TMS治疗伴焦虑共病重性抑郁精准医疗进展：fMRI聚类亚型、个体化靶向、E-field建模、TMS-EEG与AI闭环调控

There is an urgent need to derive quantitative measures based on coherent neurobiological dysfunctions or ‘biotypes’ to enable stratification of patients with depression and anxiety. We used task-free and task-evoked data from a standardized functional magnetic resonance imaging protocol conducted across multiple studies in patients with depression and anxiety when treatment free (n = 801) and after randomization to pharmacotherapy or behavioral therapy (n = 250). From these patients, we derived personalized and interpretable scores of brain circuit dysfunction grounded in a theoretical taxonomy. Participants were subdivided into six biotypes defined by distinct profiles of intrinsic task-free functional connectivity within the default mode, salience and frontoparietal attention circuits, and of activation and connectivity within frontal and subcortical regions elicited by emotional and cognitive tasks. The six biotypes showed consistency with our theoretical taxonomy and were distinguished by symptoms, behavioral performance on general and emotional cognitive computerized tests, and response to pharmacotherapy as well as behavioral therapy. Our results provide a new, theory-driven, clinically validated and interpretable quantitative method to parse the biological heterogeneity of depression and anxiety. Thus, they represent a promising approach to advance precision clinical care in psychiatry. Personalized brain circuit measures quantified using a new imaging technology in 801 patients with depression and anxiety identify six biotypes with unique symptoms, behaviors and responses to different types of treatment.

We previously identified a cognitive biotype of depression characterized by treatment resistance, impaired cognitive control behavioral performance and dysfunction in the cognitive control circuit, comprising the dorsolateral prefrontal cortex (dLPFC) and dorsal anterior cingulate cortex (dACC). Therapeutic transcranial magnetic stimulation (TMS) to the left dLPFC is a promising option for individuals whose depression does not respond to pharmacotherapy. Here, 43 veterans with treatment-resistant depression were assessed before TMS, after early TMS and post-TMS using functional magnetic resonance imaging during a Go–NoGo paradigm, behavioral cognitive control tests and symptom questionnaires. Stratifying veterans at baseline based on task-evoked dLPFC–dACC connectivity, we demonstrate that TMS-related improvement in cognitive control circuit connectivity and behavioral performance is specific to individuals with reduced connectivity at baseline (cognitive biotype +), whereas individuals with intact connectivity at baseline (cognitive biotype −) did not demonstrate significant changes. Our findings show that dLPFC–dACC connectivity during cognitive control is both a promising diagnostic biomarker for a cognitive biotype of depression and a response biomarker for cognitive improvement after TMS applied to the dLPFC. The authors investigate functional connectivity before and after transcranial magnetic stimulation in veterans with treatment-resistant depression stratified by cognitive biotype, demonstrating associated brain connectivity-mediated improvement in cognitive behavioral task performance.

… to a hierarchical clustering algorithm to identify categorical depression subtypes. … rTMS over the dorsolateral PFC or dorsomedial PFC (1). These results underscore the potential of fMRI-…

… with depression, which were clustered using k-means to identify biologically informed subtypes. Subtype-specific responses to dorsolateral prefrontal cortex rTMS … depression or anxiety …

Increasing evidence suggests that the clinical effects of transcranial magnetic stimulation (TMS) are target-dependent. Within any given symptom, precise targeting of specific brain circuits may improve clinical outcomes. This principle can also be extended across symptoms - stimulation of different circuits may lead to different symptom-level outcomes. This may include targeting different symptoms within the same disorder (such as dysphoria versus anxiety in patients with major depression) or targeting the same symptom across different disorders (such as primary major depression and depression secondary to stroke, traumatic brain injury, epilepsy, multiple sclerosis, or Parkinson's disease). Some of these symptom-specific changes may be desirable, while others may be undesirable. This Review focuses on the conceptual framework via which symptom-specific target circuits may be identified, tested, and implemented.

… 90 participants with major depression, obsessive-compulsive disorder, generalized anxiety disorder, or schizophrenia will receive 40 total sessions of accelerated intermittent theta burst …

… transcranial magnetic stimulation (TMS) (Drysdale et al., 2017a). These resting fMRI biotypes, accounting for clinical profiles of depressed … that robust cluster biotypes can be replicated …

Background Targeting methods for repetitive transcranial magnetic stimulation (rTMS) in patients with depression now include the use of individual functional scans to target specific functional connectivity (FC) patterns obtained from functional magnetic resonance imaging (fMRI). Potential biomarkers of rTMS response include target FC with the subgenual anterior cingulate cortex (SGC) or the causal depression circuit (CDC), each of which may be candidates for individualized functional targets (iFTs). We assessed the relationship of these two approaches to clinical outcomes in two large rTMS clinical trials. Methods 501 subjects with moderate to severe depression underwent 4-6 weeks of daily rTMS to the left dorsolateral prefrontal cortex (DLPFC), targeted using neuronavigation to a common group-based functional target. Resting-state scans acquired at baseline were used to retrospectively compute iFTs using either SGC-DLPFC or CDC-DLPFC FC. The Euclidean distance from the group-based target used in the trial to the centre of gravity of each iFT was computed and correlated with outcomes. Results Most subjects’ iFTs were within 2cm of their group-based target. Proximity to either the SGC- or CDC-iFT was not associated with better outcomes. Sensitivity analyses accounting for treatment target FC, methodology, data quality, or treatment parameters did not change the results. Conclusions Proximity to SGC- or CDC-derived iFTs was not associated with better outcomes in patients who received neuronavigated rTMS to a group-based target. Prospective randomized clinical trials comparing neuronavigated group-based target to neuronavigated iFTs are needed.

Key Points Question How are objective cognitive deficits in depression associated with symptoms, neural circuits, and treatment outcomes? Findings In this secondary analysis of a randomized clinical trial in 1008 patients with major depression, 27% exhibited pretreatment global cognitive impairment and significantly decreased brain response to a cognitive task as well as worse response to standard pharmacotherapy, defining what may be categorized as a cognitive biotype. Changes in cognitive symptoms over the course of treatment mediated the association between pretreatment cognitive status and improvement in overall symptoms and psychosocial functioning. Meaning These results suggest that consideration of treatments targeting cognitive dysfunction in a subset of patients with depression is warranted to attain symptomatic and psychosocial improvement.

Major depressive disorder (MDD) is a leading cause of disability worldwide, affecting over 300 million people and posing a significant burden on healthcare systems. The heterogeneity of MDD can be attributed to diverse etiologic mechanisms. Characterizing MDD subtypes with distinct clinical manifestations could improve patient care through targeted personalized interventions. Topological Data Analysis (TDA) has emerged as a promising tool for identifying homogeneous subgroups of diverse medical conditions and key disease markers. Our study applied TDA to data from a UK Biobank MDD subcohort comprising 3052 samples, leveraging genetic, environmental, and neuroimaging data to assess their differential capability in predicting clinical outcomes in MDD. TDA graphs were built from unimodal and multimodal feature sets and quantitatively compared based on their capability to predict depression severity, physical comorbidities, and treatment response outcomes. Our findings showed a key role of the environment in determining the severity of depressive symptoms. Comorbid medical conditions of MDD were best predicted by brain imaging characteristics, while brain functional patterns resulted in the best predictors of the treatment response profiles. Our results suggest that considering genetic, environmental, and brain characteristics is essential to characterize the heterogeneity of MDD, providing avenues for the definition of robust markers of health outcomes in MDD.

Summary Background The heterogeneity of major depressive disorder (MDD) significantly hinders its effective and optimal clinical outcomes. This study aimed to identify MDD subtypes by adopting a data-driven approach and assessing validity based on symptomatology and neuroimaging. Methods A total of 259 patients with MDD and 92 healthy controls were enrolled in this cross-sectional study. Latent profile analysis (LPA) was used to identify MDD subtypes based on validated clinical symptoms. To examine whether there were differences between these identified MDD subtypes, network analysis was used to test any differences in symptom patterns between these subtypes. We also compared neural activity between these identified MDD subtypes and tested whether certain neural activities were related to individual subtypes. This MDD subtyping was further tested in an independent dataset that contains 86 patients with MDD. Findings Five MDD subtypes with distinct depressive symptom patterns were identified using the LPA model, with the 5-class model selected as the optimal classification solution based on its superior fit indices (AIC = 6656.296, aBIC = 6681.030, entropy = 0.917, LMR p = 0.3267, BLRT p < 0.001). The identified subtypes include atypical-like depression, two melancholic depression (moderate and severe) subtypes with distinct patterns on feeling anxious, and two anhedonic depression subtypes (moderate and severe) with different manifestations on weight/appetite loss. The reproducibility of the classification was also confirmed. Significant differences in symptom structures between melancholic and two anhedonic subtypes, and between anhedonic and atypical subtypes were observed (all p < 0.05). Furthermore, these identified subtypes had differential neural activities in both regional spontaneous neural activity (pFWE < 0.005) and functional connectivity between different brain regions (pFDR < 0.005), linked to different clinical symptoms (FDR q < 0.05). Interpretation The network analysis and neuroimaging tests support the existence and validity of the identified MDD subtypes, each exhibiting unique clinical manifestations and neural activity patterns. The categorisation of these subtypes sheds light on the heterogeneity of depression and suggest that personalised treatment and management strategies tailored to specific subtypes may enhance intervention strategies in clinical settings. Funding 10.13039/100014717National Natural Science Foundation of China (NSFC) and 10.13039/501100004543China Scholarship Council (CSC).

OBJECTIVE The weak link between subjective symptom-based diagnostic methods for posttraumatic psychopathology and objectively measured neurobiological indices forms a barrier to the development of effective personalized treatments. To overcome this problem, recent studies have aimed to stratify psychiatric disorders by identifying consistent subgroups based on objective neural markers. Along these lines, a promising 2021 study by Stevens et al. identified distinct brain-based biotypes associated with different longitudinal patterns of posttraumatic symptoms. Here, the authors conducted a conceptual nonexact replication of that study using a comparable data set from a multimodal longitudinal study of recent trauma survivors. METHODS A total of 130 participants (mean age, 33.61 years, SD=11.21; 48% women) admitted to a general hospital emergency department following trauma exposure underwent demographic, clinical, and neuroimaging assessments 1, 6, and 14 months after trauma. All analyses followed the pipeline outlined in the original study and were conducted in collaboration with its authors. RESULTS Task-based functional MRI conducted 1 month posttrauma was used to identify four clusters of individuals based on profiles of neural activity reflecting threat and reward reactivity. These clusters were not identical to the previously identified brain-based biotypes and were not associated with prospective symptoms of posttraumatic psychopathology. CONCLUSIONS Overall, these findings suggest that the original brain-based biotypes of trauma resilience and psychopathology may not generalize to other populations. Thus, caution is warranted when attempting to define subtypes of psychiatric vulnerability using neural indices before treatment implications can be fully realized. Additional replication studies are needed to identify more stable and generalizable neuroimaging-based biotypes of posttraumatic psychopathology.

