White Matter Electric Field Maximization Guided Coil Orientation Optimization for rTMS in Alzheimer's Disease

White Matter Electric Field Maximization Guided Coil Orientation Optimization for rTMS in Alzheimer's Disease | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article White Matter Electric Field Maximization Guided Coil Orientation Optimization for rTMS in Alzheimer's Disease Nianshuang Wu, Ziyan Zhu, Zhen Wu, Shuxiang Zhu, Penghao Wang, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7319128/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Repetitive transcranial magnetic stimulation (rTMS) has emerged as a promising intervention for Alzheimer's disease (AD), yet current protocols lack standardized methodologies for correlating coil orientation with electric field (E-field) characteristics. This study establishes coil orientation optimization protocols for AD-related rTMS targets by analyzing E-field distribution patterns. Methods In a study population of 45 AD patients undergoing targeted stimulation of either the dorsolateral prefrontal cortex (DLPFC, n = 30, M:F, 11:19) or angular gyrus (AG, n = 15, M:F, 8:7), we performed E-field simulations across four critical regions: DLPFC, AG, precuneus (PC), and primary motor cortex (M1), determining optimal orientation through region-specific E-field maximization. Post hoc analysis of neuropsychological outcomes validated the proposed optimization strategy. Results Our findings revealed that white matter E-field (E WM ) and normal E-field component (E ⊥ ) demonstrated higher orientation-dependent variability compared to gray matter E-field (E GM ), despite larger GM volumes. Orientation optimization achieved more than 85% consistency rates for M1 (45°), AG (-15°) and PC (0°) through E WM maximization, whereas DLPFC required individualized coil orientation. Clinical validation demonstrated that AG-targeted stimulation maintained significantly lower orientation deviation variance compared to DLPFC protocols, corresponding to reduced coefficient of variation (CV) in the improvement of MoCA scores (55.65% vs. 100.76%, P = 0.006). Notably, patients aligned with E WM -optimized orientation in DLPFC group showed superior improvement of auditory verbal learning test (AVLT) scores compared to non-optimized cases ( P < 0.05). Conclusions These findings establish E WM as a critical determinant of rTMS efficacy and advocate for region-specific orientation protocols to enhance AD treatment outcomes. Alzheimer's disease Transcranial magnetic stimulation Electric field White matter Coil orientations Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Background Repetitive transcranial magnetic stimulation (rTMS) has been widely used to treat neuropsychiatric disorders, particularly showing significant potential in the clinical management of Alzheimer's disease (AD)[1, 2]. However, current research has not integrated the anatomical features of AD with electric field (E-field) components induced by transcranial magnetic stimulation (TMS), an approach that could optimize stimulation parameters and enhance treatment efficacy. Mounting evidence establishes the E-field magnitude as a critical mediator of neuromodulation outcomes across diverse brain stimulation modalities. In rTMS protocols, cortical E-field magnitude was correlated with depressive symptom relief, while individualized auditory-verbal hallucination network E-field strength predicts therapeutic efficacy in schizophrenia interventions[3–5]. Parallel findings in electroconvulsive therapy reveal hippocampal E-field intensity as a determinant of executive function changes, and transcranial alternating current stimulation (tACS) studies demonstrate age-dependent correlations between modeled E-field strength and cognitive performance[6, 7]. Crucially, the vector decomposition of TMS-generated E-fields into normal (perpendicular to cortical surface) and tangential components reveals distinct neurophysiological effects. Clinical data from treatment-resistant depression indicate that normal component magnitude, not tangential, mediates antidepressant responses during bilateral sequential rTMS, underscoring the therapeutic relevance of field orientation analysis[8]. Traditional E-field analyses have predominantly focused on gray matter (GM) territories, given their high neuronal soma density. However, emerging connectomic evidence positions white matter (WM) integrity as equally critical for maintaining functional network dynamics impaired in AD[9, 10]. The structural decoupling between GM nodes and WM edges, exacerbated by AD-related axonal degeneration, necessitates dual consideration of both tissue compartments in field modeling. Advanced multimodal approaches integrating diffusion tractography with E-field modeling demonstrate enhanced spatial specificity when accounting for WM anisotropy[11], while image-guided TMS paradigms show potential for transcending conventional depth limitations through network-targeted stimulation[12]. These developments highlight the imperative for combined GM-WM E-field analyses in AD therapeutic optimization. The biophysical foundation of TMS-induced E-fields comprises rotational (directly magnetic flux-derived) and irrotational (surface charge-mediated) components, computationally separable via contemporary simulation frameworks[13, 14]. For the irrotational E-field component, while existing models typically incorporate standard conductivity profiles, they frequently neglect AD-specific neuroanatomical perturbations. Progressive cerebral atrophy, a hallmark AD pathology exhibiting marked interindividual