repertoire in music therapy. The repertoire mining model based on an improved genetic algorithm is constructed, and the optimized search is carried out with the genetic algorithm to mine the most effective therapeutic repertoire for patients. Based on the recommendation degree analysis, the personalized recommended repertoire is customized according to the specific conditions of patients. The therapeutic effect of this paper’s music therapy based on the optimization of repertoire by genetic algorithm is examined through comparative experiments from the changes in the therapeutic effects of the experimental and control groups before and after the experiments. Before the experiment, the experimental and control groups are not identical. After the experiment, the therapeutic effect of the experimental group was greatly improved, while the control group remained basically unchanged, the total score of the therapeutic effect of the two groups differed by 50.09 points, and showed significant differences in all dimensions of the therapeutic effect. The experimental group received an improvement of 4-8 points in each dimension of the treatment effect after the experiment, while the control group received an improvement of no more than 0.5 points. Both groups and inter- and intra-group differences were statistically significant.

Facial Expression Recognition (FER) has created widespread interest due to its potential uses in personalized technology and mental health, notably in systems that recommend music based on emotion. These systems can improve adaptable user interfaces and support music therapy. While prior research has explored algorithms for FER and their effectiveness in identifying emotions, existing solutions often lack optimal accuracy in pairing emotion recognition with music recommendations, particularly in real-world contexts with diverse user preferences. This study introduces a robust Convolutional Neural Network (CNN)-based model specifically designed for accurate FER, complemented by an innovative music recommendation system. The model was trained and evaluated using the publicly available FER-2013 dataset, which comprises 35,887 labelled facial images, along with a curated music dataset of 1,000 songs annotated with emotions. This approach incorporates a distinctive emotion-based song classification pipeline that utilizes a pre-built text emotion classification model to categorize song lyrics into seven emotional categories: anger, fear, disgust, neutral, surprise, sadness, and happiness. Experimental results demonstrate that our model achieved a FER accuracy of 91.78% and various methodologies discussed for emotion-based music categorization, significantly outperforming baseline methods. These findings offer a scalable framework for personalized and emotionally resonant music experiences, representing a viable solution for applications in tailored entertainment and music therapy.

Abstract Music, as a carrier of emotional sustenance, can not only achieve spiritual resonance in the process of listening but also reflect the vitality of life in the rhythm of music. As an effective relief method, music intervention has far-reaching significance in the development of psychotherapy. In the practice of music therapy, music selection is particularly important, and the rationality and rigor of music selection directly affect the therapeutic effect. The process of music selection in music therapy is in common with Internet music recommendation, so it has theoretical and practical value to apply personalized music recommendation algorithms to music therapy. In this paper, driven by big data, a music recommendation model based on an improved collaborative filtering (CF) algorithm is proposed, which combines the psychological adjustment of users’ music preferences and different music rhythm features to select music for music therapy and provides theoretical support for music therapy selection. The results show that the construction of music resources for music therapy based on an improved CF algorithm can greatly improve the music selection process of music therapy.

This project presents an Emotion-Based Music Recommendation System designed to enhance mental wellness through intelligent and personalized music therapy. The system uses real-time facial ex- pression recognition via a webcam and Convolutional Neural Networks (CNNs) to identify users’ emotional states such as happy, sad, angry, neutral, or surprised. Based on the detected emotion, suitable music is recommended to positively influence mood, reduce stress, and support emotional balance. By integrating computer vision, deep learning, and an emotion- aware recommendation engine, the system offers a non-invasive, user-friendly, and cost-effective digital solution for mental health support and affective computing applications

To systematically evaluate the intervention effect of music therapy on anxiety and depression in breast cancer patients. Randomised controlled trial (RCT) on music therapy for anxiety and depression in breast cancer patients was searched from 7 major databases, PubMed, Embase, the Cochrane Library, WOS, CNIC, Wanfang, and Wipro, spanning the period of library construction to 23 October 2023, and the literature screening of music therapy for anxiety or depression in breast cancer patients was carried out by 2 experimentalists, each of whom conducted a literature screening RCT independently of the other anxiety or depression in a RCT. Methodological quality was evaluated using the PEDro scale; GRADE profiler software for quality of evidence; and RevMan 5.4 was used for effect size merging and forest plots; publication bias tests and sensitivity analyses were performed using Stata 17.0; and standardized mean difference (SMD) and 95% CI were used as the effect statistics. A total of 13 RCTs with 1326 subjects (aged 18–70 years) were included in the literature, with a mean PEDro score of 6.8, and the literature was overall of good methodological quality. Meta-analysis showed that music therapy improved anxiety in breast cancer patients (841 cases), with a combined effect size (SMD = − 0.82, 95% CI [− 1.03, − 0.61] and P < 0.001); and improved depression in breast cancer patients (387 cases) with a combined effect size (SMD = − 0.76, 95% CI [− 1.15, − 0.38], P < 0.001). Subgroup analyses showed that music intervention with off-site music (757 cases) and music choice of non-self-selected music (537 cases) had the best effect on anxiety improvement, with corresponding combined effect sizes (SMD = − 0.88, P < 0.001; SMD = − 0.83, P < 0.001), respectively; followed by an intervention length of < 30 min (589 cases), a frequency of 2 times/day (382 cases), and intervention period of 2–3 weeks (101 cases) had the best effect on anxiety improvement, and the corresponding combined effect sizes were (SMD = − 0.80, P < 0.001; SMD = − 0.91, P < 0.001; SMD = − 1.02, P < 0.001), respectively; and the music selection was the choice of one's own favourite music among the expert recommendations (219 cases) (270 cases) had the best effect on the improvement of depressed mood, with combined effect sizes of (SMD = − 1.15, P < 0.001; SMD = − 0.71, P < 0.001) and music with an intervention duration of 30 min (287 cases), an intervention frequency of 1 time/day (348 cases), and an intervention period of 2–4 weeks (120 cases), respectively, with corresponding combined effect sizes of (SMD = − 0.75, P < 0.001; SMD = − 0.86, P < 0.001; SMD = − 1.06, P < 0.001), respectively. Music therapy can improve anxiety and depression in breast cancer patients, and the level of evidence is moderate. Although the heterogeneity between studies is high, which may lead to bias in the results, we explored the source of heterogeneity through subgroup and sensitivity analyses, providing a good evidence-based basis for clinical practice. The heterogeneity of anxiety and depression was explored by subgroup analysis, with anxiety due to music duration and music cycle; and depression due to intervention cycles and music duration. Sensitivity analyses also identified music duration and music cycle as contributing to the heterogeneity. Also, this study has some limitations since the included literature did not take into account the duration of the disease, education, and family economic status and did not categorize the age stages. This study found that music therapy improves anxiety and depression in breast cancer patients and the results can be used as a basis for clinical practice and researcher enquiry. This research has been registered on the INPLASY platform (https://inplasy.com/contact/) under the number: INPLASY2023100057.

Objective: To evaluate the effects of music therapy on perioperative anxiety, depression and postoperative satisfaction in elderly patients undergoing surgery for distal radius fractures. Methods: A retrospective cohort study of 186 patients (≥65 years) was conducted from January 2022 to December 2024. The participants were divided based on the treatment period: standard care (January 2022–June 2023, n = 92) or music therapy plus standard care (July 2023–December 2024, n = 94). Outcomes measured included anxiety and depression [Hospital Anxiety and Depression Scale (HADS)]; anxiety levels (Visual Analog Scale for Anxiety [VAS-A]; sleep quality (Pittsburgh Sleep Quality Index [PSQ]; satisfaction; pain; length of hospital stay and functional recovery at preoperative, postoperative 24 hours, discharge and 2-week and 6-week follow-up timepoints. Results: The music-therapy group demonstrated progressive improvements across all timepoints. At 6 weeks, the music therapy group maintained superior outcomes compared with standard care: HADS-Anxiety (6.05 ± 1.98 vs. 7.34 ± 2.24, P < 0.001), HADS-Depression (5.86 ± 2.07 vs. 6.69 ± 2.53, P = 0.015), VAS-A (28.83 ± 12.94 vs. 33.47 ± 15.19 mm, P = 0.026) and Pittsburgh Sleep Quality Index (6.15 ± 1.94 vs. 7.33 ± 2.18, P < 0.001). Patient satisfaction scores were significantly higher (8.46 ± 0.94 vs. 7.93 ± 1.29, P < 0.001). Hospital stay was reduced (3.21 ± 0.89 vs. 4.12 ± 1.23 days, P < 0.001), opioid consumption decreased (23.59 ± 11.67 vs. 34.56 ± 13.89 oral morphine equivalents, P < 0.001) and functional recovery was superior (disability of the arm, shoulder and hand]: 28.01 ± 11.34 vs. 38.87 ± 13.78, P < 0.001). Conclusion: Music therapy considerably reduces anxiety and depression, enhances sleep quality and satisfaction and facilitates recovery of elderly patients undergoing distal radius fracture surgery.