Non-invasive brain stimulation (NIBS) treatments have gained considerable attention as a potential therapeutic intervention for psychiatric disorders. The identification of reliable biomarkers for predicting clinical response to NIBS has been a major focus of research in recent years. Neuroimaging techniques, such as electroencephalography (EEG) and (functional) magnetic resonance imaging (fMRI), have been used to identify potential biomarkers that could predict response to NIBS. However, identifying clinically actionable brain biomarkers requires robustness. In this systematic review, we aimed to summarize the current state of brain biomarker research for NIBS in depression, focusing only on well-powered studies (N≥88) and/or studies that aimed at independently replicating prior findings, either successfully or unsuccessfully. A total of 220 studies were initially identified, of which 18 MRI studies and 18 EEG studies adhered to the inclusion criteria, all focused on repetitive transcranial magnetic stimulation treatment in depression. After reviewing the included studies, we found the following MRI and EEG biomarkers to be most robust: 1) fMRI-based functional connectivity between the dorsolateral prefrontal cortex and subgenual anterior cingulate cortex, 2) fMRI-based network connectivity, 3) task-induced EEG frontal-midline theta, and 4) EEG individual alpha frequency. Future prospective studies should further investigate the clinical actionability of these specific EEG and MRI biomarkers to bring biomarkers closer to clinical reality.

Background Postpartum depression is a common and disabling condition that differs from major depressive disorder and shows marked variation in symptoms and outcomes. Identifying distinct biological subtypes could improve diagnosis and treatment. The present study aims to uncover neurophysiological subtypes of postpartum depression and explore their underlying neural and molecular features. Methods We analyzed structural brain images from a cohort of postpartum women recruited at the West China Second Hospital, Sichuan University, including 76 patients with postpartum depression (age range: 24-39 years) and 62 healthy postpartum women (age range: 23-40 years). An unsupervised clustering approach was applied to gray matter volume patterns to identify neurobiological subtypes. Individualized structural covariance networks were then constructed to compare subtype-specific connectivity. Transcriptomic profiles and neurotransmitter density maps were further integrated to examine molecular mechanisms underlying the structural alterations. Results Here we show that postpartum depression can be divided into two neurobiological subtypes. Subtype 1 displays reduced gray matter volume in the dorsal attention network, consistent with cognitive impairments. Subtype 2 shows increased gray matter volume in the default mode network, reflecting emotional dysregulation. Subtype 2 also exhibits weaker structural connectivity between the middle temporal gyrus, parahippocampus, and amygdala. Molecular analysis indicates that Subtype 1 is related to energy metabolism and the neurotransmitter receptor mGluR5, whereas Subtype 2 is associated with synaptic regulation, neuroplasticity, and neurotransmitter receptors such as 5-HT1B, dopamine D2, cholinergic M1 and μ-opioid receptor (MOR). Conclusions These findings suggest that postpartum depression comprises two biologically distinct forms with different cognitive and emotional characteristics. Recognizing these subtypes may enhance our understanding of its neuropathology and support the development of personalized therapeutic strategies.

While brain connectivity can inform coil placement in a given individual, Efield modelling helps to additionally optimize coil orientation, considering the individual cortical folding pattern.In contrast to more commonly used one-size-fits-all approaches, the aim of the multi-modal approach advocated in this symposium is to improve neural target engagement in both local and remote brain regions that are otherwise not directly reachable by TMS.We will demonstrate that the proposed approach has the potential to further improve the specificity and efficacy of TMS in, e.g., clinical applications.First, Deborah Klooster will give an introduction on the topic, presenting theoretical considerations on combining multi-modal brain imaging and E-field modelling as well as first results on the added value of E-field modelling for TMS coil positioning.Afterwards, Maximilian Lueckel will show how to integrate individual whole-brain functional connectivity and E-field modelling for indirect targeting of deep brain structures, including first TMS-fMRI validation results of this approach.Likewise, Charles Lynch will present a novel approach for precise and personalized targeting of individual functional brain networks in densely sampled individuals, also utilizing E-field modelling as well as concurrent TMS-fMRI for validation.Lastly, Bruce Luber will complement the previous talks, showing how utilizing individual structural (instead of functional) connectivity can inform TMS target definition and leads to reliable brain responses in deep, therapeutically relevant brain regions (as measured by TMS-fMRI).

The modeling of TMS-induced electric fields (E-fields) serves as a versatile technique for evaluating and refining brain targeting and dosing strategies, while also providing insights into dose-response relationships in the brain. This review outlines the methodologies employed to derive E-field estimations, covering TMS physics, modeling assumptions, and aspects of subject-specific head tissue and coil modeling. We also summarize various numerical methods for solving the E-field, and their suitability for various applications. Modeling methodologies have been optimized to efficiently execute numerous TMS simulations across diverse scalp coil configurations, facilitating the identification of optimal setups or rapid cortical E-field visualization for specific brain targets. These brain targets are extrapolated from neurophysiological measurements and neuroimaging, enabling precise and individualized E-field dosing in experimental and clinical applications. This necessitates the quantification of E-field estimates using metrics that enable the comparison of brain target engagement, functional localization, and TMS intensity adjustments across subjects. The integration of E-field modeling with empirical data has the potential to uncover pivotal insights into the aspects of E-field responsible for stimulating and modulating brain function and states, enhancing behavioral task performance, and impacting the clinical outcomes of personalized TMS interventions.

BACKGROUND Despite growing interest in personalized rTMS targeting methods using fMRI, there is a notable lack of studies examining the variability and repeatability of these techniques. OBJECTIVE This study investigates the variability and repeatability of personalized rTMS target localization methods, as well as their impact on cortical electric field distribution, using multiple repeated resting-state functional magnetic resonance imaging (rs-fMRI) scans. The goal is to provide a reliable approach for personalizing and precisely targeting rTMS to enhance its clinical efficacy. METHODS A total of 19 healthy subjects participated in the study. Each underwent five repeated rs-fMRI scans (with two runs per session, resulting in 10 runs in total). Five established personalized targeting methods were used to calculate the coordinates of single-run and multiple-run targets. The variability and repeatability of the targets were quantified as the mean Euclidean distance between targets and the probability of locating the same coordinates across scans. These were compared for both single-run and multiple-run targets. Additionally, the electric field distribution for each participant was calculated at various coordinates in individual space. RESULTS Variability in the single-run target ranged from 8.75 mm to 31.48 mm across the five personalized targeting methods. The repeatability for these single-run targets varied between 0.12% and 5.1%. Simulation results revealed that variability in single-run targeting significantly impacted cortical electric field distribution (p<0.0001). In contrast, the variability of multiple-run targets was significantly reduced (p<0.0001), ranging from 1.44 mm to 13.5 mm, and the repeatability improved significantly (p<0.0001), spanning 4.2% to 78.6%. Furthermore, multiple-run targets demonstrated greater consistency in cortical electric field distribution across the five targeting methods. CONCLUSION These findings indicate that personalized rTMS targeting methods based on a single run of rs-fMRI data introduce inherent variability. Averaging data across multiple rs-fMRI scans yields more reliable and stable rTMS targeting, producing a more consistent electric field at the targeted cortical area.

Background: Clinical transcranial magnetic stimulation (TMS) for depression routinely relies on the scalp-based Beam/F3 targeting method to identify stimulation targets in the dorsolateral prefrontal cortex (dLPFC). Scalp-based targeting offers a low-cost and easily implemented method for TMS coil placement, enhancing treatment availability. However, limited anatomical and functional specificity of the Beam/F3 method may affect treatment outcomes, motivating assessment of the clinical standard. Methods: In a naturalistic clinical trial of TMS conduced at four Veterans Affairs hospitals, the authors evaluate the Beam/F3 method using neuroimaging incorporated before TMS, after five treatment sessions, and after all thirty sessions. Personalized anatomical and electric field (E-field) models were developed to assess target location and network engagement, as well as subsequent effects on clinical outcomes. Results: Anatomical models demonstrate that the Beam/F3 method produced reliable targets in the dLPFC across individuals and repeated treatment sessions. E-field models revealed that baseline anticorrelation between the stimulation center and the sgACC was associated with antidepressant symptom response after five TMS sessions (p = 0.032, r2 = 0.100, N = 46) and at the end of treatment (p = 0.042, r2 = 0.107, N = 39). Relatedly, E-field magnitude at the sgACC-anticorrelated peak in the prefrontal cortex correlated with symptom response throughout treatment (early treatment: p = 0.001, r2 = 0.220, N = 46; end of treatment: p = 0.026, r2 = 0.127, N = 39). Conclusions: This work establishes that scalp-based targeting can produce reliable targets in the dLPFC and be successfully evaluated using a combination of neuroimaging and E-field modeling in pragmatic, multisite applications. Importantly, this investigation also found that significant network effects occur early in treatment and that Beam/F3 targets can engage functional mechanisms in TMS.

High-frequency repetitive transcranial magnetic stimulation (rTMS) to the left dorsolateral prefrontal cortex (L-DLPFC) shows promise as a treatment for treatment-resistant depression in adolescents. Conventional rTMS coil placement strategies include the 5 cm, the Beam F3, and the magnetic resonance imaging (MRI) neuronavigation methods. The purpose of this study was to use electric field (E-field) models to compare the three targeting approaches to a computational E-field optimization coil placement method in depressed adolescents. Ten depressed adolescents (4 females, age: 15.9±1.1) participated in an open-label rTMS treatment study and were offered MRI-guided rTMS five times per week over 6–8 weeks. Head models were generated based on individual MRI images, and E-fields were simulated for the four targeting approaches. Results showed a significant difference in the induced E-fields at the L-DLPFC between the four targeting methods (χ2=24.7, p<0.001). Post hoc pairwise comparisons showed that there was a significant difference between any two of the targeting methods (Holm adjusted p<0.05), with the 5 cm rule producing the weakest E-field (46.0±17.4V/m), followed by the F3 method (87.4±35.4V/m), followed by MRI-guided (112.1±14.6V/m), and followed by the computational approach (130.1±18.1V/m). Variance analysis showed that there was a significant difference in sample variance between the groups (K2=8.0, p<0.05), with F3 having the largest variance. Participants who completed the full course of treatment had median E-fields correlated with depression symptom improvement (r=−0.77, p<0.05). E-field models revealed limitations of scalp-based methods compared to MRI guidance, suggesting computational optimization could enhance dose delivery to the target.

Higher‐order cognitive and affective functions are supported by large‐scale networks in the brain. Dysfunction in different networks is proposed to associate with distinct symptoms in neuropsychiatric disorders. However, the specific networks targeted by current clinical transcranial magnetic stimulation (TMS) approaches are unclear. While standard‐of‐care TMS relies on scalp‐based landmarks, recent FDA‐approved TMS protocols use individualized functional connectivity with the subgenual anterior cingulate cortex (sgACC) to optimize TMS targeting. Leveraging previous work on precision network estimation and modeling of the TMS electric field (E‐field), we asked whether various clinical TMS approaches target different functional networks between individuals. Results revealed that modeled homotopic scalp positions (left F3 and right F4) target different networks within and across individuals, and right F4 generally favors a right‐lateralized control network. TMS coil positions over the dorsolateral prefrontal cortex (dlPFC) zone anticorrelated with the sgACC most frequently target a network coupled to the ventral striatum (reward circuitry) but largely miss that network in some individuals. We further illustrate how modeling can be used to retrospectively assess the estimated targets achieved in prior TMS sessions and also used to prospectively provide coil positions that can target distinct closely localized dlPFC network regions with spatial selectivity and maximal E‐field intensity. In a final study, precision targeting was found to be feasible in participants with Major Depressive Disorder using data derived from a single low‐burden MRI session suggesting the methods are applicable to translational efforts where limiting patient burden and ensuring robustness are critical.