variability[15], fundamentally alters the cortical geometry of key rTMS targets including dorsolateral prefrontal cortex (DLPFC), angular gyrus (AG), and precuneus (PC)[16–18]. Concurrently, amyloid-β/tau-mediated WM degeneration disrupts structural connectivity through myelin oxidative damage and neuroinflammation[19–21], creating complex interactions between pathological tissue remodeling and E-field distribution. Though direct evidence remains sparse, transcranial direct current stimulation (tDCS) studies confirm cranial anatomy's significant impact on E-field patterns[22], while finite element analyses reveal tissue-specific E-field gradients dependent on coil positioning and WM anisotropy[23]. Therefore, further research is needed to explore the impact of brain atrophy on TMS induced E-fields in different stimulation targets. For the rotational E-field component, while neuronavigation systems achieve millimeter-level spatial precision in target localization[24, 25], rotational freedom around the coil-to-target axis introduces substantial angular variability[26]. Current clinical protocols often arbitrarily adopt 45° medial-sagittal angles[27, 28] or midline-parallel orientations[29, 30] without empirical justification, a stark contrast to M1 stimulation where motor evoked potentials (MEPs) enable real-time optimization guided by cortical column cosine models[31–34]. The absence of analogous physiological feedback in cognitive targets necessitates computational approaches to orientation optimization. Though engineering studies propose generalized workflows incorporating finite element analysis and deep learning[24, 35–38], their clinical translation remains impeded by technical complexity. This disparity between motor and cognitive stimulation paradigms creates critical knowledge gaps in rTMS parameter selection for AD. In this study, we hypothesize that orientation-dependent maxima in critical E-field components serve as robust predictors of enhanced therapeutic efficacy. Initially, we quantified the interindividual variability in GM and WM volumetric distributions and characterized their impact on E-field component. Then we evaluated orientation-dependent E-field modulation profiles to identify critical E-field components for optimization. This enabled the derivation of region-specific optimal coil orientations through E-field maximization across regions of interest (ROIs). The clinical relevance and therapeutic potential of these optimized orientations were subsequently validated through two clinical trials. Collectively, this work establishes a novel region-specific coil orientation optimization strategy to maximize the therapeutic potential of rTMS in AD treatment. 2. Methods 2.1 Patients and neuropsychological assessment We recruited 45 patients diagnosed with probable AD from the outpatient clinic service of Xuanwu Hospital, Capital Medical University, Beijing, China. All patients provided written informed consent prior to participation. Diagnoses for all patients were made by two trained senior neurologists following a detailed consultation. In this study, 30 patients underwent rTMS targeting the left DLPFC from August 2022 to August 2023, while 15 patients received bilateral AG stimulation between November 2020 and September 2021. This study is registered in the Chinese Clinical Trial Registry ( https://www.chictr.org.cn/index.html ), number ChiCTR2200062564 and ChiCTR1900025045. The inclusion criteria for the DLPFC and AG groups, as well as the cognitive assessments, are detailed in eMethods in Supplement 1. 2.2 Image acquisition All T1-weighted data were collected by a Siemens 3.0 T MRI system (Siemens, Erlangen, Germany). The specific MRI parameters and scanning requirements are presented in eMethods in Supplement 2. 2.3 rTMS treatment The stimulation of the left DLPFC and bilateral AG was performed using a Magstim Rapid 2 transcranial magnetic stimulator (Magstim, Co. Ltd, UK), equipped with an air-cooled figure-of-8 coil (70 mm diameter). The patient's head was co-registered with structural MRI using the BrainSight TMS navigation system (Rogue Research, Montreal, QC) to accurately target the intracerebral markers. In the DLPFC group, the stimulation target was the left DLPFC, whereas in the AG group, it was the bilateral AG. The protocol for rTMS therapy is detailed in eMethods in Supplement 3. 2.4 E-field modeling 2.4.1 Coil positions and orientations Figure 1 illustrates the overall workflow of this study. Simulations were performed for each patient at four target regions: left M1[39], left DLPFC[40], left AG[41], and PC[18]. For each target, the coil was positioned tangentially to the skull, with the handle pointing posterolaterally at a 45° angle to the midsagittal plane. The junction point of the figure-8 coil was positioned at the scalp point closest to the intracerebral target[27]. This position and orientation were defined as the initial orientation (0°) for this study. The line connecting the coil center and the intracranial target was defined as the rotation axis. Starting from 0°, the coil was rotated counterclockwise around the rotation axis in 30° increments up to 150°. Due to the symmetry of the stimulation coil, E-field values for coil orientations from − 180° to -30° were calculated by mirroring the corresponding orientations from the first half of the rotation (0° to 150°). Insert Fig. 1 here. 2.4.2 Stimulation strength The stimulation strength is determined by the d i /d t value in SimNIBS software. For the M1 target, the d i /d t value is set to 1 A/µs, while for the DLPFC, AG, and PC targets, the d i /d t values are set to 0.8 A/µs, corresponding to 80% rMT. M1 E-field strength (E M1 ) of each patient was calculated as the top 99.9% of E-field magnitude from all voxels, serving as an estimate of the peak-induced field strength. Other simulation parameters were set to the default values. 