Introduction Patients with cancer frequently suffer from psychological distress, including anxiety and depression, which negatively impacts their quality of life. Music therapy (MT)and Tai Chi exercise are emerging as promising complementary interventions in cancer care. However, their comparative effectiveness remains uncertain. This systematic review aims to evaluate and compare the effects of MT and Tai Chi exercise on anxiety, depression and quality of life among patients with cancer . Methods and analysis Following the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) guidelines, a comprehensive search will be conducted in databases including PubMed, Web of Science, Scopus and CNKI for randomised controlled trials (RCTs) published up to March 2025. Studies assessing the effects of either intervention on cancer patients will be included. The primary outcomes are anxiety and depression, while quality of life is considered a secondary outcome. All outcomes will be measured using validated instruments. Two reviewers will independently screen, extract data and assess study quality using the Cochrane Risk of Bias tool. Meta-analyses will be performed using RevMan where appropriate, along with analyses of heterogeneity, publication bias and sensitivity. Ethics and dissemination Ethical approval is not required as the study involves secondary data analysis. Results will be disseminated through peer-reviewed journals and conferences. This review will offer evidence-based insights to guide cancer rehabilitation strategies involving mind-body and music-based therapies. PROSPERO registration number CRD42024550787.

Mental health conditions and disorders, such as anxiety and depression, are common, but factors like cost, access to professionals and stigma limit access to traditional care. Music therapy is a potential alternative approach that does not involve the use of medications. This study was a quantitative pre-test/post-test design to determine the effectiveness of a music therapy intervention for 100 adults (ages 18-60) to reduce anxiety and depression. The intervention consisted of 2–3 sessions per week that lasted for 30–45 minutes, such as listening, singing, and instrument playing, and song writing. GAD-7 and PHQ-9 were used to assess anxiety and depression levels, respectively. The analysis of paired t-tests, effect size, correlation, and perception surveys were conducted. Results indicated that both outcomes were significantly decreased. Mean GAD-7 scores decreased from 13.42 to 8.67 (mean reduction 4.75 points), t(99)=12.34, p<0.001, Cohen's d=1.23. Mean PHQ-9 scores decreased from 14.85 to 9.46 (mean reduction 5.39 points), t(99)=13.87, p<0.001, Cohen's d=1.39. 78% (anxiety) and 81% (depression) participants made clinically meaningful improvements. There was a moderate correlation between attendance and improvement (r=0.52, anxiety; r=0.48, depression; both < 0.001). The overall response rate was very favourable with a score of 86–94% agreement of the participants in therapeutic benefits and healthcare integration. The study found that structured Music therapy is highly effective and has large effect size in reduction of anxiety and depression levels, besides being highly acceptable. Music therapy is a simple, easy-to-access and low cost non-pharmacological treatment for adult mental health.

: This study aims to explore the effectiveness of various music therapy methods in alleviating academic anxiety and depression among college students by conducting a comprehensive review and analysis of relevant literature. The results indicate that music therapy, as a psychological intervention, has shown significant effects in reducing academic anxiety and depression among college students. The music therapy not only effectively lowers anxiety levels and alleviates depressive symptoms but also enhances students' self-confidence, promoting their healthy and happy growth. Additionally, the study highlights future research directions and challenges. Although music therapy has achieved significant results in alleviating academic anxiety and depression among college students, further investigation is needed into its long-term effects and mechanisms, as well as its combination with other psychological intervention methods. Moreover, the personalized design and optimization of music therapy for college students with different cultural backgrounds, personality types, and levels of academic pressure deserve in-depth research.

BACKGROUND Subjective cognitive impairment has been reported to be associated with depressive symptoms, stress and anxiety in older people. This study examines the impact of music interventions on cognitive functioning, depression, anxiety, and stress for adults with dementia or cognitive impairment. METHOD We searched Academic Search Complete, CINAHL, Medline, and PsycINFO without restriction to date. Systematic reviews with or without meta-analysis of music interventions were included. Effect sizes were estimated using standardized mean difference (SMD), weighted mean difference (WMD), mean difference (MD), and Hedges g, as reported. Effect sizes were reported as 〈 0.1 = small effect to 〉 0.5 as large effect. RESULTS Twenty systematic reviews were included. There is evidence that music interventions can have effects on cognitive abilities compared to standard care, with a small to large decrease in anxiety. CONCLUSION Music interventions might have variable effects on improved cognitive functioning, depression, anxiety and stress.

No abstract available

This study looked at the effect of group music therapy on dementia patients' levels of feelings of depression and anxiety. Quasi-experiment was carried out (N = 121) in which patients were randomly divided into a control (N = 61) and an experimental group (N = 60). Following a six-weeks, twelve-sessions trial involving both passive and active music therapy, it was discovered that the music group therapy intervention decreased the anxiety and depression levels of dementia patients in the experimental group as compared to the control group, which did not receive any music therapy. It was also noted that after three months of follow-up, participants in the experimental group still maintained a low level of anxiety and depression, but the control group's level remained high, therefore, substantiating the effect of music in lowering anxiety and depression among older ones. We urge the nursing descipline, medical professionals, carers, and care homes to include music therapy in the care they provide for patients with dementia.

Objective: To observe the effects of music therapy on negative psychology, sleep, and quality of life in elderly patients recovering from cerebral infarction accompanied by depression and anxiety. Methods: We performed a retrospective analysis of 160 patients with convalescent cerebral infarction diagnosed between December 2022 and December 2023 in the Affiliated Brain Hospital of Nanjing Medical University. Cases from the control group (n = 76) were treated with conventional rehabilitation training. Meanwhile, the music therapy group (n = 84) cases were managed with standard and music therapy. Various measures, including polysomnography, the general quality of life questionnaire-74 (GQOL-74), the self-rating anxiety scale (SAS), and the self-rating depression scale (SDS) were collected at baseline (admission) and 4 weeks post-admission. Results: After 4 weeks, the music therapy group exhibited higher GQOL-74 scores, lower SAS scores, and lower SDS scores (P < 0.001). Furthermore, compared with the control group, the music therapy group had shorter sleep latency, longer sleep duration, higher sleep efficiency, lower wake time, fewer wake times, shorter S1 stage, longer S3 + S4 stage, and longer REM period (P < 0.001). Conclusions: Music therapy appears to be effective in improving negative psychology, sleep quality, and quality of life in elderly patients with comorbid depression and anxiety during recovery from cerebral infarction.

This paper investigates the effect of music therapy on symptoms, social performance, and emotional regulation in individuals with ASD, depression, and anxiety. Through organised but flexible exercises such as rhythmic breathing, songwriting and chanting, music therapy has become a supplement to conventional treatments that treat the social, emotional and physiological dimensions of these conditions. For people with ASD, music therapy helps them to gain nonverbal communication and social skills. When it comes to depression, music therapy provides a safe space for the expression of emotion and enhances neurochemistry that stabilises mood. For anxiety management, music therapy reduces physiological symptoms and offers ways to cope in the long run. They used a mixed method, combining quantitative (SRS, HDRS, STAI scores) and qualitative feedback to assess the efficacy of the therapy. These outcomes improved dramatically in all groups, especially in the music therapy groups, demonstrating its efficacy as a treatment for mental health and neurodevelopmental disorders. Such results point to the potential for music therapy to increase social interaction, emotional well-being and wellbeing.

No abstract available

BACKGROUND Music therapy has become a notable complementary approach for psychotherapy. Nevertheless, because of the interplay between music and culture, the relative efficacy of five-element music therapy and Western music therapy remains to be determined for Chinese patients with moderate depression. OBJECTIVE To assess the relative efficacy of Western music and five-element music in alleviating negative emotions and sleep disorders of Chinese patients who had moderate depression. METHODS Two hundred and one patients with moderate depression were recruited from August 2022 to February 2023 and randomly assigned to three groups: five-element music therapy (n = 67), Western music therapy (n = 67), and treatment as usual (n = 67). The five-element and Western music therapy groups received the assigned intervention measures once a day for two weeks. The primary outcome was the Hamilton Depression Scale score. The secondary outcomes were scores for the Hamilton Anxiety Scale and the Sleep Dysfunction Rating Scale. Data for the three groups of patients were collected on the day before intervention, the seventh day of intervention, and the fourteenth day of intervention. RESULTS Of the 201 patients enrolled, 183 completed all treatments and follow-ups. The full analysis set showed that the five-element music and Western music therapies effectively mitigated depression (B=-3.075, 95 % CI: -5.188, -0.961, P = 0.004; B=-1.985, 95 %CI: -3.908, -0.062, P = 0.043) and anxiety (B=-2.433, 95 % CI: -3.992, -0.874, P = 0.002; B=-2.672, 95 % CI: -4.221, -1.122, P = 0.001). But patients in the five-element music therapy group showed greater relief of symptoms of depression and anxiety compared with patients in the Western music therapy group; the difference was statistically significant (mean difference=-1.94±0.766, 95 %CI: -3.77, -0.11, P = 0.034; mean difference=-1.37±0.509; 95 %CI: -2.59, -0.15, P = 0.021). In addition, the five-element music group exhibited significant improvements in sleep (B=-3.119, 95 %CI: -5.016, -1.223, P = 0.001). CONCLUSION Preliminary research results suggest that five-element music may be used as an adjunct therapy with conventional treatment to more effectively alleviate depression and anxiety and improve sleep quality of Chinese patients who have moderate depression. REGISTRATION The trial was registered at the Chinese Clinical Trial Registry (ChiCTR2100050682), and the initial recruitment began in August 2022.