… TMSEEG methodology, we discuss how it can be utilized as a tool for understanding individualized … series of studies utilizing E-field modeling to compare the onand off-target degree …

Transcranial magnetic stimulation (TMS) is used to study brain function and treat mental health disorders. During TMS, a coil placed on the scalp induces an E-field in the brain that modulates its activity. TMS is known to stimulate regions that are exposed to a large E-field. Clinical TMS protocols prescribe a coil placement based on scalp landmarks. There are inter-individual variations in brain anatomy that result in variations in the TMS-induced E-field at the targeted region and its outcome. These variations across individuals could in principle be minimized by developing a large database of head subjects and determining scalp landmarks that maximize E-field at the targeted brain region while minimizing its variation using computational methods. However, this approach requires repeated execution of a computational method to determine the E-field induced in the brain for a large number of subjects and coil placements. We developed a probabilistic matrix decomposition-based approach for rapidly evaluating the E-field induced during TMS for a large number of coil placements due to a pre-defined coil model. Our approach can determine the E-field induced in over 1 Million coil placements in 9.5 h, in contrast, to over 5 years using a brute-force approach. After the initial set-up stage, the E-field can be predicted over the whole brain within 2-3 ms and to 2% accuracy. We tested our approach in over 200 subjects and achieved an error of <2% in most and <3.5% in all subjects. We will present several examples of bench-marking analysis for our tool in terms of accuracy and speed. Furthermore, we will show the methods' applicability for group-level optimization of coil placement for illustration purposes only. The software implementation link is provided in the appendix.

Transcranial Magnetic Stimulation (TMS), a non-invasive neuromodulation technique based on electromagnetic induction, modulates cortical excitability by inducing currents with a magnetic field. TMS has demonstrated significant clinical potential in the treatment of various neuropsychiatric disorders, including depression, anxiety, and Parkinson’s disease. However, conventional TMS targeting methods that rely on anatomical landmarks do not adequately account for individual differences in brain structure and functional networks, leading to considerable variability in treatment responses. In recent years, advances in neuroimaging techniques–such as functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI)–together with the application of machine learning (ML) and artificial intelligence (AI) algorithms in big data analysis, have provided novel approaches for precise TMS targeting and individualized treatment. This review summarizes the latest developments in the integration of multimodal neuroimaging and AI technologies for precision neuromodulation with TMS. It focuses on critical issues such as imaging resolution, AI model generalizability, real-time feedback modulation, as well as data privacy and ethical considerations. Future prospects including closed-loop TMS control systems, cross-modal data fusion, and AI-assisted brain-computer interfaces (BCIs) are also discussed. Overall, AI-driven personalized TMS strategies hold promise for markedly enhancing treatment precision and clinical efficacy, thereby offering new theoretical and practical guidance for individualized treatment in neuropsychiatric and neurodegenerative disorders.

… near real-time TMS E-field modeling and their advantages and … can integrate E-field modeling with navigated TMS, including … comparison of the E-field intensity in the target area and the …

… The dmPFC represents a novel target for rTMS and is a key region implicated in goal-… rTMS at the left dmPFC using individualized fMRI-derived targets. 21,600 total pulses are delivered …

INTRODUCTION Precise targeting and dosing are critical for neurophysiological effectivity of repetitive transcranial magnetic stimulation (rTMS), particularly in clinical applications such as treating major depressive disorder (MDD). While neuronavigation enables accurate, individualized coil positioning, even small deviations in coil placement, e.g. during extended stimulation protocols, can significantly alter the induced electric field (E-field). In this study, we use continuous neuronavigational monitoring during stimulation to quantify motion-induced E-field variability at the target and introduce a novel methodology for compensating it. METHODS We analysed coil-target movement parameters in a sample of 200 rTMS sessions conducted in 20 adults with MDD, evaluating position, rotation and main axes of movement. In addition, we simulated induced E-fields within a target-ROI and quantified variability within- and across-sessions. To mitigate movement-related variability, we developed an algorithm which enables real-time adjustment of stimulator output based on current coil position and interpolation of precomputed E-fields. RESULTS Our results show that E-field variability in this sample was primarily driven by coil displacement along the scalp-normal and rotation. Lateral movement played a minor role. Using the new stimulation amplitude adjustment strongly reduced target E-field variability. Mean E-field coefficient of variation was reduced within-session by 41% (2.85% to 1.67%) and across-sessions by 74% (6.77% to 1.73%). DISCUSSION This study presents the first quantitative analysis of motion during rTMS treatment sessions and a practical method to compensate for it. Given its low computational cost, the proposed approach is well suited for clinical implementation, potentially enhancing treatment reliability, particularly in individuals prone to motion.

… Here, we explore the potential of electric field (e-field) based TMS … and e-field-based TMS dosing individualize the stimulation strength based on motor-cortex excitability. When targeting …

Highlights • The spatial dispersion of functional targets is 2.63 times that of anatomical targets.• E-field at the functional target is significantly higher than that at nearby anatomical target (P < 0.001), when the coil is positioned above functional target.• Coil orientation selectively modulates E-fields. With the handle perpendicular to the anatomical-functional target axis, functional target E-field was maintained while anatomical target E-field was suppressed.

Repetitive transcranial magnetic stimulation (rTMS) holds promise for treating psychiatric disorders; however, the variability in treatment efficacy among individuals underscores the need for further improvement. Growing evidence has shown that TMS induces a broad network modulatory effect, and its effectiveness may rely on accurate modulation of the pathological network specific to each disorder. Therefore, determining the optimal TMS coil setting that will engage the functional pathway delivering the stimulation is crucial. Compared to group-averaged functional connectivity (FC), individual FC provides specific information about a person’s brain functional architecture, offering the potential for more accurate network targeting for personalized TMS. However, the low signal-to-noise ratio (SNR) of FC poses a challenge when utilizing individual resting-state FC. To overcome this challenge, the proposed solutions include increasing the scan duration and employing the cluster method to enhance the stability of FC. This study aimed to evaluate the stability of a personalized FC-based network targeting model in individuals with major depressive disorder or schizophrenia with auditory verbal hallucinations. Using resting-state functional magnetic resonance imaging data from the Human Connectome Project, we assessed the model’s stability. We employed longer scan durations and cluster methodologies to improve the precision in identifying optimal individual sites. Our findings demonstrate that a scan duration of 28 minutes and the utilization of the cluster method achieved stable identification of individual sites, as evidenced by the intraindividual distance falling below the ~1cm spatial resolution of TMS. The current model provides a feasible approach to obtaining stable personalized TMS targets from the scalp, offering a more accurate method of TMS targeting in clinical applications.

fMRI-guided nTMS. A short-term ‘virtual lesion ’ caused by cTBS produced detectable changes after the stimulation. The within-session normalization of behavioral

… , individualized TMS targeting without navigation equipment. … We established a personalized TMS targeting method based … The method was validated using individualized dorsolateral …

OBJECTIVE Transcranial magnetic stimulation (TMS) can efficiently and robustly modulate synaptic plasticity, but little is known about how TMS affects functional connectivity (rs-fMRI). Accordingly, this project characterized TMS-induced rsFC changes in depressed patients who received 3 days of left prefrontal intermittent theta burst stimulation (iTBS). METHODS rs-fMRI was collected from 16 subjects before and after iTBS. Correlation matrices were constructed from the cleaned rs-fMRI data. Electric-field models were conducted and used to predict pre-post changes in rs-fMRI. Site by orientation heatmaps were created for vectors centered on the stimulation site and a control site (contralateral motor cortex). RESULTS For the stimulation site, there was a clear relationship between both site and coil orientation, and connectivity changes. As distance from the stimulation site increased, prediction accuracy decreased. Similarly, as eccentricity from the optimal orientation increased, prediction accuracy decreased. The systematic effects described above were not apparent in the heatmap centered on the control site. CONCLUSIONS These results suggest that rs-fMRI following iTBS changes systematically as a function of the distribution of electrical energy delivered from the TMS pulse, as represented by the e-field model. SIGNIFICANCE This finding lays the groundwork for future studies to individualize TMS targeting based on how predicted rs-fMRI changes might impact psychiatric symptoms.

Background: A promising treatment option for adolescents with treatment-resistant depression is high-frequency repetitive transcranial magnetic stimulation (rTMS) delivered to the left dorsolateral prefrontal cortex (L-DLPFC). Conventional coil placement strategies for rTMS in adults include the 5-cm rule, the Beam F3 method, and the magnetic resonance imaging (MRI) neuronavigation method. The purpose of this study was to compare the three targeting approaches to a computational E-field optimization coil placement method in depressed adolescents. Methods: Ten consenting and assenting depressed adolescents (4 females, age: 15.9 +/- 1.1) participated in an open-label rTMS treatment study. Participants were offered MRI-guided rTMS 5 times per week over 6 to 8 weeks. To compute the induced E-field, a head model was generated based on MRI images, and a figure-8 TMS coil (Neuronetics) was placed over the L-DLPFC using the four targeting approaches. Results: Results show that there was a significant difference in the induced E-field at the L-DLPFC between the four targeting methods (chi-squared = 24.7, p < 0.001). Post hoc pairwise comparisons show that there was a significant difference between any two of the targeting methods (Holm adjusted p < 0.05), with the 5-cm rule producing the weakest E-field (46.0 +/- 17.4 V/m), followed by the F3 method (87.4 +/- 35.4 V/m), followed by the MRI-guided (112.1 +/-14.6 V/m), and followed by the computationally optimized method (130.1 +/- 18.1 V/m). The Bartlett test of homogeneity of variances show that there was a significant difference in sample variance between the groups (K^2 = 8.0, p < 0.05), with F3 having the largest variance. In participants who completed the full course of treatment, the median E-field strength in the L-DLPFC was correlated with the change in depression severity (r = -0.77, p < 0.05). Conclusions: The E-field models revealed inadequacies of scalp-based targeting methods compared to MRI-guidance. Computational optimization may further enhance E-field dose delivery to the treatment target.

Transcranial magnetic stimulation (TMS) combined with electroencephalography (EEG) provides a powerful framework to probe and modulate human cortical and corticospinal excitability. In recent years, brain state-dependent EEG–TMS paradigms have gained increasing interest by synchronizing stimulation to ongoing neural activity. However, traditional approaches relying on single oscillatory features or fixed thresholds have yielded heterogeneous and often inconsistent results, motivating the adoption of machine learning (ML) and artificial intelligence (AI) methods to model brain state in a multivariate, data-driven manner. This review synthesizes current ML and deep learning (DL) approaches aimed at predicting cortical and corticospinal excitability from pre-stimulus EEG. We contextualize these methods within brain state-dependent EEG–TMS frameworks based on oscillatory phase, power, and network-level features, and within evolving definitions of brain state that move beyond local biomarkers toward distributed, large-scale, and dynamically evolving neural representations. The reviewed studies span feature-engineered models, data-driven decoding approaches, and emerging adaptive closed-loop frameworks. Finally, we discuss key methodological challenges, translational barriers, and future directions toward personalized, interpretable, and fully closed-loop neuromodulation systems.

… modulate the brain state based on precise individual characteristics in an adaptive closed-loop … We illustrated this vision through a practical application of a truly adaptive closed-loop …

Transcranial magnetic stimulation (TMS) is an established treatment for depression, yet response rates remain at 50% 1 month after treatment. Despite two decades of clinical use, substantial room for improvement remains. This overview examines biomarker-guided personalization of brain stimulation. The authors trace the translational path from invasive circuit-level insights through noninvasive biomarkers to clinical deployment. With validated biomarkers, systematic optimization becomes possible: stimulation parameter tuning, state-dependent approaches, augmentation strategies, and closed-loop systems. The path forward requires randomized trials demonstrating that biomarker-guided personalization improves outcomes beyond standard care, justifying increased complexity and costs. Success would mark an important pathway in interventional psychiatry's evolution to precision medicine.