2.4.3 E-field analyses Stimulation targets, defined in Montreal Neurological Institute (MNI) coordinates, were mapped to individual subject spaces and projected onto the nearest gyral crown surfaces relative to the stimulation coil. The gyral crown surface point was designated as the center of the region of interest (ROI), with a 10-mm radius. Within each ROI, the number of GM and WM voxels was calculated. The MRI T1 data processing is detailed in eMethods in Supplement 4. The E-fields in each voxel within the ROI were divided by E M1 at the initial orientation to normalize the E-field strength. Then E-field strength in ROI (E ROI ) of each patient was calculated by averaging E-field strength in all voxels within ROI. Moreover, we divided the voxels in ROI into GM and WM in the simulation model. The GM and WM E-fields (E GM and E WM ) were also calculated by averaging E-fields in their respective voxels. In addition, the E-field can be divided into two components: the normal component (E ⊥ ), which is perpendicular to the cortical surface, and the tangential component (E ∥ ), which runs parallel to it. 2.5 Statistical analyses Differences between GM and WM voxels were assessed using T-tests. The values of GM and WM voxels, along with the E-fields, were standardized by dividing the data by the mean of the all patients. Data variability was expressed as the coefficient of variation (CV), and F Test and Levene's Test is used to evaluate variance homogeneity. The changes in clinical score (Δscore) were calculated as the difference between the post-treatment and pre-treatment scores. T-tests and Wilcoxon tests were performed on clinical scales before and after rTMS treatment to assess differences. An ANOVA was performed to assess whether statistically significant differences existed in E-field values across various coil orientations. Frequency statistics were used to quantify the occurrence of maximum E-field values at various coil orientations. The coil orientation that corresponded to the maximum values of E ⊥ and E WM for each patient was defined as the patient's optimal coil orientation (pOCO) for E ⊥ and E WM , respectively. A Chi-square test was applied to examine the frequency of occurrence of pOCO across different coil orientations for the 45 patients. The coil orientations with the highest frequency of occurrence were defined as the modal optimal coil orientation (mOCO). Pearson correlation was employed to assess the relationship between E-field components and clinical outcomes. The coil orientation deviation for each patient group was assessed by calculating the absolute difference between each patient's optimal and actual treatment orientations, with the standard deviation (SD) representing the deviation. T-test performed to compare the coil orientation deviations between the two groups. The patients in the DLPFC group were divided into four subgroups: Subgroup-1 and Subgroup-3 represented the optimal coil orientation groups (treatment orientation matched the pOCO) for the E ⊥ and E WM , respectively, while Subgroup-2 and Subgroup-4 represented the non-optimal coil orientation groups (treatment orientation did not match the pOCO). The impact of optimal coil positioning on clinical outcomes was analyzed using a T-test to determine whether precise coil placement enhances treatment efficacy. Statistical significance for all analyses was set at a two-sided P value of less than 0.05. Statistical analyses were performed using SPSS Statistics V.26.0 (SPSS) and MATLAB (v.8.5, R2015a). 3. Results 3.1 Interindividual variability in GM/WM volumes and E-field components We calculated the variability of WM and GM voxels within each ROI, as well as the E-field components at the initial coil orientation, as shown in Fig. 2 . The number of GM voxels significantly exceeded that of WM voxels in all ROIs (all P < 0.001), with detailed voxel counts and CVs for four brain regions provided in Table S1 . To facilitate comparison of the CVs of GM and WM voxel counts across different brain regions, the normalized results are presented in Fig. 2 A. Quantitative analysis revealed significantly higher CVs in WM than GM across all target regions (all P < 0.001). The TMS-induced E-fields exhibited distinct region-specific distribution patterns, with spatial heterogeneity observed in the E-field components within each ROI (Table S2). Normalized comparisons of E-field component CVs across different ROIs are systematically illustrated in Fig. 2 B-E, with the intracerebral E-field distribution of patient1 shown as a representative example. Cortical surface E-field component analysis revealed no significant difference in CVs between E ⊥ and E ∥ , except M1 (all P > 0.05). Voxel-based analysis of E-field components within ROIs revealed that E WM had significantly greater variability than E GM across all regions except M1. Notably, correlation analysis demonstrated a significant association between the E-field and regional brain volume only in the DLPFC (all P 0.05), as shown in Fig. 2 F-H. Insert Fig. 2 here 3.2 Orientation-dependent modulation of E-field components We further examine how coil orientation variations impact E-fields. For each patient, we calculated the E-field components at various coil orientations within each ROI. Figure 3 shows the variability of E-field