Objective: This study aims to investigate the effects and feasibility of integrating music therapy into a family support system (FSS) for adolescent depression. Methods: In this retrospective study, 120 adolescents with depression who visited the Wenzhou Seventh People’s Hospital between June 2022 and June 2024 were selected and categorized into the music therapy group and a conventional group (60 cases in each group). The conventional group received FSS-assisted treatment and conventional medication, whereas the music therapy group was given music therapy based on the conventional treatment. The Hamilton Depression Rating Scale (HAMD), Hamilton Anxiety Rating Scale (HAMA), Family Assessment Device (FAD), Adolescent Self-Rating Life Events Checklist (ASLEC), and Pittsburgh Sleep Quality Index (PSQI) were compared between the two groups. Results: After treatment, the decrease in HAMD and HAMA scores was significantly greater in the music therapy group than in the conventional group (P < 0.05). The FAD score in the music therapy group improved more significantly in dimensions such as problem-solving, communication, emotional involvement, behavior control, and overall function than that of the conventional group (P < 0.05). The ASLEC score showed a greater reduction in aspects such as interpersonal relationships, academic stress, punishment, and loss (P < 0.05). The PSQI score improved more significantly with regard to subjective sleep quality, sleep latency, sleep duration, sleep disorders, hypnotic drug use, and daytime dysfunction (P < 0.05). Conclusion: Integrating music therapy into FSS can effectively reduce depression and anxiety symptoms in adolescents with depression, improve the family and psychosocial function, and enhance sleep quality. However, further studies are still needed to verify and improve this approach.

ABSTRACT The elderly are at a higher risk of developing depression, anxiety, and other mental illnesses due to their age, chronic diseases, financial dependencies and other factors. Access to mental health services is often limited due to high cost, a low number of psychologists/therapists, and social taboos. In this paper, we propose automating the emerging technique of ‘music therapy’ (MT) using the internet of things (IoT) to develop a novel solution ‘IoT‐MT’. IoT‐MT has the potential to deal with psychological disorders in a convenient, quick, and cost‐effective manner. The system is based on collecting body parameters (pulse rate, SPO2, and blood pressure) from the patient demonstrating anxiety levels and playing appropriate music on a handheld device if the received parameters are above a certain predefined threshold. An early version of the system prototype is developed along with the mobile app, and experiments were conducted at a general hospital in Karachi, Pakistan, with 500 patients and their corresponding cardiologists. It has been observed that the proposed MT system reduced the anxiety levels for the majority of the patients, as indicated by their pulse rate and blood pressure. Also, the anxiety detection of IoT‐MT was found to be accurate when compared to the patients’ self‐assessment performed using the standard psychology tool, state‐trait anxiety inventory (STAI).

In recent years, tailored therapies for mental health have gained popularity. The paper provides a novel framework for real-time adaptive individualized music treatment using Internet of Things (IoT) and Artificial Neural Networks (ANNs). The framework uses biometric sensors, data analytics, and music intervention methodologies to personalize treatments to physiological reactions to reduce stress and anxiety and improve mental health. The architecture starts with IoT-enabled biometric sensors that capture real-time heart rate, skin conductivity, and brainwave patterns. The main data source is these sensors, tracking physiological reactions in different situations. Data is sent to a central processing unit or cloud platform for analysis. In analysis, advanced artificial neural network models understand complex physiological-mental state relationships. The ANN model is trained using biometric data and mental health-boosting music therapy. Iterative training and optimization help the ANN model predict the optimal music therapy treatments using real-time physiological responses. Participants should assess planned music therapy sessions to improve the ANN model. A feedback and model improvement loop adapts the framework to new mental health needs.

The MAP is an innovative receptive music therapy intervention derived from psychomusical relaxation methods that aims to foster the well-being and recovery of youths with mental health problems by providing them with an adaptive and effective music-assisted means to regulate their mood states. In this quasi-experimental pilot study, we assessed the mood-enhancing potential of participation in MAP sessions delivered by a music therapist in an in-patient mental health facility for children and adolescents. Using short standardized self-reported questionnaires, 20 participants aged 9-17 years old (M = 14, SD = 2.4), mainly girls (13 = 65%), rated their affective state immediately before and after two to four MAP sessions and a similar number of regular unit activity sessions used as comparison. This created a 2 × 2 (Time × Condition) single-group within-individual design. We analyzed pre-post session changes in affect using multilevel mixed models and found participation in MAP sessions to be associated with systematic reductions in self-reported general negative affect and state anxiety. These variations were of modest-to-large magnitude and significantly greater than those associated to participation in regular unit activities. While only a first step towards the validation of the MAP as an effective intervention to foster more adaptive and effective day-to-day mood regulation in youths with mental health problems, this study supports its specific potential to alleviate negative affects and provides a rare demonstration of the putative benefits of music therapy in a pediatric mental health inpatient context.

As older adults get older, their bodies deteriorate. A large proportion of the elderly population experiences decreased attention and memory failure. Through the intervention of music, the memory of the elderly can be improved to some extent, which is of great help to their physical and mental health. In this paper, a visual music therapy rehabilitation system was designed in which the cloud processing module performs physiological signal normalization, feature extraction of heart rate variability patterns, neural network‐based emotional state classification, and data storage for adaptive visual music intervention. When the system detected a change in old people's emotions, the subsystem judged the current model of people's expressions and transmitted the results to the server subsystem, which normalized the data and afterward sent it to the visualization subsystem. The visualization subsystem loaded the corresponding music file based on the server's data and generated a dynamic video to relieve the old people's mental health condition. In this paper, a systematic review was conducted of the three aspects of physical and mental debilitation and social debilitation in the elderly population. Experimental results showed that after 10 weeks of treatment, 34.7% of the elderly were in the pre‐debilitation state and 22% were in a robust state in the visual music therapy rehabilitation system group. This observation period represents a defined intervention stage used to examine the short‐ to medium‐term emotional response of elderly participants during continuous visual music therapy interaction. In this paper, the effects of performing music therapy on the debilitated elderly were observed, which contributes to the development of intelligent medical information systems for emotion‐aware rehabilitation in elderly populations.

Mental health has emerged as a critical global concern in the twenty-first century. Despite growing awareness, access to professional psychological services remains severely limited in developing nations, particularly India, where cultural stigma, financial barriers, and a shortage of trained professionals prevent millions from seeking timely support. This paper presents Manam, an AI-powered empathetic chat application designed to provide accessible, anonymous, and culturally relevant mental health support across India, with a particular focus on the Tamil Nadu region. The system leverages the Google Gemini API for context-aware, compassionate natural language generation. Key features include a dynamic mood-adaptive user interface, an integrated music therapy module, and an automated crisis intervention system that connects users in distress to verified regional emergency helplines. The modular client-server architecture built on Python Flask ensures scalability and maintainability. Manam addresses critical limitations of existing global platforms by offering a free, localized, and emotionally adaptive digital companion that serves as a first-line mental health support system. Evaluation of the prototype demonstrates consistent empathetic response generation, reliable crisis detection performance, and positive user engagement, affirming the system's potential to reduce the mental health treatment gap across India.

The World Health Organization (WHO) reported that in 2019, nearly 970 million people worldwide suffered from mental health conditions, with anxiety and depression being the most frequently occurring. These disorders cause considerable distress, a decline in productivity, and elevated medical expenses. Although music therapy has been proven to have results in decreasing stress and stabilizing emotions, its application is limited by expense, inaccessibility, and reliance on professional practitioners. To overcome these shortcomings, this paper proposes a Generative AI approach to personalized and adaptive music therapy that incorporates cross-modal learning and large language models (LLMs). The system takes speech and text inputs, predicts Valence-Arousal-Dominance (VAD) scores with a custom Transformer, projects them onto emotion categories, and synthesizes structured prompts for music with MusicGen. By removing the necessity for human intervention, the model enables real-time therapeutic music delivery. Experimental assessment examines the Transformer's ability in VAD-based emotion prediction, and the outcomes demonstrate the potential for automated emotion support with music. The best-performing LoRA configuration achieves a Pearson Correlation Coefficient of 0.7622 for valence, 0.7811 for arousal, and 0.6724 for dominance prediction.

Children often struggle to express their emotions verbally, making early detection of mental health issues challenging and limiting timely intervention. This research proposes an AI-powered mobile application designed to support children’s mental well-being through multimodal emotional analysis and personalized interventions. The significance of this study lies in addressing the lack of accessible, child-friendly, and culturally relevant digital mental health solutions, particularly for Sinhala-speaking users, while reducing dependence on subjective expert interpretation. The proposed system integrates four main components, including a culturally grounded therapeutic story generator using a GRU-based deep learning model, an emotion analysis module for children’s drawings utilizing DenseNet121 and YOLOv8 with colour and spatial feature extraction, an emotion-aware music recommendation and adaptive puzzle game powered by EfficientNet-based facial emotion recognition, and a Sinhala voice-first mood prediction system using fine-tuned XLM-RoBERTa .Experimental results demonstrate strong performance, including improved contextual coherence and adaptive therapeutic storytelling capabilities, 79% accuracy in drawing-based emotion classification, 87% accuracy in emotion-aware music recommendation, and 96% accuracy in Sinhala text-based mood prediction, along with improved user engagement through personalized storytelling and adaptive feedback mechanisms. The system further enhances interpretability by generating parent-friendly emotional reports using large language models. In conclusion, this research contributes a comprehensive, multimodal, and culturally adaptive AI solution for supporting children’s mental health, enabling early emotional detection, improved communication, and scalable intervention through an engaging mobile platform.

