Background: Transcranial Magnetic Stimulation (TMS) is an effective treatment for Major Depressive Disorder; however, meta-analyses report greater clinical improvement in females than males. One hypothesis is that anatomical differences between sexes create variations in the "dose" of electric field (e-field) reaching the cortex, potentially influencing treatment outcomes. Methods: Structural MRIs from 109 adults (72f) were analyzed using SimNIBS to model TMS-induced e-fields generated by a MagVenture figure-of-eight coil with standard tissue conductivities. E-fields were evaluated across hemispheres at the dorsolateral prefrontal cortex (EEG 10–20 sites F3/F4) and the motor cortex (group-level MNI functional hotspot). Average e-field was quantified across two methods: the 99th-percentile of voxels and within a 30-mm radius sphere. Results: A generalized estimating equations model for the DLPFC revealed a significant interaction between measure and sex (p = .001). Post-hoc estimated marginal means showed that females had greater average e-field values than males via the 99th-percentile measure (p = .035), but no sex differences were observed with the 30-mm measure (p = .893). These effects likely reflect method-dependent voxel count differences; a parallel statistical approach revealed a significant measure × sex interaction for voxel count, with males having greater voxel counts than females in the 99th-percentile mask (p < .001) but not in the 30-mm sphere (p = .19). Conclusion: Both sexes receive comparable cortical stimulation from TMS at the DLPFC and motor cortex. Apparent sex differences in e-field "dose" may arise from sex-related differences in cortical volume, underscoring the importance of e-field quantification methods and highlighting the need to explore other contributors to sex-related variability in TMS treatment response.

Background: Although trauma-focused psychotherapy remains the first-line treatment for PTSD, 14% to 35% of individuals fail to achieve symptom remission. Transcranial magnetic stimulation (TMS) has shown promise as an adjunctive intervention, but its efficacy for PTSD remains inconsistent, partly due to the lack of symptom-specific neuromodulation targets. To address this, we developed a connectivity-based modeling framework to identify functional networks that predict symptom severity and to estimate where stimulation would most effectively reduce PTSD symptoms. Methods: We used clinical, structural, and resting state functional data from 350 veterans with and without PTSD. Principal component analysis (PCA) was applied to functional connectivity matrices to reduce dimensionality and resulting component scores were entered into a lasso regression to identify connections associated with symptom severity. To simulate TMS effects, we generated electric field models for 200 randomly distributed cortical sites per participant and used these distributions to estimate connectivity changes proportional to the modeled fields. The estimated changes were combined with PCA regression weights to predict stimulation induced symptom changes at each site and were mapped across the cortex to create a whole brain TMS targeting atlas. Results: The model identified two broad and opposing patterns. Stimulation expected to increase connectivity in the left prefrontal cortex were predicted to reduce symptoms, whereas stimulation expected to decrease connectivity in bilateral parietal and occipital regions were also predicted to reduce symptoms. Conclusions: These findings provide a framework for individualized TMS targeting in PTSD and generate potential therapeutic targets for future clinical trials.

Anxious misery is a transdiagnostic cluster of mood and anxiety symptoms with substantial comorbidity, spanning depression, anxiety, and trauma-related conditions. fMRI in this population has shown distributed dysfunction across default mode, frontoparietal, and salience networks. Transcranial Magnetic Stimulation (TMS) is a promising tool with clinical efficacy for several neuropsychiatric conditions, such as Depression and Obsessive-Compulsive Disorder. However, clinical outcomes vary, highlighting the need for personalized protocols, informed by patients' symptomology and network organization. Here we developed a computational pipeline to derive symptom-specific TMS targets from whole-brain resting-state functional connectivity and electrical-field (E-field) modeling. Clinical, structural, and resting state fMRI (rs-fMRI) data from 118 participants were pulled from the Dimensional Connectomics of Anxious Misery dataset, including healthy controls. Rs-fMRI was parcellated and principal components analysis was applied to pairwise connectivity matrices to reduce dimensionality. Components were entered into a lasso regression to identify which connectivity features predicted symptom severity, yielding sparse, symptom-specific connectivity weight maps. To simulate TMS we used SimNIBS to generate E-field models across 200 cortical sites per participant then projected these E-field maps onto symptom-weighted networks to predict post-stimulation symptom change, creating a whole-brain TMS targeting atlas. Preliminary results demonstrate dissociable patterns such as anxiety and rumination targets clustering anteriorly towards prefrontal regions, while depression and anhedonia targets favored frontoparietal cortex and more posterior regions. By mapping symptom dimensions onto cortical stimulation sites, our pipeline supports personalized TMS targeting protocols which account for network-level heterogeneity and propose potential therapeutic targets to be tested in future clinical trials.

Transcranial Magnetic Stimulation (TMS) is an effective non-invasive therapy for treating depression; however, clinical outcomes vary substantially, likely driven by individual brain differences. We aimed to reduce this variability by implementing a neural network-based brain decoder to enable adaptive, closed-loop stimulation optimization. Real-time decoder outputs were generated by analyzing neural patterns during a worry and rumination task performed between blocks of repetitive TMS (rTMS) administered inside the MRI to an individualized stimulation target. Optimal and suboptimal frequencies for each participant were determined using both real-time decoder readouts and emotion self-report obtained during the interleaved TMS/fMRI scan. Participants with anxiety/depression symptoms then completed a randomized crossover design consisting of three consecutive days of neuromodulation with either their optimal or suboptimal frequency, administered on separate weeks with a minimum two-week washout. Following each neuromodulation session, participants completed a perseverative thought task, providing continuous joystick ratings of thought valence and intensity across scenario types. Across participants, optimal neuromodulation was associated with a significantly greater shift in final joystick position, reflecting more positive end-of-task thoughts (N=11, p = 0.035). Further analyses revealed a positive correlation between the difference in decoder outputs (optimal vs. suboptimal) and the corresponding difference in change in final joystick position (N=9, R²=.53, p = 0.025), supporting the use of neuroimaging-based machine learning to optimize behavioral responses to neuromodulation through individualized TMS targeting and frequency selection. These findings demonstrate the feasibility of real-time brain decoding to enhance TMS effectiveness and highlight the potential for neuroimaging-guided, machine learning-based personalization of TMS treatment.