components across coil orientations. Representative data from patient 1 (Fig. 3 A-H) demonstrate characteristic elliptical modulation patterns for all E-field components, where the major and minor axes of the ellipses correspond to the orientation-dependent maximum and minimum E-field magnitudes, respectively. Notably, the coil orientations yielding peak values for E ROI , E GM and E WM showed strong spatial consistency, with angular deviations limited to less than 30° across these parameters. The E-field components (mean ± SD) and their CVs for 45 patients across six coil orientations are shown in Table S3. Cortical surface E-field components analysis revealed significantly higher CVs for E ⊥ compared to E ∥ in all regions except AG (all P < 0.05; Fig. 3 I). Given the orthogonal maxima of E ⊥ and E ∥ across ROIs, coupled with CV difference less than 0.5% for both components in the AG, we prioritized E ⊥ for orientation-specific optimization. Voxel-based analysis of E-field components within ROIs revealed that E WM exhibited significantly higher CV than E GM in all regions except PC (all P < 0.05; Fig. 3 J). Intriguingly, while the PC exhibited maximal CV in GM, we still chose E WM for voxel-based optimization to maintain consistency in parameter selection across regions. Insert Fig. 3 here 3.3 Optimal coil orientation determination To determine the mOCO, E ⊥ and E WM were selected as the key E-field components on the cortical surface and within the voxels of the ROI, respectively. We analyzed the distribution of maximum E ⊥ and E WM across six coil orientations for each patient, as well as the frequency of occurrence of these maximum values as shown in Fig. 4 . Figure 4 A-H show the distributions of E ⊥ and E WM at different coil orientations for all patients. ANOVA tests revealed significant variation in E ⊥ across all coil orientations (all P < 0.001), except for DLPFC ( P = 0.69), while E WM showed significant orientation dependence in all regions ( P < 0.05). These findings confirm orientation-specific E-field variations across brain regions, with DLPFC exhibiting unique insensitivity to E ⊥ modulation. Figure 4 I and J, along with Table S4, show the occurrence of frequency of pOCO for E ⊥ and E WM across different coil orientations. In the M1 region, significant differences were found in the occurrence of pOCO for E ⊥ and E WM (E ⊥ : P < 0.001; E WM : P < 0.001). For both E ⊥ and E WM , about 50% of patients achieved maxima at 0°, with 86.7% of maxima clustered within − 30°~30°. This convergence motivated defining M1 mOCO as -30°~30°. In the DLPFC region, E ⊥ maxima exhibited non-significant dispersion across orientations ( P = 0.12), whereas E WM demonstrated significant orientation preference ( P = 0.008), peaking at -30° (33.3% occurrence) with 84.4% of maxima distributed between − 60°~30°. The discordant E ⊥ profiles and broad E WM distribution precluded definitive mOCO determination. In the AG region, both E ⊥ and E WM showed strong orientation tuning: 88.9% of E ⊥ maxima concentrated at 30°~60° ( P < 0.001), while 86.7% of E WM maxima spanned 0°~60° ( P < 0.001). This partial overlap justified defining AG mOCO as 0°~60°. In the PC region, significant orientation effects emerged for both E-field components (E ⊥ : P < 0.001; E WM : P = 0.001). Remarkably, 88.9% of E ⊥ and 91.1% of E WM maxima co-localized at 30°~60°, establishing 30°~60° as the PC mOCO (The E GM optimization result remained consistent, as detailed in eResults, Supplement 1). A comprehensive summary of regional mOCO ranges is provided in Table 1 . Table 1 Optimal coil orientation distribution Brain regions E ⊥ range a E WM range a E ⊥ range b E WM range b M1 0° (-30°~30°) 0° (-30°~30°) 45° (15°~75°) 45° (15°~75°) DLPFC \ \ \ \ AG 45° (30°~60°) 30° (0°~60°) 0° (-15°~15°) -15° (-45°~15°) PC 45° (30°~60°) 45° (30°~60°) 0° (-15°~15°) 0° (-15°~15°) Note: a. The optimal orientation range refers to counterclockwise rotation angle from the initial coil handle orientation, which points posterolaterally at a 45-degree angle to the midsagittal plane. b. The optimal orientation range refers to the angle between the coil handle, which points posterolaterally, and the midsagittal plane. Insert Fig. 4 here 3.4 Clinical correlates of E-field optimization 3.4.1 Coil orientation governs neuromodulation outcome uniformity Patient demographics are detailed in eResults in Supplement 2. Despite the uniform 0° coil orientation applied to all patients in the clinical trials, orientation discrepancies were observed between the actual coil orientations and the computationally derived optimal orientations in both the DLPFC and AG groups. The DLPFC group demonstrated pronounced variability of orientation discrepancies under both E ⊥ (42°±29.05°) and E WM (42°±25.65°) paradigms, whereas AG group exhibited markedly reduced deviations for E WM (22°±21.11°, P = 0.013) despite comparable fluctuations (46°±19.20°, P = 0.63) for E ⊥ . This refined directional specificity in AG group corresponded to superior clinical stabilization, manifesting as 39.19% reduced CV in improvement of MoCA scores relative to DLPFC group (101.03% vs 166.14%, P = 0.006, Fig. S1 ), with more detailed information provided in Table S5. The robust association between optimized field orientation fidelity and reduced outcome variability underscores precise coil orientation as a critical determinant of response predictability in rTMS targeting distinct cerebral regions. 