Abstract Mental-health-oriented interventions increasingly require structured, scalable, and nonpharmacological formats that can strengthen psychological functioning without reducing clinician oversight. This study examined whether a therapist-led, AI-assisted classical piano intervention could produce stronger process-level psychological gains than matched therapist-led piano instruction without adaptive AI feedback among young adults diagnosed with mild-to-moderate anxiety or depression. A randomized controlled mixed-methods design was used with 120 Mandarin-speaking participants aged 18–30 years assigned to either an AI-assisted condition or a non-AI control condition. Over a 10-week program, both groups completed weekly 90-min supervised sessions, and pretest/post-test changes were assessed in resilience, mindfulness, and music-related emotional processing. Quantitative data were analyzed using 2 × 2 mixed repeated-measures ANOVA and follow-up tests, while post-intervention interviews from a stratified experimental subsample were analyzed using interpretative phenomenological analysis. The findings showed significant improvement in both groups; however, the AI-assisted condition demonstrated larger gains across all three primary outcomes, together with a steeper trajectory of weekly musical performance development. Engagement indicators within the AI-assisted group were positively associated with psychological improvement. Qualitative findings reinforced this pattern, with emotional release, increased self-awareness, and stronger motivation to continue practice emerging as the most frequent themes. The study suggests that AI can contribute meaningfully to music-based intervention in supervised contexts, not as an autonomous agent but as an adaptive feedback layer. From a practical viewpoint, the model provides a viable framework for structured mental health-focused practice in various contexts.

Academic stress is a common issue experienced by university students and is closely associated with mental health conditions and a decline in academic performance. Emotion regulation plays a crucial role in stress management and has been proven to be an adaptive ability that enhances well-being. Interventions involving arts and music have increasingly been explored as effective approaches to improving emotion regulation skills among university students. This paper aims to systematically review existing studies related to Arts and music-based interventions in enhancing emotion regulation and reducing academic stress among university students, as well as to identify the intervention methods implemented. The method used in this paper is a systematic review of experimental studies, meta-analyses, and qualitative research focusing on arts and music interventions among university students. The analysis of results is based on measurements of academic stress, emotion regulation, and psychological well-being. The main findings indicate that Arts and music-based interventions consistently reduce academic stress levels and help improve students’ emotion regulation skills. In conclusion, Arts and music-based interventions are effective in enhancing emotion regulation and reducing academic stress among university students. Such approaches can be integrated into mental health support programs, including counseling and group guidance services, for students at the university level.

Abstract EP1.4, e-Poster Terminal 1, September 4, 2025, 11:35 - 13:00 Aims Some migrant women face numerous vulnerabilities, such as poor levels of integration and lack of social support, leading to an increased risk of developing postpartum depression. If left untreated, it may have serious consequences, both for the mother and the baby. Pharmacological treatment can be effective but is often limited by low uptake and adherence among mothers. Psychotherapy shows mixed results and is also hindered by low uptake or delayed treatment. “Music & Motherhood” aims to bridge the gap in non-pharmacological treatment options through a music-based intervention. Initially implemented in the UK with promising results, the intervention is now being adapted to the Portuguese context to explore its acceptability, feasibility and transferability. Methods In a 10-week intervention, mothers from diverse socioeconomic and cultural backgrounds, including those of migrant origin, and their babies will engage in guided singing activities led by a singing leader. In these sessions, mothers will learn songs that are not language-specific to ensure inclusiveness for different cultural groups. To evaluate the implementation and impact of the intervention, we will use a mixed-methods approach, guided by implementation science frameworks, through a longitudinal pre- and post-assessment study. Results We expect this intervention to reduce depressive symptoms, strengthen mother-baby bonding, and foster a sense of community and mutual support among participating mothers. Moreover, evaluating its implementation and outcomes will offer valuable insights to refine and expand future music-based approaches in maternal mental health. Conclusions This study may highlight the potential of music-based interventions as an accessible, culturally inclusive approach to supporting maternal mental health, especially for migrant women. By fostering mother-baby bonding and creating a supportive community, this initiative could address some of the challenges faced by these groups experiencing vulnerabilities.

Mental health and emotional well-being are crucial to academic success and workplace productivity. The rising stress levels amongst students due to prolonged exposure to stressors in schools and workplaces calls for early detection and intervention strategies, owing to the demands of modern education and professional settings. Classic evaluation methods, like self-reported surveys, are frequently centered on speculation and omit on-the-spot adaptability. This work introduces a system that employs machine learning models for the evaluation of mental health disorders and their control.For students, it uses the Depression, Anxiety and Stress Scale (DASS-21) to compute real-time psychological, behavioral and physiological indicators of user stress levels. When detected, automated alerts and intervention strategies (psychiatrist consultations, parental notifications, adaptive learning. In corporate environments, a real-time emotion detection system uses computer vision, speech analysis, and biometric sensors to assess employees' emotional states, allowing for tailored workplace interventions such as personalized workload management and mindfulness initiatives. Moreover, a music recommendation system creates personalized playlists tailored to their needs, curbing workplace stress and improving productivity.

Women of Global Majority ethnicities have an increased risk of developing and sustaining perinatal mental health problems in the UK. This is partially explained by the ethnic inequalities experienced at an individual, societal and systemic level. Previous research highlights the benefits of engaging with participatory music-based interventions to alleviate symptoms of postnatal depression, stress and anxiety, but current provision lacks cultural inclusivity. This qualitative study focuses on fourteen women of African, Caribbean and Mixed heritages living in South East London to explore how their perinatal experiences, coping strategies and preferences regarding music-based support could inform a future culturally inclusive perinatal mental health participatory music intervention. Participants took part in either an online focus group or interview, led by a local community leader. An inductive reflexive thematic analysis identified four overarching themes: (1) supportive mechanisms during the perinatal period, (2) the overwhelming pressures and expectations of motherhood, (3) systemic barriers to accessing perinatal mental healthcare and (4) suggestions for future perinatal mental health music-based support groups. This study reveals the individual experiences of the perinatal period for women of African, Caribbean and Mixed heritages, exploring themes of sociocultural pressures, barriers to care and individual activities used to support mental health. The sociocultural, logistical and musical considerations outlined in this study highlight gaps in current community provision and offer practical suggestions for facilitating inclusive music-based interventions for perinatal mental health in South East London.

Background Just-In-Time Adaptive Interventions (JITAIs) are interventions designed to deliver timely tailored support by adjusting to changes in users' internal states and external contexts. To accomplish this, JITAIs often apply complex analytic techniques, such as machine learning or Bayesian algorithms to real- or near-time data acquired from smartphones and other sensors. Given the idiosyncratic, dynamic, and context dependent nature of mental health symptoms, JITAIs hold promise for mental health. However, the development of JITAIs is still in the early stages and is complex due to the multifactorial nature of JITAIs. Considering this complexity, Nahum-Shani et al. developed a conceptual framework for developing and testing JITAIs for health-related problems. This review evaluates the current state of JITAIs in the field of mental health including their alignment with Nahum-Shani et al.'s framework. Methods Nine databases were systematically searched in August 2023. Protocol or empirical studies self-identifying their intervention as a “JITAI” targeting mental health were included in the qualitative synthesis if they were published in peer-reviewed journals and written in English. Results Of the 1,419 records initially screened, 9 papers reporting on 5 JITAIs were included (sample size range: 5 to an expected 264). Two JITAIs were for bulimia nervosa, one for depression, one for insomnia, and one for maternal prenatal stress. Although most core components of Nahum-Shani's et al.'s framework were incorporated in the JITAIs, essential elements (e.g., adaptivity and receptivity) within the core components were missing and the core components were only partly substantiated by empirical evidence (e.g., interventions were supported, but the decision rules and points were not). Complex analytical techniques such as data from passive monitoring of individuals' states and contexts were hardly used. Regarding the current state of studies, initial findings on usability, feasibility, and effectiveness appear positive. Conclusions JITAIs for mental health are still in their early stages of development, with opportunities for improvement in both development and testing. For future development, it is recommended that developers utilize complex analytical techniques that can handle real-or near-time data such as machine learning, passive monitoring, and conduct further research into empirical-based decision rules and points for optimization in terms of enhanced effectiveness and user-engagement.