Transcranial magnetic stimulation (TMS) is a non-invasive FDA-approved therapy for major depressive disorder (MDD), specifically for treatment-resistant depression (TRD). Though offering promise for those with TRD, its effectiveness is less than one in two patients (i.e., less than 50%). Limits on efficacy may be due to individual patient variability, but to date, there are no established biomarkers or measures of target engagement that can predict efficacy. Additionally, TMS efficacy is typically not assessed until a six-week treatment ends, precluding interim re-evaluations of the treatment. Here, we report results using a closed-loop phase-locked repetitive TMS (rTMS) treatment that synchronizes the delivery of rTMS based on the timing of the pulses relative to a patient’s individual electroencephalographic (EEG) prefrontal alpha oscillation indexed by functional magnetic resonance imaging (fMRI). Among responders, synchronized rTMS produces two systematic changes in brain dynamics: a reduction in global cortical excitability and enhanced phase entrainment of cortical dynamics. These effects predict clinical outcomes in the synchronized treatment group but not in an active-treatment unsynchronized control group. The systematic decrease in excitability and increase in entrainment correlated with treatment efficacy at the endpoint and intermediate weeks during the synchronized treatment. Specifically, we show that weekly biomarker tracking enables efficacy prediction and dynamic adjustments through a treatment course, improving the overall response rates. This innovative approach advances the prospects of individualized medicine in MDD and holds potential for application in other neuropsychiatric disorders.

One of the most critical challenges in using noninvasive brain stimulation (NIBS) techniques for the treatment of psychiatric and neurologic disorders is inter- and intra-individual variability in response to NIBS. Response variations in previous findings suggest that the one-size-fits-all approach does not seem the most appropriate option for enhancing stimulation outcomes. While there is a growing body of evidence for the feasibility and effectiveness of individualized NIBS approaches, the optimal way to achieve this is yet to be determined. Transcranial electrical stimulation (tES) is one of the NIBS techniques showing promising results in modulating treatment outcomes in several psychiatric and neurologic disorders, but it faces the same challenge for individual optimization. With new computational and methodological advances, tES can be integrated with real-time functional magnetic resonance imaging (rtfMRI) to establish closed-loop tES-fMRI for individually optimized neuromodulation. Closed-loop tES-fMRI systems aim to optimize stimulation parameters based on minimizing differences between the model of the current brain state and the desired value to maximize the expected clinical outcome. The methodological space to optimize closed-loop tES fMRI for clinical applications includes (1) stimulation vs. data acquisition timing, (2) fMRI context (task-based or resting-state), (3) inherent brain oscillations, (4) dose-response function, (5) brain target trait and state and (6) optimization algorithm. Closed-loop tES-fMRI technology has several advantages over non-individualized or open-loop systems to reshape the future of neuromodulation with objective optimization in a clinically relevant context such as drug cue reactivity for substance use disorder considering both inter and intra-individual variations. Using multi-level brain and behavior measures as input and desired outcomes to individualize stimulation parameters provides a framework for designing personalized tES protocols in precision psychiatry.

In the same way that beauty lies in the eye of the beholder, what a stimulus does to the brain is determined not simply by the nature of the stimulus but by the nature of the brain that is receiving the stimulus at that instant in time. Over the past decades, therapeutic brain stimulation has typically applied open-loop fixed protocols and has largely ignored this principle. Only recent neurotechnological advancements have enabled us to predict the nature of the brain (i.e., the electrophysiological brain state in the next instance in time) with sufficient temporal precision in the range of milliseconds using feedforward algorithms applied to electroencephalography time-series data. This allows stimulation exclusively whenever the targeted brain area is in a prespecified excitability or connectivity state. Preclinical studies have shown that repetitive stimulation during a particular brain state (e.g., high-excitability state), but not during other states, results in lasting modification (e.g., long-term potentiation) of the stimulated circuits. Here, we survey the evidence that this is also possible at the systems level of the human cortex using electroencephalography-informed transcranial magnetic stimulation. We critically discuss opportunities and difficulties in developing brain state–dependent stimulation for more effective long-term modification of pathological brain networks (e.g., in major depressive disorder) than is achievable with conventional fixed protocols. The same real-time electroencephalography-informed transcranial magnetic stimulation technology will allow closing of the loop by recording the effects of stimulation. This information may enable stimulation protocol adaptation that maximizes treatment response. This way, brain states control brain stimulation, thereby introducing a paradigm shift from open-loop to closed-loop stimulation.

BACKGROUND Entrainment (increase) and modulation (shift) of intrinsic brain oscillations via rhythmic-TMS (rh-TMS) enables to either increase the amplitude of the individual peak oscillatory frequency, or experimentally slowing/accelerating this intrinsic peak oscillatory frequency by slightly shifting it. Both entrainment, and modulation of brain oscillations can lead to different measurable perceptual and cognitive changes. However, there are noticeable between-participant differences in such experimental entrainment outcomes. OBJECTIVE/HYPOTHESIS The current study aimed at explaining these inter-individual differences in entrainment/frequency shift success. Here we hypothesize that the width and the height of the Arnold tongue, i.e., the frequency offsets that can still lead to oscillatory change, can be individually modelled via resting-state neural markers, and may explain and predict efficacy and limitation of successful rhythmic-TMS (rh-TMS) manipulation. METHODS Spectral decomposition of resting-state data was used to extract the spectral curve of alpha activity, serving as a proxy of an individual Arnold tongue. These parameters were then used as predictors of the rh-TMS outcome, when increasing alpha-amplitude (i.e., applying pulse train tuned to the individual alpha frequency, IAF), or modulating the alpha-frequency (i.e., making alpha faster or slower by stimulating at IAF±1Hz frequencies). RESULTS Our results showed that the height of the at-rest alpha curve predicted how well the entrainment increased the intrinsic oscillatory peak frequency, with a higher at-rest spectral curve negatively predicting amplitude-enhancement during entrainment selectively during IAF-stimulation. In contrast, the wider the resting-state alpha curve, the higher the modulation effects aiming to shift the intrinsic frequency towards faster or slower rhythms. CONCLUSION These results not only offer a theoretical and experimental model for explaining the variance across different rh-TMS studies reporting heterogenous rh-TMS outcomes, but also introduce a potential biomarker and corresponding evaluative tool to develop most optimal and personalized rh-TMS protocols, both in research and clinical applications.

Despite considerable advancement of first choice treatment (pharmacological, physical therapy, etc.) over many decades, neurological disorders still represent a major portion of the worldwide disease burden. Particularly concerning, the trend is that this scenario will worsen given an ever expanding and aging population. The many different methods of brain stimulation (electrical, magnetic, etc.) are, on the other hand, one of the most promising alternatives to mitigate the suffering of patients and families when conventional treatment fall short of delivering efficacious treatment. With applications in virtually all neurological conditions, neurostimulation has seen considerable success in providing relief of symptoms. On the other hand, a large variability of therapeutic outcomes has also been observed, particularly in the usage of non-invasive brain stimulation (NIBS) modalities. Borrowing inspiration and concepts from its pharmacological counterpart and empowered by unprecedented neurotechnological advancement, the neurostimulation field has seen in recent years a widespread of methods aimed at the personalization of its parameters, based on biomarkers of the individuals being treated. The rationale is that, by taking into account important factors influencing the outcome, personalized stimulation can yield a much-improved therapy. Here, we review the literature to delineate the state-of-the-art of personalized stimulation, while also considering the important aspects of the type of informing parameter (anatomy, function, hybrid), invasiveness, and level of development (pre-clinical experimentation versus clinical trials). Moreover, by reviewing relevant literature on closed loop neuroengineering solutions in general and on activity dependent stimulation method in particular, we put forward the idea that improved personalization may be achieved when the method is able to track in real time brain dynamics and adjust its stimulation parameters accordingly. We conclude that such approaches have great potential of promoting the recovery of lost functions and enhance the quality of life for patients.

… 本文结合这两个特点提出 并系统梳理了同步 TMS-EEG 在心理学… 为应用同步 TMS-EEG 技术提供清 晰的理论框架与实践指南. … : An emerging tool to study the neurophysiologic biomarkers …

… PostTMS EEG analysis further confirmed network-specific modulation in the sensitive group. … represents a significant leap towards a closed-loop digital healthcare paradigm in mental …

… represents a promising biomarker and outcome metric to … responsiveness, particularly in network modulation. Additionally, I … on our human work using closed-loop TMS-EEG at the motor …

Transcranial magnetic stimulation combined with electroencephalography (TMS-EEG) has emerged as a transformative tool for probing cortical dynamics with millisecond precision. This review examines the state-dependent nature of TMS-EEG, a critical yet underexplored dimension influencing measurement reliability and clinical applicability. By integrating TMS’s neuromodulatory capacity with EEG’s temporal resolution, this synergy enables real-time analysis of brain network dynamics under varying neural states. We delineate foundational mechanisms of TMS-evoked potentials (TEPs), discuss challenges posed by temporal and inter-individual variability, and evaluate advanced paradigms such as closed-loop and task-embedded TMS-EEG. The former leverages real-time EEG feedback to synchronize stimulation with oscillatory phases, while the latter aligns TMS pulses with task-specific cognitive phases to map transient network activations. Current limitations—including hardware constraints, signal artifacts, and inconsistent preprocessing pipelines—are critically analyzed. Future directions emphasize adaptive algorithms for neural state prediction, phase-specific stimulation protocols, and standardized methodologies to enhance reproducibility. By bridging mechanistic insights with personalized neuromodulation strategies, state-dependent TMS-EEG holds promise for advancing both basic neuroscience and precision medicine, particularly in psychiatric and neurological disorders characterized by dynamic neural dysregulation.

Symptom provocation is a well-established component of psychiatric research and therapy. It is hypothesized that specific activation of those brain circuits involved in the symptomatic expression of a brain pathology makes the relevant neural substrate accessible as a target for therapeutic interventions. For example, in the treatment of obsessive-compulsive disorder (OCD), symptom provocation is an important part of psychotherapy and is also performed prior to therapeutic brain stimulation with transcranial magnetic stimulation (TMS). Here, we discuss the potential of symptom provocation to isolate neurophysiological biomarkers reflecting the fluctuating activity of relevant brain networks with the goal of subsequently using these markers as targets to guide therapy. We put forward a general experimental framework based on the rapid switching between psychiatric symptom states. This enable neurophysiological measures to be derived from EEG and/or TMS-evoked EEG measures of brain activity during both states. By subtracting the data recorded during the baseline state from that recorded during the provoked state, the resulting contrast would ideally isolate the specific neural circuits differentially activated during the expression of symptoms. A similar approach enables the design of effective classifiers of brain activity from EEG data in Brain-Computer Interfaces (BCI). To obtain reliable contrast data, psychiatric state switching needs to be achieved multiple times during a continuous recording so that slow changes of brain activity affect both conditions equally. This is achieved easily for conditions that can be controlled intentionally, such as motor imagery, attention, or memory retention. With regard to psychiatric symptoms, an increase can often be provoked effectively relatively easily, however, it can be difficult to reliably and rapidly return to a baseline state. Here, we review different approaches to return from a provoked state to a baseline state and how these may be applied to different symptoms occurring in different psychiatric disorders.