3.4.2 E-field strength and clinical outcomes Given the absence of definitive mOCO in DLPFC, we investigated E-field strength correlations with cognitive changes. Significant associations emerged between specific E-field components and neuropsychological task performance in the DLPFC group (all P 0.05). Complete correlation matrices are provided in Table S6. 3.4.3 Relationship between individualized optimal coil orientation and clinical outcomes Subgroup stratification by E WM -optimized and E ⊥ -optimized coil orientation revealed critical orientation-dependent cognitive effects in DLPFC neuromodulation. Patients in Subgroup-3 demonstrated significantly greater improvements in AVLT3 scores (2.29 ± 2.64 vs. 0.69 ± 0.75, P = 0.043) compared to Subgroup-4. Although the mean of AVLT_all score in Subgroup-3 (7.00 ± 7.04) was greater than that in subgroup-4 (3.08 ± 3.75), this difference did not reach statistical significance ( P > 0.05). Patients in Subgroup-1 demonstrated greater improvements in AVLT_all (6.50 ± 7.21 vs. 3.93 ± 4.34) and AVLT3 (2.19 ± 2.69 vs. 0.93 ± 1.14) compared to Subgroup-2, however, these differences not statistically significant (all P > 0.05). The aggregate cognitive profile, as visualized in Fig. 5 E-F, further supports the selective efficacy of E WM -guided orientation. Insert Fig. 5 here 4. Discussion This study establishes a neurobiologically grounded framework for optimizing TMS coil orientation in AD, demonstrating that target-specific maximization E WM significantly enhances therapeutic outcomes. Our findings extend prior motor cortex research to cognitive networks through three key advances: (1) identification of WM anatomical variations as primary contributors to TMS-induced E-field variability, (2) demonstration of superior angular modulation sensitivity of E WM compared to that of E GM , and (3) validation that precise alignment between computational optimization and clinical coil orientation reduces outcome variability while enhancing efficacy. These insights provide critical theoretical support for personalized neuromodulation in AD. Unlike conventional E ROI /E ⊥ maximization approaches[24, 35], our E WM -targeted optimization considers AD-specific neuropathology. While prior coil orientation studies predominantly focused on cortical geometry’s influence on E ROI and E ⊥ [8, 26, 35, 42–45], emerging computational evidence underscores the critical role of orientation-dependent brain network engagement[46, 47]. These findings collectively emphasize the necessity of addressing individual WM variability for optimizing cognitive modulation and therapeutic outcomes[10]. By integrating E WM quantification with E ROI /E ⊥ metrics within ROIs, our approach fills a crucial gap in TMS research, offering a more comprehensive framework. While the methodological differences between E WM optimization and conventional E ⊥ maximization seem minor, their biophysical implications are fundamentally different. The E WM averages the E-field across WM voxels within ROIs, whereas E ⊥ is limited to CSF-GM interfaces, making E WM a more comprehensive biomarker for axonal activation patterns rather than just cortical effects. Notably, our study focuses on atrophic AD brains, which differ from studies on healthy or depressed individuals[3, 26, 36]. These AD brains exhibit significant GM/WM volumetric alterations, with GM predominance in ROI[48], coupled with substantial inter-individual anatomical variability in cortical thickness and sulcal geometry[49]. These factors collectively enhance E WM fluctuations. Remarkably, despite WM being deeper, where conventional distance decay principles would predict attenuated E-field intensity[50], E WM consistently demonstrated higher field strength than E GM across all ROIs. This apparent paradox can be explained by tissue conductivity gradients, where under quasi-static conditions governed by ∇·J = 0 and J = σE, the lower conductivity (σ) of WM results in elevated E-field magnitude at GM-WM interfaces relative to adjacent GM regions[51]. Although external stimulation parameters (e.g., intensity, frequency, and pulse count) are typically guided by protocols, the selection of coil orientation has received less attention. We resolve this gap by combining computational modeling with data from two independent clinical trials, ensuring comprehensive validation. Our motor cortex findings demonstrate that the maximum E WM and E ⊥ in M1 are optimally clustered within − 30°~30° of the central sulcus perpendicular, aligning with both established neurophysiological principles of MEP optimization[26, 52] and clinical evidence showing that the posteroanterior coil orientation, perpendicular to the central sulcus, is effective for rTMS pain therapy, while the lateromedial orientation is not[53]. Crucially, we extend our coil orientation optimization principle to cognitive relevant stimulation targets, i.e. DLPFC, AG, and PC, where no feedback signal similar to MEP can be used to help optimize coil orientation. The analysis of angular deviations between actual and optimal coil orientations revealed group differences based on optimization criteria, with E ⊥ maximization showed no significant variation between AG and DLPFC groups, whereas E WM maximization demonstrated reduced variability in the AG group. These findings support the reliability of E WM -based optimization. Consistent alignment between the actual and optimal coil orientations reduces variability in treatment outcomes and enhances efficacy. Our E WM maximization premise aligns with established E-field/efficacy correlations[3–5, 8], supported by significant relationships between E-field and clinical outcome. These insights provide critical theoretical support for coil orientation optimization in AD. Although