Restoration of upper limb motor function and patient functional independence are crucial treatment targets in neurological rehabilitation. Growing evidence indicates that music-based intervention is a promising therapeutic approach for the restoration of upper extremity functional abilities in neurologic conditions such as cerebral palsy, stroke, and Parkinson’s Disease. In this context, music technology may be particularly useful to increase the availability and accessibility of music-based therapy and assist therapists in the implementation and assessment of targeted therapeutic goals. In the present study, we conducted a pre-clinical, single-arm trial to evaluate a novel music-based therapeutic device (SONATA) for upper limb extremity movement training. The device consists of a graphical user interface generated by a single-board computer displayed on a 32” touchscreen with built-in speakers controlled wirelessly by a computer tablet. The system includes two operational modes that allow users to play musical melodies on a virtual keyboard or draw figures/shapes whereby every action input results in controllable sensory feedback. Four motor tasks involving hand/finger movement were performed with 21 healthy individuals (13 males, aged 26.4 ± 3.5 years) to evaluate the device’s operational modes and main features. The results of the functional tests suggest that the device is a reliable system to present pre-defined sequences of audiovisual stimuli and shapes and to record response and movement data. This preliminary study also suggests that the device is feasible and adequate for use with healthy individuals. These findings open new avenues for future clinical research to further investigate the feasibility and usability of the SONATA as a tool for upper extremity motor function training in neurological rehabilitation. Directions for future clinical research are discussed.

This article synthesizes contemporary research on the multifaceted impacts of piano interaction on human cognition, mental health, and neural processes, and explores its translation into intelligent human-computer systems. Evidence from clinical psychology demonstrates that structured piano training can significantly alleviate anxiety and depression in the elderly, enhance executive functions and working memory in aging populations, and serve as an effective component within multi-element interventions for severe mental illness, as exemplified by the GET UP PIANO trial. Neuroscientific investigations reveal that these benefits are supported by a specialized auditory-motor network, encompassing premotor and parietal cortices, which exhibits significant plasticity in experts and is engaged during both music perception and mental imagery. Unique patterns of musical processing in special populations, such as preserved fronto-temporal connectivity for song in autism spectrum disorder, provide a neurobiological rationale for music-based therapies. Critically, these findings are now informing the development of intelligent technologies. We explore how principles of cross-modal correspondence and personalization are being leveraged to create adaptive systems for health-tech, including closed-loop neurofeedback for neuromodulation, smart keyboards for motor rehabilitation using biofeedback, and AI-driven analysis of musical improvisation for mental health assessment. Finally, the review addresses key challenges, including the need for methodological standardization, mechanistic elucidation, and ethical HCI design focused on data security and long-term engagement. The convergence of neuroscience, clinical practice, and technology positions the piano as a powerful and evolving interface for enhancing human health and cognitive resilience.

12003 Background: Anxiety is prevalent, disruptive, and under-treated among cancer survivors. While the pandemic exacerbated the mental health crisis, it also accelerated telehealth adoption, creating opportunities to expand access. Cognitive behavioral therapy (CBT) is the gold standard for anxiety, but not all survivors respond to it, have access, or prefer this option due to stigma. Music therapy (MT) has demonstrated short-term anxiety reduction, but its long-term effectiveness relative to other treatments is unknown. This study evaluated whether MT is noninferior to first-line CBT when both are delivered remotely. Methods: In this randomized clinical trial, English- or Spanish-speaking survivors of any cancer type or stage with anxiety lasting ≥1 month were randomized 1:1 to MT or CBT. Participants received 7 weekly sessions of a standardized protocol via Zoom and were followed for 26 weeks. Co-primary endpoints were Hospital Anxiety and Depression Scale (HADS) anxiety subscale changes at weeks 8 and 26. Secondary outcomes included depression, fatigue, insomnia, pain, cognitive function, and quality of life. Data were analyzed with linear mixed-effects models following intention-to-treat. Assuming 15% attrition and 1-sided p < 0.025, the trial had 80% power to detect noninferiority within a margin of D = 0.35*standard deviations (SDs). The margin was informed by an expected SD of 4.2 and the minimum clinically important difference (MCID) of 1.7 points for the HADS anxiety subscale. Thus, establishing noninferiority would indicate differences between MT and CBT are not clinically meaningful. Results: Among 300 (147 MT; 153 CBT), mean age was 56.9 (SD 13.2) years, 224 (74.7%) were female, 228 (76.5%) were white, and 57 (19.1%) were Hispanic. The most common cancer types were breast (45.3%) and hematologic (15.7%). At week 8, mean change in HADS anxiety score was -3.12 (95% CI -3.59 to -2.65) in MT and -2.97 (95% CI -3.45 to -2.50) in CBT; between-group difference was -0.15 (95% CI -0.78 to 0.49), which was within margin of 1.20 (P < 0.001 for noninferiority of MT; calculated from SD of 3.42). At week 26, mean change was -3.31 (95% CI -3.78 to -2.85) in MT and -3.00 (95% CI -3.47, -2.53) in CBT; between-group difference was -0.31 (95% CI -0.95 to 0.32), which was within margin of 1.30 (P < 0.001 for noninferiority of MT; calculated from SD of 3.65). Both groups produced anxiety reductions exceeding MCID of 1.7 and showed similar improvements in secondary outcomes. Conclusions: MT was non-inferior to CBT for short- and long-term anxiety reduction among diverse survivors of various cancer types. Both treatments produced clinically meaningful, durable anxiety reduction and were delivered remotely using standardized protocols, thereby increasing their reach and scalability. MT should be considered alongside first-line CBT to expand treatment options for anxiety during cancer survivorship. Clinical trial information: NCT05215353 .

BACKGROUND Preoperative anxiety is a prevalent concern among patients undergoing surgery, potentially elevating the risk of postoperative complications. This study aimed to compare the effectiveness of inhalation aromatherapy using rosemary essential oil and music therapy in reducing anxiety levels among general surgery patients. METHODS This study employed a randomized controlled clinical trial design involving 236 patients scheduled for general surgery. The participants were randomly assigned to one of four groups: aromatherapy with Rosmarinus officinalis L. essential oil (n = 59), music therapy (n = 59), combined aromatherapy and music therapy (n = 59), and a control group (n = 59). Data collection instruments consisted of the Spielberger State Anxiety Inventory, a personal information form, a music player, and wireless headphones. In the aromatherapy group, patients were instructed to inhale three drops of 25% rosemary essential oil. The music therapy group listened to instrumental music through headphones, while the combined group received both inhalation aromatherapy with rosemary essential oil and music therapy. The control group received three drops of an aromatic placebo. The interventions were administered for a duration of 30 min, starting one hour before surgery. Inferential statistics, including the Kolmogorov-Smirnov test, chi-square test, paired t-test, one-way analysis of variance, and LSD post-hoc test, were employed for data analysis. RESULTS The study findings demonstrated significant effectiveness in reducing anxiety among surgical patients for aromatherapy (P = 0.001), music therapy (P < 0.001), and combined aromatherapy and music therapy (P = 0.001). Aromatherapy showed the highest anxiety-reducing effect, whereas combined therapy exhibited the lowest effect. CONCLUSION The results indicate that inhalation aromatherapy, music therapy, and combined aromatherapy and music therapy are all effective approaches for anxiety reduction in surgical patients. Hence, the implementation of these cost-effective and non-invasive interventions is recommended to alleviate anxiety among surgical patients.

Pain is a prevalent issue across various medical conditions, and numerous methods have been employed to manage it. In developed countries, cancer ranks as the second leading cause of mortality after cardiovascular diseases. This study aims to compare the effectiveness of hypnotherapy, relaxation therapy, and Music Therapy (MT) with control groups in alleviating pain in children with cancer. The research involved a single-blind clinical trial with a study population comprising children diagnosed with an abdominal mass and acute lymphoblastic leukemia. Fifty-eight patients were selected through blocked randomization. Data were collected through a demographic checklist and the Visual Analog Scale (VAS) to measure pain intensity. Data analysis was carried out using repeated measures ANOVA in SPSS version 22 to compare mean pain intensity among the study groups. A total of 58 patients, with an average age of 9.28±4.02 years, participated in the study. Significant differences were observed between the hypnotherapy group and the other study groups (F=14.51; P≤0.001), as well as between the MT group and the other study groups (F=12.81; P≤0.001). Moreover, significant differences were found in terms of time between the relaxation therapy group and the other study groups (F=8.46; P≤0.001) and between the control group and the other groups (F=5.506; P≤0.001). Based on the findings, relaxation therapy, hypnotherapy, and Music Therapy (MT) have shown significant effectiveness in alleviating pain in cancer patients during their treatment. Trial registration Clinical trial. gov Identifier: IRCTID website (code: IRCT20120905010744N2; http://irct.ir).

Background/Objectives: Depression is a prevalent and debilitating illness that significantly affects psychological and physical well-being. Apart from conventional therapies such as psychotherapy and medication, individuals with depression often lack opportunities for activities that are generally perceived as enjoyable, such as music, meditation, and arts, which have demonstrated therapeutic effectiveness. TaKeTiNa music therapy has been employed as a therapeutic intervention for more than two decades. However, there is a notable absence of well-designed clinical trials investigating its antidepressant effects, a gap we aim to address in our current study. Furthermore, shifts in the progression of depression may manifest both psychologically, by influencing emotional states, and physiologically, by leading to alterations in lipid and sphingolipid metabolism, cortisol levels, and immune system function. Our study seeks to analyze the impact of TaKeTiNa music therapy on both levels. Methods: This is a prospective monocentric randomized waitlist-controlled clinical trial. It investigates the influence of TaKeTiNa music therapy on patients with major depression in an outpatient setting. Therefore, interested persons are randomly assigned to two groups, an intervention group or a control group, after completing a screening procedure. The intervention group starts with an eight-week TaKeTiNa music therapy intervention. The waiting group receives the same therapy program after completing the follow-up period. Blood and saliva sampling as well as responses to questionnaires are obtained at specific time points. Discussion: Our study investigates the effects of TaKeTiNa music therapy, a non-pharmacological antidepressant treatment option, on depressive symptoms. We also address functional and causal immunological changes; hormonal changes, such as changes in cortisol levels; and metabolic changes, such as changes in serum lipids and sphingolipids, during the course of depression. We expect that this study will provide evidence to expand the range of treatment options available for depression.