Background: Transcranial Magnetic Stimulation is an FDA-approved intervention for the treatment of Treatment-Resistant Depression. The exact mechanism(s) by which TMS triggers remission are unclear. These mechanisms could be present across multiple scales, i.e., cellular, microcircuit, and network levels. By understanding these mechanisms, we could possibly personalize the treatment course for each individual. In this poster, we reviewed the literature on biophysical models used to understand the effects of TMS. Methods: We searched PubMed, Embase, and Web of Science using "Transcranial Magnetic Stimulation" and "biophysical" terms. Results: We identified the following biophysical model concepts: (a) Electrical (E-field) and microcircuit models in the primary motor cortex, (b) models to understand the motor evoked potential, (c) biophysical models of TMS-induced synaptic plasticity, (d) network-based biophysical models, and (e) a simple neuronal model to explain theta burst stimulation. We will explain the benefits and limitations of each model in the poster.

Parkinson's disease (PD) is the world's second most common neurodegenerative disorder. The premature death of dopaminergic (DA) neurons leads to denervation of the striatum that receives dopamine as the modulatory neurotransmitter for direct and indirect basal ganglia pathways. To re-establish dopamine levels in the striatum, we are developing a novel neural engineering solution whereby custom-built dopaminergic Tissue Engineered Living Electrodes (dTELEs) could extend from the cortical surface to deep brain targets to serve as a biologically-active, axon-based "living DBS". The dTELEs consist of aggregated human induced pluripotent stem cell (iPSC)-derived DA neurons that are transduced to express channelrhodopsin (ChR2) with long axon tracts enclosed in a hydrogel. To simulate in vivo conditions of integration of DA axonal tracts with the striatal tissue, we added human iPSC derived striatal neurons as the end target. A detailed in vitro characterization was performed to optimize viability and functionality of the microtissue. Assessments included immunocytochemistry for dopaminergic markers (tyrosine hydroxylase, dopamine transporter, A9/A10 markers), and dopamine release to confirm functional synaptic transmission. The dTELEs were then stimulated using light to allow for optically controlled release of dopamine in the striatal compartment. Our results showed that over 40-50% of cells expressed key dopaminergic markers and extended long axonal projections. Further comparisons based on light- versus electrical-induced evoked dopamine release are currently ongoing. This study underscores the potential of a bioengineered functional microtissue that mimics the natural biological inputs of the lost nigrostriatal pathway and can be used as a promising cell-based therapeutic strategy for PD.

Numerous studies have administered non-invasive brain stimulation (NIBS) techniques, such as transcranial direct current stimulation (tDCS) or transcranial magnetic stimulation (TMS), at C3 (right hand) or C4 (left hand) in the international 10-20 system, particularly in situations where the use of TMS with electromyography (EMG) is not feasible, demonstrating physiological or behavioral effects in the contralateral hand area. However, due to the difficulty of delivering stimulation exclusively to the targeted region with these NIBS techniques, there remains a considerable possibility that stimulation outside the hand area indirectly affects the hand representation in the primary motor cortex (M1). Therefore, it is uncertain whether C3 and C4 more accurately reflect the cortical motor hand representation compared to other electrode locations. The purpose of this study was to verify, through electric field (e-field) modeling, a concern previously raised in the literature that the C3 region may not be the optimal target for the motor hand area (Kim et al., 2023; Silva et al., 2021). Based on previous studies, we hypothesized that stimulation at C3 would not yield the highest mean e-field at the M1 compared to other electrode locations. In this study, a total of 144 T1-weighted MRI scans obtained from OpenNeuro, McLean Hospital, Brigham and Women's Hospital, and Butler Hospital were analyzed. The mean e-field was computed at two MNI coordinates considered to represent the motor hand area, [-37, -21, 58] (Mayka et al., 2006) and [-36, -24, 56] (Hardwick et al., 2013), using SimNIBS 4.1.0. For tDCS modeling, the anode was placed at one of seven locations (FC1, FC3, C1, C3h, C3, CP1, or CP3), while the cathode was fixed at Fp2, following the most commonly used montage (i.e., anode: C3, cathode: Fp2) for motor learning studies (Buch et al., 2017). The stimulation intensity was set to 1 mA. For TMS modeling, simulations were performed with the coil centered at FC1, FC3, C1, C3h, C3, CP1, or CP3 and angled 45 degrees towards the midline. TMS modeling was based on the Magstim D70 coil with stimulation intensity was set to 50% of maximum stimulator output. In the tDCS modeling, higher e-fields were calculated at CP3 and CP1 compared to C3 at both MNI coordinates. In the TMS modeling, higher e-fields were observed at C3h than at C3 at both MNI coordinates. These results suggest that when it is difficult to determine the location that elicits the largest motor-evoked potential (i.e., hotspot) using a TMS and EMG system, stimulating C3 by default may not be the best approach, and better alternatives may exist. Future studies should verify whether commonly used EEG-based targets, such as F3 for depression treatment, most effectively reach their intended cortical targets and subsequently yield optimal therapeutic efficacy.