our study does not include clinical trials targeting PC, recent studies have demonstrated promising therapeutic outcomes by stimulating the PC with a midline-parallel coil orientation, inducing posterior-anterior currents[18]. Notably, this coil orientation aligns with the mOCO identified in our computational modeling study. Contrary to prevailing individualization trends[24, 36], we demonstrate non-personalized approaches suffice for AG, M1, and PC stimulation, significantly streamlining clinical workflows. For DLPFC, despite significant E WM variation across orientations ( P = 0.008) with peak frequencies at -60° (26.67%), -30° (33.33%), 0° (11.11%), and 30° (13.33%), no universal mOCO exists within the 90° range, necessitating individualization. The 30° sampling interval balances practicality with field stability, because smaller increments (10°) show negligible field changes[26]. In addition, our left-hemisphere optimizations can be mirrored for right-sided targets. TMS targets specific regions in the brain, which are part of different brain networks[54]. Though the exact mechanisms behind rTMS's therapeutic effects are still unclear, the DLPFC is thought to be a node in a brain network linked to cognitive impairment in AD[55]. Ongoing research aims to optimize and personalize rTMS targets based on functional connectivity[56, 57]. Emerging connectomic evidence positions WM integrity as equally critical for maintaining functional network dynamics impaired in AD[9, 10]. Functional connectivity is established upon the basis of structural connectivity[58], which is primarily composed of WM fibers. Our findings indicate that targeting the DLPFC necessitates personalized optimization, which may underlie the differential activation of neural networks through distinct coil orientations. These results establish a framework for future investigations to systematically examine how E WM modulate TMS-induced network effects. Our study has several limitations. First, we calculated E-fields only from 0° to 150° due to mirror symmetry beyond 150° (180°-330°). This approach is practical for studying intracerebral E-field distribution, but previous studies on M1 have shown that rotating the TMS coil by 180° can alter motor response, including amplitude and latency[59, 60]. This is likely because different neural populations are recruited depending on the direction of the E-field[61, 62]. Second, the "mri2mesh" command in SimNIBS was used for image segmentation based on templates from healthy individuals. Future research should explore differences between segmented structures and actual anatomy of patients with brain atrophy. Third, the accuracy of conductivity values affects the simulation results, so more work is needed to optimize these values. Since WM is anisotropic, future studies should include DTI-based conductivity models to improve the accuracy of the simulations. Moreover, current methods don’t consider the low-conductivity Aβ and tau proteins, which could impact the accuracy of the simulations. Fourth, increasing the sample size is important for validating and generalizing findings. Finally, the AG group received both rehabilitation training and rTMS, and without a control group, it’s hard to separate the effects of rehabilitation training on clinical improvements. Although this study focuses on AD, the optimized coil orientations could also be applied to rTMS therapy for other neuropsychiatric disorders with some adjustments. 5 Conclusion In conclusion, this study highlights the importance of optimizing coil orientation in TMS for AD. We identified optimal coil orientations with the coil handle pointing posterolaterally at a -15° and 0° angle to the midsagittal plane for AG and PC respectively, while the DLPFC requires a more individualized approach. We also demonstrated E WM as a reliable indicator for determining the optimal coil orientations. These findings provide strong theoretical support for coil orientation optimization in rTMS protocols for AD patients. Declarations Ethics approval and consent to participate The study was approved by the Ethics Committee of Xuanwu Hospital of Capital Medical University, and written informed consent was obtained from each participant. All participants were informed of the study’s objectives and procedures. Consent for publication Consent for the publication of these data was not obtained from the participants. Data availability The datasets used or analyzed during the current study are available from the corresponding author on reasonable request. Acknowledgements The authors thank all the participants who participated in this study. Funding This research was funded by several projects. The National Key R&D Program of China supported this work with the project numbers 2022YFC2402200. The National Natural Science Foundation of China provided support through the project numbers 51977205, 82471498, and NSFC - AF 82211530041. Additionally, the Capital's Funds for Health Improvement and Research also contributed to this work with the project number CFH2022 - 2 - 2014. Neither of the mentioned funders was involved in the conception and performance of the study. Author contributions Nianshuang Wu: Conceptualization, Investigation, Methodology, Writing – original draft, review and editing, Supervision and Visualization. Ziyan Zhu: Data curation, Investigation, Methodology, Supervision and Visualization. Zhen Wu: Conceptualization, Investigation, Methodology, Supervision and Visualization. Shuxiang Zhu: Data curation. Penghao Wang: Conceptualization. Yuxuan Shao: Data curation. Cheng Zhang: Conceptualization. Changzhe Wu: Conceptualization. Xiaolin Huo: Conceptualization. Hua Lin: Conceptualization, Data curation, Investigation , Methodology, Supervision and Visualization, Project administration and Funding acquisition. Guanghao Zhang: Conceptualization, Investigation, Methodology, Writing – original draft, review and editing, Supervision, Visualization, Project administration and Funding acquisition. Competing interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References Li X, Qi G, Yu C, Lian G, Zheng H, Wu S, et al. Cortical plasticity is correlated with cognitive improvement in Alzheimer's disease patients after rTMS treatment. Brain Stimul. 