The healthcare profession is highly challenging and distinctive, requiring students to possess exceptional academic performance and clinical competencies in addition to interpersonal skills. Accumulating stress developed by these factors increases anxiety among students affecting their performance in numerous aspects. Music therapy and guided imagery have proven to be effective in reducing stress and anxiety. However, no studies have been done comparing these two interventions. The aims and objectives of this study were to evaluate the effect of music therapy versus guided imagery on healthcare professional students suffering from stress and anxiety. A randomized clinical trial study was conducted at the healthcare institution. Ethical clearance was granted by the institutional research and ethics committee before the study commenced. Fifty individuals with stress and anxiety were recruited in the study, and informed consent was taken. They were randomly allotted into two Groups A and B. “A” received music therapy from Peter Sterling’s album “The Angels Gift” and “B” received guided imagery audio by Johns Hopkins. Both groups received intervention for 15 min for 4 days a week for 3 weeks. Pre- and post-assessment was taken of both the groups by measuring their stress and anxiety scoring based on the Perceived Stress Scale and Hamilton Anxiety Rating Scale. Version 29.0.10 Statistical Package for the Social Studies (SPSS) program (SPSS Inc.; Chicago, IL, USA) was used for analysis. Independent sample t-test and Paired t-test were used for analysis. P < 0.05 was selected. Music therapy and guided imagery were both equally effective in the reduction of stress and anxiety levels with P > 0.05.in healthcare professional students. Music therapy and guided imagery showed significant effects in reducing stress and anxiety among healthcare professional students. Outcomes showed that both were equally reliable.

Highlights • NMT in CAREN VR could innovate neurorehabilitation for PD.• Motor NMT helps cognitive rehab in PD by engaging executive function networks.• Attention, inhibition, and processing speed improved with combined training.

Introduction: Fear and anxiety are the most prevalent concerns among young children, often leading to uncooperative behaviour that poses challenges in providing effective dental care. Hence, implementing appropriate strategies can significantly improve a child’s dental experience. Music therapy as audio distraction technique is one such promising non-pharmacological methods to reduce the anxiety in children. Different kinds of music influence human brain waves, leading to deep relaxation and calmness, and thereby reducing pain and anxiety. Aim: To evaluate and compare the Tell-Show-Do (TSD), Music of Choice and traditional music therapy as audio distraction techniques on anxiety levels in children undergoing restorative treatment. Materials and Methods: The present randomised three arm controlled clinical trial was carried out in the Department of Paediatric and Preventive Dentistry, KAHER’S KLE VK. Institute of Dental Sciences, Belagavi, Karnataka, India from May 2024 to November 2024. A total of 66 children aged 6-10 years were included in the study and allocated into three groups: Group I TSD, Group II Music of Choice and Group III Traditional Music Therapy. All Children by single operator received class I restorative procedure in primary/permanent molars. Anxiety was evaluated with pulse rate and Venham’s Picture Test at pre and post interventions. The child’s behaviour was also determined by the Frankl’s behaviour rating scale before and after the restorative procedure. The pre and post intervention scores of all the assessed parameters were recorded. Descriptive statistics were used to summarise the demographic data. For inferential analysis, the Kruskal-Wallis test was used to compare anxietyscore changes between the three groups (inter-comparison), the Wilcoxon matched-pairs test for intra-group comparison, and the Mann-Whitney U test was used for pair-wise comparison among the groups. The level of significance was set at p<0.05. Results: 66 patients were recruited for the study in which females were 36 (54.55%) and males were 30 (45.46%) with the mean age of 8.20 years. The results revealed no significant differences in pre intervention scores among the three groups, confirming comparable baseline levels. For Venham’s Picture Test, post intervention, significant differences were observed between TSD and Music of Choice (p=0.003*) and between TSD and traditional music (p=0.001*). However, Music of Choice and traditional music exhibited no significant difference (p=0.1490). Pulse rate mean scores differed significantly between the groups TSD (97.00), Music of Choice (91.05), and traditional music (89.35) with p=0.0370*. Intra-group comparison of pre and post intervention scores illustrated that Group II (Music of Choice) showed statistically significant reduction in both Venham’s Picture Test (p=0.001*) and pulse rate (p=0.0016*). Group III (traditional music) showed statistically significant changes for both measures with the p-value of 0.001*. In contrast, Group I (TSD) there were no statistically significant changes. Overall, all three groups showed a significant reduction in anxiety. Among them, traditional music yielded the lowest anxiety scores followed by Music of Choice group and highest with the TSD group. Additionally, there was a significant improvement in patients’ behaviour measured by Frankl’s behaviour rating scale from negative to positive and definitely positive across the three groups. Conclusion: Traditional music therapy and Music of Choice showed statistical significant changes in anxiety reduction over TSD group. Traditional music therapy had a significant impact over Music of Choice. Both the audio distraction techniques are promising, safer, non-invasive and effective behaviour guidance techniques in children.

PURPOSE Anxiety is prevalent, disruptive, and undertreated among survivors of cancer. Cognitive behavioral therapy (CBT) is the first-line treatment, but not all individuals have access, respond to treatment, or prefer this option because of stigma. Music therapy is effective for short-term anxiety reduction, but it is unknown whether it is noninferior to first-line CBT for long-term anxiety reduction. METHODS This comparative effectiveness trial randomly assigned English- or Spanish-speaking survivors of cancer to seven weekly telehealth sessions of music therapy or CBT. The coprimary end points were changes in the Hospital Anxiety and Depression Scale (HADS) anxiety score at weeks 8 and 26. The noninferiority margin was 0.35 standard deviations, informed by a minimal clinically important difference (MCID) of 1.7 points. Secondary outcomes included fatigue, depression, insomnia, pain, cognitive dysfunction, and health-related quality of life. RESULTS Among N = 300 patients, 74.7% was female, 76.5% was White, and 19.0% was Hispanic. At week 8, the mean change in HADS anxiety score was –3.12 (95% CI, –3.59 to –2.65) in music therapy and –2.97 (95% CI, –3.45 to –2.50) in CBT; the between-group difference was –0.15 (95% CI, –0.78 to 0.49), within the noninferiority margin of 1.20 (P < .001). At week 26, the mean change was –3.31 (95% CI, –3.78 to –2.85) in music therapy and –3.00 (95% CI, –3.47 to –2.53) in CBT; the between-group difference was –0.31 (95% CI, –0.95 to 0.32), within the noninferiority margin of 1.28 (P < .001). Both groups produced anxiety reductions exceeding the MCID and showed similar improvements in secondary outcomes. CONCLUSION Music therapy is noninferior to CBT for anxiety in survivors of cancer. Both telehealth interventions produced clinically meaningful, durable improvements in anxiety.

Objective This study aimed to assess how listening to music after cardiac valve replacements affected patients' pain, anxiety, and vital signs. Method In Fuzhou, China's Fujian Medical University Union Hospital, the cardiac surgery division conducted a randomized controlled clinical experiment. 86 patients were enrolled, and 43 were assigned randomly to each group (control and experimental group). The standard treatment was given to the control group, while the experimental group was given standard treatment + a 15-min music intervention 3 times. Indicators include pain, anxiety and vital signs (respiratory rate, heart rate, and blood pressure). Results In comparison to the control group, the experimental group, over time, demonstrated a statistically substantial decrease in pain, anxiety, systolic blood pressure, heart rate and respiratory rate (all P  < 0.001), yet, there were no discernible variations ( P  > 0.05) in diastolic blood pressure. Conclusions In conclusion, these results provide additional proof for using music therapy to minimize cardiac postoperative pain and anxiety, as well as systolic blood pressure, heart rate and respiratory rate. Moreover, it should be regarded as a supplementary treatment for pain and anxiety after cardiac valve replacement and other medical procedures with comparable postoperative pain.

OBJECTIVE To assess whether music therapy is effective to reduce pain during daily personal hygiene care (DPHC), a procedure performed in all patients in a pediatric intensive care unit. METHODS Fifty critically ill children were enrolled in a crossover controlled clinical trial with random ordering of the intervention, i.e., passive music therapy, and standard conditions, and blind assessment of pain on film recordings. Primary outcome was variation of the Face Legs Activity Cry Consolability (FLACC) score (range 0 to 10) comparing before and during DPHC. Secondary outcomes were changes in heart rate, respiratory rate, and mean arterial blood pressure, and administration of analgesic or sedative drugs during DPHC. Mixed-effects linear model analysis was used to assess effect size (95% CI). RESULTS Median (Q25-Q75) age and weight of the patients were 3.5 (1.0, 7.6) years and 15.0 (10.0, 26.8) kg. Consecutive DPHC were assessed the 3rd (2, 5) and 4th days (3, 7) of hospitalization. In standard conditions, FLACC score was 0.0 (0.0, 3.0) at baseline and 3.0 (1.0, 5.5) during DPHC. With music therapy, these values were, respectively, 0.0 (0.0, 1.0) and 2.0 (0.5, 4.0). Rates of FLACC scores >4 during DPHC, which indicates severe pain, were 42% in standard conditions and 17% with music therapy (p=0.013). Mixed-effects model analysis found smaller increases in FLACC scores (-0.54 [-1.08 to -0.01]; p= 0.04) and heart rate (-9; [-14.53; -3.4]; p=0.001) with music therapy. CONCLUSION Music therapy is effective to improve analgesia in critically ill children exposed to DPHC.