Neurological and psychiatric disorders such as Parkinson’s disease, essential tremor, epilepsy, Tourette’s syndrome, depression, and chronic pain remain major causes of disability worldwide. For patients who fail to respond to medication, neuromodulation, particularly deep-brain stimulation (DBS), has become a cornerstone therapy. Traditional open-loop DBS delivers continuous stimulation using pre-set parameters, yielding substantial clinical benefits but also limitations, including side effects, energy inefficiency, and lack of adaptability to dynamic brain states. These drawbacks have motivated the development of closed-loop, or adaptive, DBS systems, which incorporate real-time biomarkers to adjust stimulation in response to neural or physiological signals. Emerging clinical studies demonstrate that closed-loop approaches can improve symptom control in selected disorders, while consistently reducing stimulation time and prolonging device longevity. Despite promising results, outcomes remain heterogeneous across patients, largely due to variability in biomarkers, algorithms, and methodological approaches. Ethical considerations and technical challenges also remain significant barriers to widespread implementation. This narrative review synthesizes evidence on open- and closed-loop neuromodulation across neurological and psychiatric disorders, emphasizing their comparative advantages, limitations, and translational challenges. We highlight the role of biomarkers, adaptive algorithms, and machine learning in shaping personalized neuromodulation and argue that closed-loop stimulation represents a paradigm shift toward precision medicine. Ultimately, the integration of robust biomarkers, predictive algorithms, and scalable clinical frameworks will be critical to realizing the full potential of closed-loop neuromodulation in transforming brain stimulation therapies.

Introduction: Transcranial magnetic stimulation (TMS) over the left dorsolateral prefrontal cortex (L-DLPFC) is an established intervention for treatment-resistant depression (TRD), yet the underlying therapeutic mechanisms remain not fully understood. Methods: This study employs an integrative approach that combines TMS with concurrent functional magnetic resonance imaging (fMRI) and electroencephalography (EEG), aimed at assessing the acute/immediate effects of TMS on brain network dynamics and their correlation with clinical outcomes. Furthermore, this study explored the brain-state dependent effects of TMS, as the brain-state indexed by the phase of EEG prefrontal alpha oscillation. Results: Our study demonstrates that TMS acutely modulates connectivity within vital brain circuits, particularly the cognitive control and default mode networks. We found that the baseline TMS-evoked responses in the cognitive control and limbic networks significantly predicted clinical improvement in patients receiving a novel EEG-synchronized repetitive TMS treatment. Clinical outcomes in this novel treatment were linked to state-specific TMS-modulated functional connectivity within a pivotal brain circuit of the L-DLPFC and the posterior subgenual anterior cingulate cortex within the limbic system. Conclusions: These findings contribute to our understanding of the therapeutic effects underlying TMS treatment in depression and support the potential of assessing state-dependent TMS effects. This study emphasizes the importance of personalized timing of TMS for optimizing target engagement of specific clinically relevant brain circuits. Our results are crucial for future research into the development of personalized neuromodulation therapies for TRD patients.

… Our results are consistent with previous literature on TMS–EEG biomarkers in MDD, which reported abnormalities in inhibitory components (N45, N100) across multiple cohorts, …

Background Recovery of hand function after stroke is often slow and varies widely between patients. Emerging evidence suggests that rehabilitation may involve “tipping points,” at which neural reorganization accelerates functional gains. In this study, we combined EEG monitoring with closed-loop transcranial magnetic stimulation (TMS) to explore whether such tipping points can be identified and modulated, framed through a quantum-inspired perspective. Methods Sixty stroke patients were enrolled and randomized into a closed-loop TMS group (n = 30) or control group (n = 30). All participants underwent baseline clinical and neurophysiological assessments before an 8-week upper-limb rehabilitation program. In the intervention group, training was supplemented with EEG-triggered closed-loop TMS. Serial EEG recordings, motor evoked potentials (MEP), and TMS-evoked potentials (TEP) were collected alongside clinical scales (FMA-UE, ARAT, grip strength). Assessments were conducted at baseline, Weeks 2, 4, and 6, and at 3-month follow-up. Results Both groups demonstrated gradual improvements, but the closed-loop group showed a marked inflection between Weeks 4–6, with accelerated gains in FMA-UE, ARAT, and grip strength. These functional surges were closely associated with stronger µ-rhythm desynchronization, increased MEP amplitudes, and enhanced TEP N100 responses, reflecting heightened cortical excitability and reorganization. Between-group comparisons confirmed greater improvements in the closed-loop group (all P &lt; 0.05). The recovery trajectory aligns with a quantum-inspired model, where rehabilitation resembles a superposition of neural pathways until a collapse into the optimal route drives rapid functional transition. Conclusions Closed-loop TMS can catalyze tipping points in post-stroke hand recovery. EEG and neurophysiological markers may serve as predictive signals for these critical transitions. A quantum-inspired framework provides a novel lens for interpreting abrupt accelerations in recovery and may guide precision rehabilitation strategies.

… Precision medicine can reshape psychiatric and neurological care … information to uncover disease subtypes and refine treatment … developments of rTMS applications in depression. …

Depression is a highly disabling mental disorder imposing a substantial burden on global public health. Repetitive Transcranial Magnetic Stimulation (rTMS), as a non-invasive physical treatment modality, demonstrates favorable efficacy and safety in treating depression. However, significant inter-individual variability in treatment response exists, with the reliability of target localization being a key factor influencing efficacy. Traditional non-neuronavigated localization methods (e.g., 5-cm rule, Beam F3 method), while operationally convenient, suffer from limited reliability due to failure to account for individual variations in brain anatomy (e.g., cortical folding) and functional connectivity patterns. In recent years, driven by advances in magnetic resonance imaging (MRI) technology and individualized treatment paradigms, neuronavigated localization methods based on clinical symptom subtypes and patient-specific brain structural/functional connectivity profiles have significantly enhanced localization reliability and personalization, offering novel approaches to overcome efficacy variability. This review systematically summarizes the mechanisms of action and standard treatment protocols of rTMS for depression, with a primary focus on research advances in target localization methodologies. It encompasses the principles, clinical applications, efficacy comparisons, and optimized integration of both non-neuronavigated and neuronavigated techniques across different populations (adolescents, elderly) and symptom subtypes. By critically analyzing current research achievements and challenges, this review aims to provide clinicians with theoretical foundations and practical references for optimizing rTMS treatment protocols, enhancing response rates, and advancing individualized neuronavigated protocols.

Depressive spectrum disorders are considered among the most common in the general population. Major depressive disorder and persistent depressive disorder (or dysthymia) are the most recognized, but other depressive disorders exist with varying or no specificity. The main difference between major depressive disorder and dysthymia lies in the duration and intensity of symptoms. Improving our understanding of its etiology and pathogenesis must be a priority for health and safety. Given the complexity of the evidence in the literature, it was deemed useful to provide a comprehensive summary of the neuroanatomical dysfunctions currently identified, with particular attention to the anterior and medial cingulate cortex, dorsolateral and ventromedial prefrontal cortex, posterior parietal cortex, insula, amygdala, and hippocampus. Significant neural network alterations include hyperconnectivity of the default mode network (DMN), impairment of the executive control network (ECN), and dysfunction of the salience network (Salience Network). Neurophysiological markers reveal frontal alpha asymmetries and front-striatal metabolic alterations. Studying neural correlates is essential to deepen our understanding of the depressive spectrum and the development of personalized therapeutic interventions, including noninvasive neurostimulation techniques and target-specific pharmacological therapies, opening new avenues for translational research in neuropsychiatric settings.

While the left subgenual anterior cingulate cortex (sgACC) is a validated target for treating major depressive disorder (MDD) (e.g., with rTMS), the therapeutic relevance of the right sgACC is unknown. Given emerging evidence that treatment‐resistant depression (TRD) and non‐TRD (nTRD) may involve distinct neural circuits, a more precise, circuit‐based understanding is needed. This study, therefore, sought to determine the degree of sgACC functional lateralization and to elucidate TRD‐specific neural mechanisms, thereby informing the development of targeted neuromodulation therapies. Resting‐state fMRI (rs‐fMRI) data were acquired for 20 patients with TRD, 53 patients with nTRD, and 51 healthy controls (HCs). Analysis was performed to compare the static functional connectivity (sFC), dynamic functional connectivity (dFC), and effective connectivity via Granger causality analysis (GCA) of left and right sgACC between the three groups, followed by clinical correlation analyses. Group analysis revealed significant sFC and dFC between left sgACC and temporal pole (TP), fusiform gyrus (FG), parahippocampal gyrus (PHG), and orbitofrontal cortex (OFC), and between right sgACC and cerebellum posterior lobule (CPL) and right dorsolateral prefrontal cortex (DLPFC.R). Furthermore, GCA showed increased effective connectivity between bilateral sgACC and left caudate (CAU.L), left thalamus (THA.L), right postcentral gyrus (PoCG.R), and DLPFC.L. Compared to the nTRD group the TRD group showed decreased sFC and dFC between sgACC and DLPFC and cerebellum, along with decreased effective connectivity between sgACC and CAU and DLPFC, indicating disrupted fronto‐limbic and cerebellar circuits in TRD. The distinct sgACC‐centered connectivity profiles of TRD and nTRD delineate a neurobiological continuum across MDD subtypes, pinpointing the sgACC, DLPFC, CAU, and cerebellum as key circuit targets for individualized neuromodulation in TRD.

Major depressive disorder (MDD) represents a serious public health concern, negatively affecting individuals’ quality of life and making a substantial contribution to the global burden of disease. Anhedonia is a core symptom of MDD and is associated with poor treatment outcomes. Variability in anhedonia components within MDD has been observed, suggesting heterogeneity in psychopathology across subgroups. However, little is known about anhedonia subgroups in MDD and their underlying neural correlates across subgroups. To address this question, we employed a hierarchical cluster analysis based on Temporal Experience of Pleasure Scale subscales in 60 first-episode, drug-naive MDD patients and 32 healthy controls. Then we conducted a connectome-wide association study and whole-brain voxel-wise functional analyses for identified subgroups. There were three main findings: (1) three subgroups with different anhedonia profiles were identified using a data mining approach; (2) several parts of the reward network (especially pallidum and dorsal striatum) were associated with anticipatory and consummatory pleasure; (3) different patterns of within- and between-network connectivity contributed to the disparities of anhedonia profiles across three MDD subgroups. Here, we show that anhedonia in MDD is not uniform and can be categorized into distinct subgroups, and our research contributes to the understanding of neural underpinnings, offering potential treatment directions. This work emphasizes the need for tailored approaches in the complex landscape of MDD. The identification of homogeneous, stable, and neurobiologically valid MDD subtypes could significantly enhance our comprehension and management of this multifaceted condition.

Abstract Psychiatric disorders are associated with significant social and economic burdens, many of which are related to issues with current diagnosis and treatments. The coronavirus (COVID-19) pandemic is estimated to have increased the prevalence and burden of major depressive and anxiety disorders, indicating an urgent need to strengthen mental health systems globally. To date, current approaches adopted in drug discovery and development for psychiatric disorders have been relatively unsuccessful. Precision psychiatry aims to tailor healthcare more closely to the needs of individual patients and, when informed by neuroscience, can offer the opportunity to improve the accuracy of disease classification, treatment decisions, and prevention efforts. In this review, we highlight the growing global interest in precision psychiatry and the potential for the National Institute of Health-devised Research Domain Criteria (RDoC) to facilitate the implementation of transdiagnostic and improved treatment approaches. The need for current psychiatric nosology to evolve with recent scientific advancements and increase awareness in emerging investigators/clinicians of the value of this approach is essential. Finally, we examine current challenges and future opportunities of adopting the RDoC-associated translational and transdiagnostic approaches in clinical studies, acknowledging that the strength of RDoC is that they form a dynamic framework of guiding principles that is intended to evolve continuously with scientific developments into the future. A collaborative approach that recruits expertise from multiple disciplines, while also considering the patient perspective, is needed to pave the way for precision psychiatry that can improve the prognosis and quality of life of psychiatric patients.