2021;14(3):503-10. Sabbagh M, Sadowsky C, Tousi B, Agronin ME, Alva G, Armon C, et al. Effects of a combined transcranial magnetic stimulation (TMS) and cognitive training intervention in patients with Alzheimer's disease. Alzheimers Dement. 2020;16(4):641-50. Deng ZD, Robins PL, Dannhauer M, Haugen LM, Port JD, Croarkin PE. Optimizing TMS Coil Placement Approaches for Targeting the Dorsolateral Prefrontal Cortex in Depressed Adolescents: An Electric Field Modeling Study. Biomedicines. 2023;11(8). Quinn DK, Upston J, Jones TR, Gibson BC, Olmstead TA, Yang J, et al. Electric field distribution predicts efficacy of accelerated intermittent theta burst stimulation for late-life depression. Front Psychiatry. 2023;14:1215093. Hua Q, Wang L, He K, Sun J, Xu W, Zhang L, et al. Repetitive Transcranial Magnetic Stimulation for Auditory Verbal Hallucinations in Schizophrenia: A Randomized Clinical Trial. JAMA Netw Open. 2024;7(11):e2444215. Deng ZD, Argyelan M, Miller J, Quinn DK, Lloyd M, Jones TR, et al. Electroconvulsive therapy, electric field, neuroplasticity, and clinical outcomes. Mol Psychiatry. 2022;27(3):1676-82. Zanto TP, Jones KT, Ostrand AE, Hsu WY, Campusano R, Gazzaley A. Individual differences in neuroanatomy and neurophysiology predict effects of transcranial alternating current stimulation. Brain Stimul. 2021;14(5):1317-29. Zhang BBB, Stöhrmann P, Godbersen GM, Unterholzner J, Kasper S, Kranz GS, et al. Normal component of TMS-induced electric field is correlated with depressive symptom relief in treatment-resistant depression. Brain Stimul. 2022;15(5):1318-20. Qin T, Wang L, Xu H, Liu C, Shao Y, Li F, et al. rTMS concurrent with cognitive training rewires AD brain by enhancing GM-WM functional connectivity: a preliminary study. Cereb Cortex. 2024;34(1). Martín-Signes M, Rodríguez-San Esteban P, Narganes-Pineda C, Caracuel A, Mata JL, Martín-Arévalo E, et al. The role of white matter variability in TMS neuromodulatory effects. Brain Stimul. 2024;17(6):1265-76. Nummenmaa A, McNab JA, Savadjiev P, Okada Y, Hämäläinen MS, Wang R, et al. Targeting of white matter tracts with transcranial magnetic stimulation. Brain Stimul. 2014;7(1):80-4. Luber B, Davis SW, Deng ZD, Murphy D, Martella A, Peterchev AV, et al. Using diffusion tensor imaging to effectively target TMS to deep brain structures. Neuroimage. 2022;249:118863. Tofts PS, Branston NM. The measurement of electric field, and the influence of surface charge, in magnetic stimulation. Electroencephalogr Clin Neurophysiol. 1991;81(3):238-9. Thielscher A, Antunes A, Saturnino GB. Field modeling for transcranial magnetic stimulation: A useful tool to understand the physiological effects of TMS? Annu Int Conf IEEE Eng Med Biol Soc. 2015;2015:222-5. Marshall GA, Lorius N, Locascio JJ, Hyman BT, Rentz DM, Johnson KA, et al. Regional cortical thinning and cerebrospinal biomarkers predict worsening daily functioning across the Alzheimer's disease spectrum. J Alzheimers Dis. 2014;41(3):719-28. Chen H, Li M, Qin Z, Yang Z, Lv T, Yao W, et al. Functional network connectivity patterns predicting the efficacy of repetitive transcranial magnetic stimulation in the spectrum of Alzheimer's disease. Eur Radiol Exp. 2023;7(1):63. Liu C, Han T, Xu Z, Liu J, Zhang M, Du J, et al. Modulating Gamma Oscillations Promotes Brain Connectivity to Improve Cognitive Impairment. Cereb Cortex. 2022;32(12):2644-56. Koch G, Casula EP, Bonnì S, Borghi I, Assogna M, Minei M, et al. Precuneus magnetic stimulation for Alzheimer's disease: a randomized, sham-controlled trial. Brain. 2022;145(11):3776-86. Pereira JB, Janelidze S, Ossenkoppele R, Kvartsberg H, Brinkmalm A, Mattsson-Carlgren N, et al. Untangling the association of amyloid-β and tau with synaptic and axonal loss in Alzheimer's disease. Brain. 2021;144(1):310-24. Huang Z, Jordan JD, Zhang Q. Myelin Pathology in Alzheimer's Disease: Potential Therapeutic Opportunities. Aging Dis. 2024;15(2):698-713. Bartzokis G, Cummings JL, Sultzer D, Henderson VW, Nuechterlein KH, Mintz J. White matter structural integrity in healthy aging adults and patients with Alzheimer disease: a magnetic resonance imaging study. Arch Neurol. 2003;60(3):393-8. Laakso I, Tanaka S, Mikkonen M, Koyama S, Sadato N, Hirata A. Electric fields of motor and frontal tDCS in a standard brain space: A computer simulation study. Neuroimage. 2016;137:140-51. Opitz A, Windhoff M, Heidemann RM, Turner R, Thielscher A. How the brain tissue shapes the electric field induced by transcranial magnetic stimulation. Neuroimage. 2011;58(3):849-59. Balderston NL, Roberts C, Beydler EM, Deng ZD, Radman T, Luber B, et al. A generalized workflow for conducting electric field-optimized, fMRI-guided, transcranial magnetic stimulation. Nat Protoc. 2020;15(11):3595-614. Weise K, Numssen O, Kalloch B, Zier AL, Thielscher A, Haueisen J, et al. Precise motor mapping with transcranial magnetic stimulation. Nat Protoc. 