BACKGROUND Children with autism spectrum disorder (ASD) experience deficits in social skills. Music therapy (MT) has been used as a therapeutic aid for clinical disorders. This study aims to explore the effect of MT on the social skills of children with ASD and to provide evidence for clinical intervention in the social skills of children with ASD. METHODS Children with ASD who were admitted to the Department of Children's Health Care of Zhongshan Torch Development Zone People's Hospital from April 2023 to March 2024 were continuously recruited and assigned to the experimental group and the control group by random number table. The control group received standard care only, while the experimental group added MT to standard care. The MT program is led by an occupational therapist and combines social skills training with musical activities. The training was conducted in small groups of 3-5 children for 30 minutes, three times a week for 12 weeks. The Social Responsiveness Scale (SRS-2), the Autism Treatment Evaluation Checklist (ATEC), and the Gesell Development Schedules (GDS) were performed before and after the intervention. RESULTS A total of 29 children with ASD were included and randomly assigned to the MT group (n = 15) and the control group (n = 14). All participants completed the whole treatment protocol. There was no significant difference in the scores of SRS-2, ATEC, and GDS between the two groups before intervention. After 12 weeks of intervention, the scores of SRS-2 of the MT group were decreased in the social communication subscale (P < 0.05 compared to baseline and the control group) and total scores (P < 0.05 compared to baseline and the control group). The score of the ATEC scale of the MT group decreased in the speech/language/communication subscale (P < 0.05 compared to baseline and the control group), the sociability subscale (P < 0.05 compared to baseline and the control group), and the total score (P < 0.05 compared to baseline). The development quotient score of the social domain of GDS in the MT group was significantly higher than that before intervention (P < 0.05) and that in the control group (P < 0.05). CONCLUSION This study suggests that MT could effectively improve the social skills of children with ASD, and has a positive effect on language ability. MT has the potential to be an effective complement to regular social skill training.

Background: Music-based intervention has been advocated as a nonpharmacologic approach for the perioperative control of pain and anxiety in surgical patients. However, its impact on patients with preoperative anxiety has not been clearly established. Our study aimed to examine the impact of music-based intervention administered before, during, and after surgery on postoperative opioid consumption and pain levels, as well as preoperative anxiety, depression, and pain catastrophizing. We hypothesized that, compared to a control group, music-based intervention would be effective in reducing opioid requirements and mood disorders. Methods: This study was a single-center, prospective, single-blinded, randomized controlled trial. Inclusion criteria isame-day or observation surgery. Exclusion criteria included American Society of Anesthesiologists physical status IV, use of spinal anesthesia, PROMIS Anxiety T-scores ≤ 57.4 and ≥74.1, preoperative chronic opioid use, transgender surgery, and history of drug or alcohol abuse. Music-based intervention was developed by a certified music therapist. Each patient was randomized to receive standard of care (SC) or SC plus music-based intervention before, during, and after surgery. The primary end point was postoperative oral morphine equivalents (OMEs) over 5 days following surgery using the area under the curve (AUC)Secondary end points were PROMIS Anxiety, PROMIS Depression scores Pain Catastrophizing Scale scores, postoperative nausea and vomiting, time of hospital discharge, and patient satisfaction (0 = totally unsatisfied to 10 = completely satisfied). Results: A total of 75 patients were randomized to a music-based intervention (n = 33) or control (n = 42) group. Patients in the music-based intervention group consumed 56.7% less opioids than those in the control group (AUC was 2.8 in the music-based intervention group vs. 6.4 in the control group, absolute standardized mean difference (aSMD) = 0.34 (−0.17, 0.85)). No difference in pain scores was recorded between groups. Music-based intervention also reduced anxiety on postoperative day (POD)2 (aSMD = 0.38 (−0.16, 0.91)), depression on POD2 (aSMD = 0.31 (−0.23, 0.84)) and POD4 (aSMD = 0.24 (−0.29, 0.77)), and pain catastrophizing on POD1 (aSMD = 0.24 (−0.3, 0.77)). Conclusions: Our data support the use of music-based intervention to reduce postoperative opioid requirements. Music-based intervention may also reduce anxiety, depression, and pain catastrophizing.

Patients undergoing shock wave lithotripsy (SWL) have a high level of anxiety. This study was aimed to compare the effect of music therapy and aromatherapy on patients' anxiety levels during SWL. In this clinical trial, 176 patients undergoing SWL were randomly assigned to four groups: music therapy, aromatherapy, combined music therapy and aromatherapy, and control. The state-trait anxiety inventory (STAI) was used to collect data. In the aromatherapy group, three drops of rosemary essential oil were used during SWL. For the music therapy group, soft music was played during SWL. In the music therapy and aromatherapy group, soft music and three drops of rosemary essential oil were used simultaneously. In the control group, three drops of aromatic distilled water were used. The results showed that all patients had moderate anxiety before the intervention. The analysis of variance test did not show a statistically significant difference between the groups, but this difference was significant after the intervention. The levels of anxiety after the intervention were higher in the music therapy, aromatherapy, and combined music therapy and aromatherapy groups than in the control group. The results of the paired t-test showed the level of anxiety after the intervention was significantly increased in the intervention group (P < 001). In conclusion, the results indicated that music therapy, aromatherapy, and combined music therapy and aromatherapy had no effect on reducing patients' anxiety during SWL.

As artificial intelligence becomes more and more ingrained in daily life, we present a novel system that uses deep learning for music recommendation and emotion-based detection. Through the use of facial recognition and the DeepFace framework, our method analyses human emotions in real-time and then plays music that reflects the mood it has discovered. The system uses a webcam to take pictures, analyses the most common facial expression, and then pulls a playlist from local storage that corresponds to the mood it has detected. An engaging and customised experience is ensured by allowing users to manually change the song selection via a dropdown menu or navigation buttons. By continuously looping over the playlist, the technology guarantees continuity. The objective of our system is to improve emotional well-being through music therapy by offering a responsive and automated music-selection experience.

Many practices have been presented in music generation recently. While stylistic music generation using deep learning techniques has became the main stream, these models still struggle to generate music with high musicality, different levels of music structure, and controllability. In addition, more application scenarios such as music therapy require imitating more specific musical styles from a few given music examples, rather than capturing the overall genre style of a large data corpus. To address requirements that challenge current deep learning methods, we propose a statistical machine learning model that is able to capture and imitate the structure, melody, chord, and bass style from a given example seed song. An evaluation using 10 pop songs shows that our new representations and methods are able to create high-quality stylistic music that is similar to a given input song. We also discuss potential uses of our approach in music evaluation and music therapy.

Existing digital mental wellness tools often overlook the nuanced emotional states underlying everyday challenges. For example, pre-sleep anxiety affects more than 1.5 billion people worldwide, yet current approaches remain largely static and"one-size-fits-all", failing to adapt to individual needs. In this work, we present EmoHeal, an end-to-end system that delivers personalized, three-stage supportive narratives. EmoHeal detects 27 fine-grained emotions from user text with a fine-tuned XLM-RoBERTa model, mapping them to musical parameters via a knowledge graph grounded in music therapy principles (GEMS, iso-principle). EmoHeal retrieves audiovisual content using the CLAMP3 model to guide users from their current state toward a calmer one ("match-guide-target"). A within-subjects study (N=40) demonstrated significant supportive effects, with participants reporting substantial mood improvement (M=4.12, p<0.001) and high perceived emotion recognition accuracy (M=4.05, p<0.001). A strong correlation between perceived accuracy and therapeutic outcome (r=0.72, p<0.001) validates our fine-grained approach. These findings establish the viability of theory-driven, emotion-aware digital wellness tools and provides a scalable AI blueprint for operationalizing music therapy principles.

Introduction: Music provides an incredible avenue for individuals to express their thoughts and emotions, while also serving as a delightful mode of entertainment for enthusiasts and music lovers. Objectives: This paper presents a comprehensive approach to enhancing the user experience through the integration of emotion recognition, music recommendation, and explainable AI using GRAD-CAM. Methods: The proposed methodology utilizes a ResNet50 model trained on the Facial Expression Recognition (FER) dataset, consisting of real images of individuals expressing various emotions. Results: The system achieves an accuracy of 82% in emotion classification. By leveraging GRAD-CAM, the model provides explanations for its predictions, allowing users to understand the reasoning behind the system's recommendations. The model is trained on both FER and real user datasets, which include labelled facial expressions, and real images of individuals expressing various emotions. The training process involves pre-processing the input images, extracting features through convolutional layers, reasoning with dense layers, and generating emotion predictions through the output layer Conclusion: The proposed methodology, leveraging the Resnet50 model with ROI-based analysis and explainable AI techniques, offers a robust and interpretable solution for facial emotion detection paper.