Aim The cognitive biotype of depression has been conceptualized as a distinct subtype characterized by unique distinct neural correlates and specific clinical features. Abnormal neural oscillations related to cognitive dysfunction have been extensively studied, with particular attention given to gamma oscillations due to their crucial role in neurocircuit operations, emotional processing, and cognitive functions. Nevertheless, cross-frequency coupling between low frequency and gamma oscillations in the cognitive biotype of depression have yet to be fully elucidated. Method The study identified the cognitive biotype in depression by MATRICS Consensus Cognitive Battery (MCCB). We enrolled 141 depressed patients in remission, including 56 identified as cognitive biotype and 85 as the non-cognitive impairment subgroup. Cross-frequency coupling between low frequency and gamma oscillations were analyzed using specific computational methods based on the data collected by Electroencephalogram (EEG). Furthermore, we did correlation analysis to explore the relationship between cross-frequency coupling of neural oscillations with cognitive function in depression. Results We found that phase-amplitude coupling (PAC) values decreased in cognitive biotype. Specifically, cross-frequency coupling between theta (Pz: t =-3.512, FDR-corrected p = 0.011), alpha (P3: t =-3.377, FDR-corrected p = 0.009; Pz: t =-3.451, FDR-corrected p = 0.009), beta (P3: t =-3.129, FDR-corrected p = 0.020; Pz: t =-3.333, FDR-corrected p = 0.020) with low gamma decreased at eyes-closed state in cognitive biotype. However, cross-frequency coupling between delta with gamma increased in cognitive biotype (P4: t = 3.314, FDR-corrected p = 0.022) While cross-frequency coupling exhibited no significant differences at eyes-opened state in two subgroups (FDR-corrected p > 0.05). Furthermore, significant correlations between cognitive function and cross-frequency coupling at eyes-closed state were observed. Conclusion These results indicated that the cross-frequency coupling between low frequency and gamma occurred in the parietal lobe in cognitive biotype of depression. These results advance the understanding of neurophysiological mechanisms underlying cognitive deficits and highlight potential biomarkers for precision depression.

Highlights • rTMS of DLPFC or DMPFC improved clinical symptoms and modulated inFC with DMN/FPN, aligning with symptom-specific circuits.• Target-dependent inFC changes correlated with symptom improvement, offering insight into neural mechanisms of rTMS.• Individualized targeting highlights the value of inFC for future studies on rTMS treatment mechanisms.

Current interventions for major depressive disorder (MDD) demonstrate limited and heterogeneous efficacy, highlighting the need for improving the precision of treatment. Although findings have been mixed, resting-state functional connectivity (rsFC) at baseline shows promise as a predictive biomarker. This meta-analysis evaluates the evidence for baseline rsFC as a predictor of treatment outcomes of MDD interventions. We included MDD literature published between 2012 and 2024 that used antidepressants, non-invasive brain stimulation, and cognitive behavioral therapy. Pearson correlations or their equivalents were analyzed between baseline rsFC and treatment outcome. Nodes were categorized according to the type of brain networks they belong to, and pooled coefficients were generated for rsFC connections reported by more than three studies. Among the 16 included studies and 892 MDD patients, data from nine studies were used to generate pooled coefficients for the rsFC connection between the frontoparietal network (FPN) and default mode network (DMN), and within the DMN (six studies each, with three overlapping studies, involving 534 and 300 patients, respectively). The rsFC between the DMN and FPN had a pooled predictability of -0.060 (p = 0.171, fixed effect model), and the rsFC within the DMN had a pooled predictability of 0.207 (p < 0.001, fixed effect model). The rsFC between the DMN and FPN and the rsFC within the DMN had a larger effect in predicting the outcome of non-invasive brain stimulation (-0.215, p < 0.001, fixed effect model) and antidepressants (0.315, p < 0.001, fixed effect model), respectively. Heterogeneity was observed in both types of rsFC, study design, sample characteristics and data analysis pipeline. Baseline rsFC within the DMN and between the DMN and FPN demonstrated a small but differential predictive effect on the outcome of antidepressants and non-invasive brain stimulation, respectively. The small predictability of rsFC suggested that rsFC between the FPN and DMN and the rsFC within the DMN might not be a good biomarker for predicting treatment outcome. Future research should focus on exploring treatment-specific predictions of baseline rsFC and its predictive utility for other types of MDD interventions. The review was pre-registered at PROSPERO CRD42022370235 (33).

Background: Major Depressive Disorder (MDD) is a significant challenge in modern medicine due to its unclear underlying causes. Brain network dysfunction is believed to play a key role in its pathophysiology. Resting-state functional MRI (rs-fMRI), a neuroimaging technique, enables the in vivo assessment of functional connectivity (FC) between brain regions, offering insights into these network dysfunctions. The aim of this study was to evaluate abnormalities in FC within major brain networks in patients with drug-resistant MDD. Methods: The study group consisted of 26 patients with drug-resistant MDD and an age-matched control group (CG) of 26 healthy subjects. The rs-fMRI studies were performed on a 3T MR scanner (Philips, Ingenia) using a 32-channel head and neck coil. Imaging data were statistically analyzed, focusing on the intra- and inter-network FC of the following networks: default mode (DMN), sensorimotor (SMN), visual (VN), salience (SN), cerebellar (CN), dorsal attention (DAN), language (LN), and frontoparietal (FPN). Results: In patients with MDD, the intra-network analysis showed significantly decreased FC between nodes within VN compared to CG. In contrast, the inter-network analysis showed significantly increased FC between nodes from VN and SN or VN and DAN compared to CG. Decreased FC was found between SN and CN or SN and FPN as well as VN and DAN nodes compared to CG. Conclusions: Patients with MDD showed significant abnormalities in resting-state cortical activity, mainly regarding inter-network functional connectivity. These results contribute to the knowledge on the pathomechanism of MDD and may also be useful for developing new treatments.

Major Depressive Disorder (MDD) poses significant health risks, yet diagnosis lacks objective biomarkers. This systematic review synthesizes functional Magnetic Resonance Imaging (fMRI) studies (2020-2025, n = 52) on functional connectivity (FC) in MDD. We found robust FC alterations within and between core networks (Default Mode, Salience, Central Executive), linked to rumination, emotion dysregulation, and cognitive deficits. These alterations varied with suicidal ideation, comorbidities, childhood trauma, and biological sex. Treatments (antidepressants, rTMS, ECT) demonstrated distinct normalization effects on specific networks. This review consolidates evidence for MDD as a "network interaction disorder," moving beyond single-network foci. It highlights the translational potential of fMRI-based FC for refining diagnosis, personalizing treatment, and provides a novel integrative framework for future research.

Background Neuro-navigated repetitive transcranial magnetic stimulation (rTMS) is potentially effective in enhancing cognitive performance in the spectrum of Alzheimer’s disease (AD). We explored the effect of rTMS-induced network reorganization and its predictive value for individual treatment response. Methods Sixty-two amnestic mild cognitive impairment (aMCI) and AD patients were recruited. These subjects were assigned to multimodal magnetic resonance imaging scanning before and after a 4-week stimulation. Then, we investigated the neural mechanism underlying rTMS treatment based on static functional network connectivity (sFNC) and dynamic functional network connectivity (dFNC) analyses. Finally, the support vector regression was used to predict the individual rTMS treatment response through these functional features at baseline. Results We found that rTMS at the left angular gyrus significantly induced cognitive improvement in multiple cognitive domains. Participants after rTMS treatment exhibited significantly the increased sFNC between the right frontoparietal network (rFPN) and left frontoparietal network (lFPN) and decreased sFNC between posterior visual network and medial visual network. We revealed remarkable dFNC characteristics of brain connectivity, which was increased mainly in higher-order cognitive networks and decreased in primary networks or between primary networks and higher-order cognitive networks. dFNC characteristics in state 1 and state 4 could further predict individual higher memory improvement after rTMS treatment (state 1, R  = 0.58; state 4, R  = 0.54). Conclusion Our findings highlight that neuro-navigated rTMS could suppress primary network connections to compensate for higher-order cognitive networks. Crucially, dynamic regulation of brain networks at baseline may serve as an individualized predictor of rTMS treatment response. Relevance statement Dynamic reorganization of brain networks could predict the efficacy of repetitive transcranial magnetic stimulation in the spectrum of Alzheimer’s disease. Key points • rTMS at the left angular gyrus could induce cognitive improvement. • rTMS could suppress primary network connections to compensate for higher-order networks. • Dynamic reorganization of brain networks could predict individual treatment response to rTMS. Graphical Abstract

Depression is a common mental disorder characterized by heterogeneous cognitive and behavioral symptoms. The emerging research paradigm of functional connectomics has provided a quantitative theoretical framework and analytic tools for parsing variations in the organization and function of brain networks in depression. In this review, we first discuss recent progress in depression-associated functional connectome variations. We then discuss treatment-specific brain network outcomes in depression and propose a hypothetical model highlighting the advantages and uniqueness of each treatment in relation to the modulation of specific brain network connectivity and symptoms of depression. Finally, we look to the future promise of combining multiple treatment types in clinical practice, using multisite datasets and multimodal neuroimaging approaches, and identifying biological depression subtypes.

Patients with bipolar disorder (BD) and major depressive disorder (MDD) exhibit depressive episodes with similar symptoms despite having different and poorly understood underlying neurobiology, often leading to misdiagnosis and improper treatment. This exploratory study examined whole-brain functional connectivity (FC) using FC multivariate pattern analysis (fc-MVPA) to identify the FC patterns with the greatest ability to distinguish between currently depressed patients with BD type I (BD I) and those with MDD. In a cross-sectional design, 41 BD I, 40 MDD patients and 63 control participants completed resting state functional magnetic resonance imaging scans. Data-driven fc-MVPA, as implemented in the CONN toolbox, was used to identify clusters with differential FC patterns between BD patients and MDD patients. The identified cluster was used as a seed in a post hoc seed-based analysis (SBA) to reveal associated connectivity patterns, followed by a secondary ROI-to-ROI analysis to characterize differences in connectivity between these patterns among BD I patients, MDD patients and controls. FC-MVPA identified one cluster located in the right frontal pole (RFP). The subsequent SBA revealed greater FC between the RFP and posterior cingulate cortex (PCC) and between the RFP and the left inferior/middle temporal gyrus (LI/MTG) and lower FC between the RFP and the left precentral gyrus (LPCG), left lingual gyrus/occipital cortex (LLG/OCC) and right occipital cortex (ROCC) in MDD patients than in BD patients. Compared with the controls, ROI-to-ROI analysis revealed lower FC between the RFP and the PCC and greater FC between the RFP and the LPCG, LLG/OCC and ROCC in BD patients; in MDD patients, the analysis revealed lower FC between the RFP and the LLG/OCC and ROCC and greater FC between the RFP and the LI/MTG. Differences in the RFP FC patterns between currently depressed patients with BD and those with MDD suggest potential neuroimaging markers that should be further examined. Specifically, BD patients exhibit increased FC between the RFP and the motor and visual networks, which is associated with psychomotor symptoms and heightened compensatory frontoparietal FC to counter distractibility. In contrast, MDD patients exhibit increased FC between the RFP and the default mode network, corresponding to sustained self-focus and rumination.

… population-based FPN-optimized TMS coil placement and its … of the FPN, we compared them against a standard clinical TMS … resembling the default mode network (DMN) appears to be …

Suicide is a complex behavior strongly associated with depression. Despite extensive research, an objective biomarker for evaluating suicide risk precisely and timely is still lacking. Using the precision resting-state fMRI method, we studied 61 depressive patients with suicide ideation (SI) or suicide attempt (SA), and 35 patients without SI to explore functional biomarkers of suicide risk. Among them, 21 participants also completed electroconvulsive therapy (ECT) treatment, allowing the examination of functional changes across different risk states within the same individual. Functional networks were localized in each subject using resting-state fMRI and then an individualized connectome was constructed to represent the subject’s functional brain organization. We identified a set of connections that track suicide risk (r = 0.41, p = 0.001) and found that these risk-associated connections were hyper-connected in the frontoparietal network (FPN, p = 0.008, Cohen’s d = 0.58) in patients with suicide risk compared to those without. Moreover, ECT treatment significantly reduced (p = 0.001, Cohen’s d = 0.56) and normalized these FPN hyper-connections. These findings suggest that connections involving FPN may constitute an important biomarker for evaluating suicide risk and may provide potential targets for interventions such as non-invasive brain stimulation.

Major depressive disorder (MDD) is a devastating mental disorder characterized by considerable clinical and biological heterogeneity. While comparable clinical symptoms may represent a common pathological endpoint, it is conceivable that distinct neurophysiological mechanisms underlie their manifestation. In this study, both static and model-based dynamic functional connectivity were employed as predictive variables in the normative model to map multilevel functional developmental trajectories and determined clusters of distinguishable MDD subgroups in a large multi-site resting fMRI dataset of 2428 participants (healthy controls: N = 1128; MDD: N = 1300). An independent cohort of 72 participants (healthy controls: N = 35; MDD: N = 37) with both resting fMRI and task-based fMRI data was utilized to validate the identified MDD subtypes and explore subtype-specific task-based neural representations. Our findings indicated brain-wide, interpatient heterogeneous multilevel brain functional deviations in MDD. We identified two distinct and reproducible MDD subtypes, exhibiting comparable severity of clinical symptoms but opposing patterns of multilevel functional deviations. Specifically, MDD subtype 1 displayed positive deviations in the frontoparietal and default mode networks, coupled with negative deviations in the occipital and sensorimotor networks. Conversely, MDD subtype 2 exhibited a significantly contrasting deviation pattern. Additionally, we found that these two identified MDD subtypes exhibited different neural representations during empathic processing, while the subtypes did not differ during implicit face processing. These findings underscore the neurobiological complexity of MDD and highlights the need for a multifaceted approach to diagnosis and treatment that can be tailored specifically to individual subtypes, facilitating personalized and more effective interventions for individuals with MDD.

BACKGROUND Evidence of whether disruptions in functional connectivity (FC), particularly within the default-mode network (DMN) and frontoparietal network (FPN), differ between first-episode bipolar disorder (BD) and major depressive disorder (MDD) is inconsistent. METHODS This study included 31 individuals (mean age, 25 years) with first-episode BD, 31 with MDD, and 31 healthy controls. Resting-state FC was assessed using magnetic resonance imaging. FC was analyzed in the DMN and FPN by using Dosenbach's 160 functional regions of interest. RESULTS FC within the DMN (ventromedial prefrontal cortex [vmPFC] and occipital region) and the FPN (ventral anterior PFC and intraparietal sulcus) as well as between the FPN and DMN, specifically between the ventral anterior PFC and occipital region and between the ventral PFC and precuneus, was higher in participants with BD than in healthy controls. Participants with MDD exhibited increased FC only within the anterior DMN, particularly between the vmPFC, superior frontal cortex, and ventrolateral PFC. DISCUSSION First-episode BD is characterized by more extensive FC alterations, with increased connectivity both within and across the FPN and DMN. By contrast, first-episode MDD presents with more localized FC disruptions, confined to the anterior DMN. These distinct patterns of functional dysconnectivity may underlie the greater cognitive impairments observed early in the course of BD compared with MDD.

Rumination is closely linked to the onset and maintenance of major depressive disorder (MDD). Prior neuroimaging studies have identified the association between self‐reported rumination trait and the functional coupling among a network of brain regions using resting‐state functional magnetic resonance imaging (MRI). However, little is known about the underlying neural circuitry mechanism during active rumination in MDD. Degree centrality (DC) is a simple metric to denote network integration, which is critical for higher‐order psychological processes such as rumination. During an MRI scan, individuals with MDD (N = 45) and healthy controls (HC, N = 46) completed a rumination state task. We examined the interaction effect between the group (MDD vs. HC) and condition (rumination vs. distraction) on vertex‐wise DC. We further characterized the identified brain region's functional involvement with Neurosynth and BrainMap. Network‐wise seed‐based functional connectivity (FC) analysis was also conducted for the identified region of interest. Finally, exploratory correlation analysis was conducted between the identified region of interest's network FCs and self‐reported in‐scanner affect levels. We found that a left superior frontal gyrus (SFG) region, generally overlapped with the frontal eye field, showed a significant interaction effect. Further analysis revealed its involvement with executive functions. FCs between this region, the frontoparietal, and the dorsal attention network (DAN) also showed significant interaction effects. Furthermore, its FC to DAN during distraction showed a marginally significant negative association with in‐scanner affect level at the baseline. Our results implicated an essential role of the left SFG in the rumination's underlying neural circuitry mechanism in MDD and provided novel evidence for the conceptualization of rumination in terms of impaired executive control.

… –DMN or DLPFC–FPN functional connectivity mediate the observed improvements in DMN–… -individual functional connectivity variability in TMS targets for major depressive disorder. …

Recognizing facial expressions is dependent on multiple brain networks specialized for different cognitive functions. In the current study participants (N=20) were scanned using functional magnetic resonance imaging (fMRI) while they performed a covert facial expression naming task. Immediately prior to scanning thetaburst transcranial magnetic stimulation (TMS) was delivered over the right lateral prefrontal cortex (PFC), or the vertex control site. A group whole-brain analysis revealed that TMS induced opposite effects in the neural responses across different brain networks. Stimulation of the right PFC (compared to stimulation of the vertex) decreased neural activity in the left lateral PFC but increased neural activity in three nodes of the default mode network (DMN): the right superior frontal gyrus (SFG), right angular gyrus and the bilateral middle cingulate gyrus. A region of interest (ROI) analysis showed that TMS delivered over the right PFC reduced neural activity across all functionally localised face areas (including in the PFC) compared to TMS delivered over the vertex. These results causally demonstrate that visually recognizing facial expressions is dependent on the dynamic interaction of the face processing network and the DMN. Our study also demonstrates the utility of combined TMS / fMRI studies for revealing the dynamic interactions between different functional brain networks.

… are often chosen to achieve a specified E-field dose on targeted brain regions. TMS … E-field dose on the cortex. We introduce a method and develop software for computing brain E-field …

Electromagnetic fields are widely used to investigate brain function and to treat neurological disorders. In particular, electric fields (E-fields) influence the exchange of ions between neurons and their surrounding environment, thereby modulating the activity of targeted brain regions. The design and optimization of brain stimulation technologies increasingly depend on computational electromagnetics. This article provides an overview of E-field modeling in brain stimulation. First, we introduce the computational methods commonly used to determine the E-fields induced during brain stimulation and summarize key findings. Second, we review computational techniques for coil design and electrode placement. Finally, we discuss existing methods to understand and predict how E-fields affect individual neurons. Taken together, these modeling techniques provide a comprehensive perspective on electromagnetic brain stimulation and highlight key challenges and opportunities for future research and technological development.

… optimized both coil and mesh density sampling to achieve a faster solution. However, to maintain fast computation, … We propose a methodology to predict the TMS-induced E-field on the …

Background: Transcranial electrical stimulation (tES) generates an electric field (or current density) in the brain through surface electrodes attached to the scalp. Clinical significance has been demonstrated, although with moderate and heterogeneous results partly due to a lack of control of the delivered electric currents. In the last decade, computational electric field analysis has allowed the estimation and optimization of the electric field using accurate anatomical head models. This review examines recent tES computational studies, providing a comprehensive background on the technical aspects of adopting computational electric field analysis as a standardized procedure in medical applications. Methods: Specific search strategies were designed to retrieve papers from the Web of Science database. The papers were initially screened based on the soundness of the title and abstract and then on their full contents, resulting in a total of 57 studies. Results: Recent trends were identified in individual- and population-level analysis of the electric field, including head models from non-neurotypical individuals. Advanced optimization techniques that allow a high degree of control with the required focality and direction of the electric field were also summarized. There is also growing evidence of a correlation between the computationally estimated electric field and the observed responses in real experiments. Conclusions: Computational pipelines and optimization algorithms have reached a degree of maturity that provides a rationale to improve tES experimental design and a posteriori analysis of the responses for supporting clinical studies.

… balance between computational efficiency and accuracy [5]. Here, we strongly optimize this … As the visualization is currently restricted to the E-field magnitude, update rates were only …

The growing interest in research regarding emerging non-invasive brain stimulation techniques such as transcranial temporal interference stimulation (tIS), imposes a systematic identification of potential hazards and the quantification of the associated risks for which existing safety guidelines may be inadequate.To this purpose, the use of in silico investigations is fundamental, enabling mechanistic investigations and well controlled dosimetric analysis of a vast range of exposure conditions that are impractical or impossible to achieve in clinical environments.We report the results of two systematic in silico studies employing electromagnetic, thermal and electrophysiological simulations to investigate the conditions for the safe application of transcranial electric stimulation (tES) and tIS in healthy subjects (first study) and to assess their safety in subjects with metallic implants (second study).In the first study, skin and brain exposure-related quantities such as E-field, current-density, charge injection, and activating function were evaluated for several exposure conditions to quantify the risk of unwanted stimulation and electrochemical damage, while temperature increase was calculated to quantify the risk of thermal damage.On that basis and by comparison with exposure conditions from established technologies, frequency-dependent input current thresholds for safe tIS application were derived.In the second study, the risks of unwanted stimulation and thermal damage in proximity of active and passive implants (e.g., DBS, SEEG, abandoned leads, stents) was systematically investigated, along with the risk of current rerouting by associated low-impedance pathways and capacitive current injection in leads passing near electrodes.Our results are in agreement with the (limited) data and experience from human experiments and trials by early technology adopters.We will discuss the key findings from both investigations, detailing the methodologies employed, the study limitations, and the broader implications of these results for safety protocols in established and emerging brain stimulation technologies.