Home-based music therapy devices require accessible and cost-effective solutions for users to understand and track their therapeutic progress. Traditional physiological signal analysis, particularly EEG interpretation, relies heavily on domain experts, creating barriers to scalability and home adoption. Meanwhile, few experts are capable of interpreting physiological signal data while also making targeted music recommendations. While large language models (LLMs) have shown promise in various domains, their application to automated physiological report generation for music therapy represents an unexplored task. We present a prototype system that leverages LLMs to bridge this gap -- transforming raw EEG and cardiovascular data into human-readable therapeutic reports and personalized music recommendations. Unlike prior work focusing on real-time physiological adaptation during listening, our approach emphasizes post-session analysis and interpretable reporting, enabling non-expert users to comprehend their psychophysiological states and track therapeutic outcomes over time. By integrating signal processing modules with LLM-based reasoning agents, the system provides a practical and low-cost solution for short-term progress monitoring in home music therapy contexts. This work demonstrates the feasibility of applying LLMs to a novel task -- democratizing access to physiology-driven music therapy through automated, interpretable reporting.

This study explores the application of recurrent neural networks to recognize emotions conveyed in music, aiming to enhance music recommendation systems and support therapeutic interventions by tailoring music to fit listeners' emotional states. We utilize Russell's Emotion Quadrant to categorize music into four distinct emotional regions and develop models capable of accurately predicting these categories. Our approach involves extracting a comprehensive set of audio features using Librosa and applying various recurrent neural network architectures, including standard RNNs, Bidirectional RNNs, and Long Short-Term Memory (LSTM) networks. Initial experiments are conducted using a dataset of 900 audio clips, labeled according to the emotional quadrants. We compare the performance of our neural network models against a set of baseline classifiers and analyze their effectiveness in capturing the temporal dynamics inherent in musical expression. The results indicate that simpler RNN architectures may perform comparably or even superiorly to more complex models, particularly in smaller datasets. We've also applied the following experiments on larger datasets: one is augmented based on our original dataset, and the other is from other sources. This research not only enhances our understanding of the emotional impact of music but also demonstrates the potential of neural networks in creating more personalized and emotionally resonant music recommendation and therapy systems.

Driven by the escalating global burden of mental health conditions, music-based interventions have attracted significant attention as a non-invasive, cost-effective modality for emotion regulation and psychological stress relief. However, current digital music services rely on static preferences and fail to adapt to users'instantaneous psychological states. Furthermore, directly mapping electroencephalography (EEG) to music generation remains challenging due to severe paired-data scarcity and a lack of interpretability. To address these limitations, we propose MindMelody, a fully functional, closed-loop real-time system for EEG-driven personalized music intervention. MindMelody introduces an emotion-mediated semantic bridge. Specifically, a hybrid Transformer-GNN first decodes real-time EEG signals into global Valence-Arousal states and local temporal affect trajectories. These states are then fed into a Retrieval-Augmented Generation (RAG)-equipped Large Language Model (LLM) to formulate structured intervention plans. Subsequently, a novel Hierarchical EEG Controller injects global affect prefixes and local temporal guidance into a pretrained music backbone, enabling fine-grained controllable audio synthesis. Crucially, the system incorporates a continuous feedback loop that updates generation parameters on the fly based on the user's evolving EEG dynamics. Extensive experiments show that MindMelody improves control adherence and emotional alignment, and receives higher perceived helpfulness in a short-term listening setting, suggesting its promise as an adaptive affect-aware music generation framework.

[This retracts the article DOI: 10.1155/2022/9960589.].

Abstract Machine learning (ML) deals with algorithms able to learn from data, with the primary aim of finding optimum solutions to perform tasks autonomously. In recent years there has been development in integrating ML algorithms with live coding practices, raising questions about what to optimize or automate, the agency of the algorithms, and in which parts of the ML processes one might intervene midperformance. Live coding performance practices typically involve conversational interaction with algorithmic processes in real time. In analyzing systems integrating live coding and ML, we consider the musical and performative implications of the “moment of intervention” in the ML model and workflow, and the channels for real-time intervention. We propose a framework for analysis, through which we reflect on the domain-specific algorithms and practices being developed that combine these two practices.

Abstract Background Mobile devices for remote monitoring are inevitable tools to support treatment and patient care, especially in recurrent diseases such as major depressive disorder. The aim of this study was to learn if machine learning (ML) models based on longitudinal speech data are helpful in predicting momentary depression severity. Data analyses were based on a dataset including 30 inpatients during an acute depressive episode receiving sleep deprivation therapy in stationary care, an intervention inducing a rapid change in depressive symptoms in a relatively short period of time. Using an ambulatory assessment approach, we captured speech samples and assessed concomitant depression severity via self-report questionnaire over the course of 3 weeks (before, during, and after therapy). We extracted 89 speech features from the speech samples using the Extended Geneva Minimalistic Acoustic Parameter Set from the Open-Source Speech and Music Interpretation by Large-Space Extraction (audEERING) toolkit and the additional parameter speech rate. Objective We aimed to understand if a multiparameter ML approach would significantly improve the prediction compared to previous statistical analyses, and, in addition, which mechanism for splitting training and test data was most successful, especially focusing on the idea of personalized prediction. Methods To do so, we trained and evaluated a set of >500 ML pipelines including random forest, linear regression, support vector regression, and Extreme Gradient Boosting regression models and tested them on 5 different train-test split scenarios: a group 5-fold nested cross-validation at the subject level, a leave-one-subject-out approach, a chronological split, an odd-even split, and a random split. Results In the 5-fold cross-validation, the leave-one-subject-out, and the chronological split approaches, none of the models were statistically different from random chance. The other two approaches produced significant results for at least one of the models tested, with similar performance. In total, the superior model was an Extreme Gradient Boosting in the odd-even split approach (R²=0.339, mean absolute error=0.38; both P<.001), indicating that 33.9% of the variance in depression severity could be predicted by the speech features. Conclusions Overall, our analyses highlight that ML fails to predict depression scores of unseen patients, but prediction performance increased strongly compared to our previous analyses with multilevel models. We conclude that future personalized ML models might improve prediction performance even more, leading to better patient management and care.

Postpartum depression (PPD) significantly impacts women’s mental health and social functioning, yet effective therapies remain limited. This study investigates the preventive effects of music therapy on PPD-like behaviors and the underlying neurobiological mechanisms in a mouse model subjected to ovarian hormone withdrawal (HW). Mice exposed to daily music sessions exhibited markedly reduced depression- and anxiety-like behaviors, as evidenced by enhanced performance in behavioral tests such as the open field test (OFT), forced swim test (FST), elevated plus maze test (EPM), sucrose preference test (SPT), novelty-suppressed feeding (NSF) test, and tail suspension test (TST). Furthermore, music therapy normalized oxidative stress indicators (NO, MDA, SOD, CAT, GSH-Px, T-AOC, ATP, and glutamate) in the serum, hippocampus, and prefrontal cortex. Additionally, music exposure reduced levels of proinflammatory factors (IL-6, IL-1β, iNOS, TNF-α, and TGF-β) and the activation of microglia and astrocytes in these brain regions. Notably, music therapy preserved neuronal integrity, promoted neurogenesis, and maintained synaptic plasticity, evidenced by the restoration of dendritic spines. Transcriptome sequencing identified differential gene expression in pathways related to synaptic plasticity, inflammation, and oxidative stress. These findings suggest that music therapy prevents PPD by modulating oxidative stress, inflammation, and synaptic integrity, providing robust preclinical evidence for its potential as a natural preventive intervention for PPD. This study underscores the need for further clinical research to validate the therapeutic efficacy of music in preventing PPD in humans, highlighting its promise as a non-invasive and accessible treatment modality.

INTRODUCTION Burn patients often experience pain and mental health difficulties in the Intensive Care Unit (ICU). This study investigated the effects of live Music-Assisted Relaxation (MAR) on background pain, mental health, vital signs, and medication administration in adult ICU burn patients. METHOD This was a randomized controlled trial with two parallel arms: standard care + up to six MAR interventions over two weeks, versus standard care alone. Participants were 82 adult ICU burn patients. Primary outcome was perceived background pain. Secondary outcomes were anxiety and depression levels, vital signs, and pain medication usage. Additionally, nine patients participated in electroencephalography, electrocardiography, and electromyography recordings during MAR. RESULTS Results showed that background pain was significantly lower in the intervention group than in the control group after MAR sessions 1 and 2, with a large effect for patients reporting severe pain. Time had a significant effect on both anxiety and depression, but this was unrelated to the treatment condition. No differences were found for vital signs or pain medication administration. Analysis of electrophysiological signals during MAR revealed significant changes in delta, theta, beta, and alpha frequency bands during electroencephalography, a significant decrease of the low/high-frequency ratio of the tachogram, and significant decrease of mean frequency of facial electromyography. CONCLUSIONS The results suggest that music therapy might be effective in reducing background pain during the early stages of hospitalization and specifically for patients with severe pain. Further multicenter studies with a larger sample size are needed to confirm these results. TRIAL REGISTRATION www. CLINICALTRIALS gov, Identifier: NCT04571255.