… the effects of neurostimulation on brain activity or to optimize stimulation delivery typically … API to compute E-field distributions in head and brain tissues and to optimize NIBS treatment …

… also envisaged in modern adaptive neurofeedback methods. In these … However, a considerable drawback of neurofeedback technologies … Thus, the key feature of closed-loop adaptive …

… phase, forcing current closed-loop methods to rely on 58 … Building on this, we adapted the system for non- 68 … comparability between the two closed-loop runs, we attempted to match the …

… Finally, we explore how the closed-loop framework will … the feedback loop and the adaptive loop that uses the output … In neurofeedback the stimulus is represented by the brain …

Depressive disorders (including both unipolar and bipolar depression) continue to present treatment challenges, with many patients failing to achieve adequate symptom relief. Non-invasive neuromodulation techniques (NINTs) have emerged as promising alternative interventions, particularly valuable in resource-limited settings. Our systematic review, drawing from PubMed and Scopus databases (2016-January 2024), evaluates current evidence on NINTs protocols for depression treatment. Key findings indicate that while most clinical applications still employ open-loop systems, closed-loop approaches utilizing EEG biomarkers (notably alpha peak frequency and frontal-midline theta power) demonstrate enhanced treatment precision. However, significant challenges remain in establishing reliable biomarker protocols for closed-loop implementation. NINTs offer particular advantages for lower middle-income countries (LMICs) through their cost-effectiveness, scalability, and minimal infrastructure requirements, though initial costs and regulatory barriers limit widespread adoption. This review highlights the urgent need for standardized protocols and identifies promising research directions to optimize NINTs' therapeutic potential. The transition from open-loop to biomarker-guided closed-loop systems represents a critical frontier in depression treatment innovation.

Major depressive disorder (MDD) is associated with widespread disruptions in brain network dynamics. Although noninvasive brain stimulation (NIBS) has shown promise as an alternative treatment, its efficacy remains limited due to a lack of individualized targeting strategies that account for functional and topological heterogeneity in brain networks. This study developed a novel EEG-based framework to personalize NIBS strategies in MDD. Resting-state EEG data from 30 healthy controls and 34 MDD patients were analyzed. Functional connectivity was estimated across five frequency bands using phase locking value (PLV), amplitude envelope correlation (AEC), and weighted phase lag index (wPLI). Spectral graph embedding and structural controllability theory were applied to identify candidate stimulation targets. A multi-objective optimization algorithm (NSGA-II) was used to select optimal node–frequency–amplitude combinations minimizing control energy while maximizing network efficiency gain and structural restoration. Kuramoto-based neural simulations were conducted to evaluate stimulation efficacy in silico, quantifying changes in global synchrony, modularity, and local efficiency. MDD patients exhibited hyperconnectivity in PLV and AEC and reduced wPLI compared to controls. Control nodes in MDD were more centrally distributed, particularly around Cz in alpha and beta bands. NSGA-II optimization yielded subject-specific stimulation strategies with favorable trade-offs. Simulated stimulation significantly enhanced global synchrony (median R = 0.68, SD = 0.30), reduced network modularity (median ΔQ = − 0.0017, SD = 2.93), and improved local efficiency (median ΔEff = 0.0158, SD = 0.0038). Individualized stimulation plans consistently outperformed random controls in restoring network-level metrics. The proposed framework enables data-driven, mathematically interpretable, and simulation-validated planning of personalized brain stimulation strategies for MDD. These findings highlight the potential of EEG-based network analysis and multi-objective optimization in guiding precision neuromodulation interventions.

Emerging brain technologies have significantly transformed human life in recent decades. For instance, the closed-loop brain-computer interface (BCI) is an advanced software-hardware system that interprets electrical signals from neurons, allowing communication with and control of the environment. The system then transmits these signals as controlled commands and provides feedback to the brain to execute specific tasks. This paper analyzes and presents the latest research on closed-loop BCI that utilizes electric/magnetic stimulation, optogenetic, and sonogenetic techniques. These techniques have demonstrated great potential in improving the quality of life for patients suffering from neurodegenerative or psychiatric diseases. We provide a comprehensive and systematic review of research on the modalities of closed-loop BCI in recent decades. To achieve this, the authors used a set of defined criteria to shortlist studies from well-known research databases into categories of brain stimulation techniques. These categories include deep brain stimulation, transcranial magnetic stimulation, transcranial direct-current stimulation, transcranial alternating-current stimulation, and optogenetics. These techniques have been useful in treating a wide range of disorders, such as Alzheimer's and Parkinson's disease, dementia, and depression. In total, 76 studies were shortlisted and analyzed to illustrate how closed-loop BCI can considerably improve, enhance, and restore specific brain functions. The analysis revealed that literature in the area has not adequately covered closed-loop BCI in the context of cognitive neural prosthetics and implanted neural devices. However, the authors demonstrate that the applications of closed-loop BCI are highly beneficial, and the technology is continually evolving to improve the lives of individuals with various ailments, including those with sensory-motor issues or cognitive deficiencies. By utilizing emerging techniques of stimulation, closed-loop BCI can safely improve patients' cognitive and affective skills, resulting in better healthcare outcomes.

… low-frequency (LF) rTMS of contralesional M1 for hand motor … applied neurofeedback is obviously a typical closed-loop … sometime called adaptive neuromodulation, too [Citation14]. …

… This is adaptive—it allows learning from past events without … Athletes and performers use biofeedback and neurofeedback to … The closed-loop framework would make this systematic …

Deep brain stimulation involves the administration of electrical stimulation to targeted brain regions for therapeutic benefit. In the context of major depressive disorder (MDD), most studies to date have administered continuous or open-loop stimulation with promising but mixed results. One factor contributing to these mixed results may stem from when the stimulation is applied. Stimulation administration specific to high-symptom states in a personalized and responsive manner may be more effective at reducing symptoms compared to continuous stimulation and may avoid diminished therapeutic effects related to habituation. Additionally, a lower total duration of stimulation per day is advantageous for reducing device energy consumption. This protocol describes an experimental workflow using a chronically implanted neurostimulation device to achieve closed-loop stimulation for individuals with treatment-refractory MDD. This paradigm hinges on determining a patient-specific neural biomarker that is related to states of high symptoms and programming the device detectors, such that stimulation is triggered by this read-out of symptom state. The described procedures include how to obtain neural recordings concurrent with patient symptom reports, how to use these data in a state-space model approach to differentiate low- and high-symptom states and corresponding neural features, and how to subsequently program and tune the device to deliver closed-loop stimulation therapy.

… Currently, personalized treatment of MDD is based on the recognition of the predominant symptoms within a specific clinical setting, distinguishing the subtypes of diagnosis described …

The quest to develop and improve neuromodulatory therapies for treatment-resistant psychiatric disorders has been fueled by the discovery of intracranial neural biomarkers of symptom dimensions. These neural correlates shed light on the underlying neurophysiology of the disorder and may even be useful in guiding therapy delivery. This systematic review summarizes recent efforts in this field relating neural activity to behavior and symptomatology. For years, the majority of these neurobehavioral relationships had been studied in the hospital or clinic environment. Recent technological advances in implanted neuromodulation devices that permit not only stimulation, but also intracranial neural recording have enabled this research to move into natural settings, recording for longer periods of time in the real world. We review this combined literature to identify neurobehavioral relationships that show commonalities across these different recording strategies and environments. We also discuss potential ways to use this information for guiding neuromodulation therapy. The success of "closed loop" stimulation strategies for movement disorders and epilepsy has led to interest in exploring similar approaches for psychiatric disorders. Such efforts, however, need to consider the disorder-specific time constant relating changes in a neural biomarker to changes in symptoms and behavior. This relationship likely differs between Parkinson's disease and depression, OCD, or addiction. We interpret the results of our systematic review in this light to offer suggestions for future closed-loop or "clinician in the loop" implementations to inform the next generation of neuromodulatory therapies.

γ neuromodulation has emerged as a promising strategy for addressing neurological and psychiatric disorders, particularly in regulating executive and cognitive functions. This review explores the latest neuromodulation techniques, focusing on the critical role of γ oscillations in various brain disorders. Direct γ neuromodulation induces γ-frequency oscillations to synchronize disrupted brain networks, while indirect methods influence γ oscillations by modulating cortical excitability. We investigate how monitoring dynamic features of γ oscillations allows for detailed evaluations of neuromodulation effectiveness. By targeting γ oscillatory patterns and restoring healthy cross-frequency coupling, interventions may alleviate cognitive and behavioral symptoms linked to disrupted communication. This review examines clinical applications of γ neuromodulations, including enhancing cognitive function through 40 Hz multisensory stimulation in Alzheimer’s disease, improving motor function in Parkinson’s disease, controlling seizures in epilepsy, and modulating emotional dysfunctions in depression. Additionally, these neuromodulation strategies aim to regulate excitatory-inhibitory imbalances and restore γ synchrony across neurological and psychiatric disorders. The review highlights the potential of γ oscillations as biomarkers to boost restorative results in clinical applications of neuromodulation. Future studies might focus on integrating multimodal personalized protocols, artificial intelligence (AI) driven frameworks for neural decoding, and global multicenter collaborations to standardize and scale precision treatments across diverse disorders.

Major depressive disorder (MDD) is a heterogeneous condition with varied responses to pharmacological, psychotherapeutic, and neuromodulation interventions. Identifying neuroimaging biomarkers predictive of treatment response could facilitate personalized treatment selection. This umbrella review synthesized findings from systematic reviews and meta‐analyses evaluating neuroimaging biomarkers predictive of treatment response in MDD. A comprehensive search was conducted across PubMed, Scopus, Web of Science, and Embase. Fourteen systematic reviews and meta‐analyses, encompassing 17,855 individuals with MDD, were included. Imaging modalities assessed included structural MRI, functional MRI (resting‐state and task‐based), diffusion tensor imaging, PET, MEG, and fNIRS. Methodological quality was evaluated using the Measurement Tool to Assess Systematic Reviews (AMSTAR 2 tool). The most consistently predictive biomarkers were increased volume and activity in the anterior cingulate cortex (ACC) and hippocampus and altered functional connectivity in the default mode network (DMN) and fronto‐limbic circuits. Predictive patterns varied by treatment modality: for example, larger hippocampal volume predicted pharmacotherapy response, while smaller hippocampal volume was associated with better outcomes in ECT. Machine learning models integrating multimodal data achieved high predictive accuracy (AUC >0.85), though most lacked external validation. Evidence quality was low to very low among the included studies due to methodological heterogeneity. Neuroimaging biomarkers, particularly involving ACC, hippocampus, and large‐scale functional networks, hold promise for guiding treatment selection in MDD. Integration of multimodal imaging and computational approaches may enhance predictive accuracy. However, standardization and prospective validation in clinical settings are needed for translation into practice.

Background: Deep brain stimulation (DBS) is currently a promising technique for psychiatric patients with severe and treatment-resistant symptoms. However, the results to date have been quite heterogeneous, and the indications for psychosurgery with DBS remain in an experimental phase. One of the major challenges limiting the advancement of DBS in psychiatric disorders is the lack of objective criteria for diagnosing certain conditions, which are often based more on clinical scales rather than measurable biological markers. Additionally, there is a limited capacity to objectively assess treatment outcomes. Methods: This overview examines the literature on the available biomarkers in psychosurgery in relation to DBS, as well as other relevant biomarkers in psychiatry with potential applicability for this treatment modality. Results: There are five types of biomarkers: clinical/behavioral, omic, neuroimaging, electrophysiological, and neurobiochemical. The information provided by each biomarker within these categories is highly variable and may be relevant for diagnosis, response prediction, target selection, program adjustment, etc. Conclusions: A better understanding of biomarkers and their applications would allow DBS in psychosurgery to advance on a more objective basis, guided by the information provided by them and within the context of precision psychiatry.

… That is, this perspective positions AI as the core engine for personalized neuromodulation to target the dysfunctional neural circuitry of MDD. We first outline three synergistic application …