2023;18(2):293-318. Janssen AM, Oostendorp TF, Stegeman DF. The coil orientation dependency of the electric field induced by TMS for M1 and other brain areas. J Neuroeng Rehabil. 2015;12:47. Wu X, Ji GJ, Geng Z, Wang L, Yan Y, Wu Y, et al. Accelerated intermittent theta-burst stimulation broadly ameliorates symptoms and cognition in Alzheimer's disease: A randomized controlled trial. Brain Stimul. 2022;15(1):35-45. Bagattini C, Zanni M, Barocco F, Caffarra P, Brignani D, Miniussi C, et al. Enhancing cognitive training effects in Alzheimer's disease: rTMS as an add-on treatment. Brain Stimul. 2020;13(6):1655-64. Koch G, Casula EP, Bonni S, Borghi I, Assogna M, Minei M, et al. Precuneus magnetic stimulation for Alzheimer's disease: a randomized, sham-controlled trial. Brain. 2022;145(11):3776-86. Mencarelli L, Torso M, Borghi I, Assogna M, Pezzopane V, Bonni S, et al. Macro and micro structural preservation of grey matter integrity after 24 weeks of rTMS in Alzheimer's disease patients: a pilot study. Alzheimers Res Ther. 2024;16(1):152. Brasilneto JP, Cohen LG, Panizza M, Nilsson J, Roth BJ, Hallett M. Optimal Focal Transcranial Magnetic Activation of the Human Motor Cortex - Effects of Coil Orientation, Shape of the Induced Current Pulse, and Stimulus-Intensity. J Clin Neurophysiol. 1992;9(1):132-6. Mills KR, Boniface SJ, Schubert M. Magnetic brain stimulation with a double coil: the importance of coil orientation. Electroencephalogr Clin Neurophysiol. 1992;85(1):17-21. Bashir S, Perez JM, Horvath JC, Pascual-Leone A. Differentiation of motor cortical representation of hand muscles by navigated mapping of optimal TMS current directions in healthy subjects. J Clin Neurophysiol. 2013;30(4):390-5. Balslev D, Braet W, McAllister C, Miall RC. Inter-individual variability in optimal current direction for transcranial magnetic stimulation of the motor cortex. J Neurosci Methods. 2007;162(1-2):309-13. Gomez-Tames J, Hamasaka A, Laakso I, Hirata A, Ugawa Y. Atlas of optimal coil orientation and position for TMS: A computational study. Brain Stimul. 2018;11(4):839-48. Cerins A, Thomas EHX, Barbour T, Taylor JJ, Siddiqi SH, Trapp N, et al. A New Angle on Transcranial Magnetic Stimulation Coil Orientation: A Targeted Narrative Review. Biol Psychiatry Cogn Neurosci Neuroimaging. 2024;9(8):744-53. Dannhauer M, Gomez LJ, Robins PL, Wang D, Hasan NI, Thielscher A, et al. Electric Field Modeling in Personalizing Transcranial Magnetic Stimulation Interventions. Biol Psychiatry. 2024;95(6):494-501. Moser P, Reishofer G, Prückl R, Schaffelhofer S, Freigang S, Thumfart S, et al. Real-time estimation of the optimal coil placement in transcranial magnetic stimulation using multi-task deep learning. Sci Rep. 2024;14(1):19361. Numssen O, Zier AL, Thielscher A, Hartwigsen G, Knösche TR, Weise K. Efficient high-resolution TMS mapping of the human motor cortex by nonlinear regression. Neuroimage. 2021;245:118654. Mir-Moghtadaei A, Caballero R, Fried P, Fox MD, Lee K, Giacobbe P, et al. Concordance Between BeamF3 and MRI-neuronavigated Target Sites for Repetitive Transcranial Magnetic Stimulation of the Left Dorsolateral Prefrontal Cortex. Brain Stimul. 2015;8(5):965-73. Schroën JAM, Gunter TC, Numssen O, Kroczek LOH, Hartwigsen G, Friederici AD. Causal evidence for a coordinated temporal interplay within the language network. Proc Natl Acad Sci U S A. 2023;120(47):e2306279120. Thielscher A, Opitz A, Windhoff M. Impact of the gyral geometry on the electric field induced by transcranial magnetic stimulation. Neuroimage. 2011;54(1):234-43. Bijsterbosch JD, Barker AT, Lee KH, Woodruff PW. Where does transcranial magnetic stimulation (TMS) stimulate? Modelling of induced field maps for some common cortical and cerebellar targets. Med Biol Eng Comput. 2012;50(7):671-81. Laakso I, Hirata A, Ugawa Y. Effects of coil orientation on the electric field induced by TMS over the hand motor area. Phys Med Biol. 2014;59(1):203-18. Gomez LJ, Dannhauer M, Peterchev AV. Fast computational optimization of TMS coil placement for individualized electric field targeting. Neuroimage. 2021;228:117696. Opitz A, Fox MD, Craddock RC, Colcombe S, Milham MP. An integrated framework for targeting functional networks via transcranial magnetic stimulation. Neuroimage. 2016;127:86-96. Lynch CJ, Elbau IG, Ng TH, Wolk D, Zhu S, Ayaz A, et al. Automated optimization of TMS coil placement for personalized functional network engagement. Neuron. 2022;110(20):3263-77.e4. Ge Y, Grossman RI, Babb JS, Rabin ML, Mannon LJ, Kolson DL. Age-related total gray matter and white matter changes in normal adult brain. Part I: volumetric MR imaging analysis. AJNR Am J Neuroradiol. 2002;23(8):1327-33. Im K, Lee JM, Lyttelton O, Kim SH, Evans AC, Kim SI. Brain size and cortical structure in the adult human brain. Cereb Cortex. 2008;18(9):2181-91. Thielscher A, Kammer T. Linking physics with physiology in TMS: a sphere field model to determine the cortical stimulation site in TMS. Neuroimage. 2002;17(3):1117-30. Huang Y, Liu AA, Lafon B, Friedman D, Dayan M, Wang X, et al. Measurements and models of electric fields in the in vivo human brain during transcranial electric stimulation. Elife. 2017;6. Valero-Cabré A, Amengual JL, Stengel C, Pascual-Leone A, Coubard OA. Corrigendum to "Transcranial magnetic stimulation in basic and clinical neuroscience: A comprehensive review of fundamental principles and novel insights" [Neurosci. Biobehav. Rev. 83 (2017) 381-404]. Neurosci Biobehav Rev. 2019;96:414. André-Obadia N, Mertens P, Gueguen A, Peyron R, Garcia-Larrea L. Pain relief by rTMS: differential effect of current flow but no specific action on pain subtypes. Neurology. 2008;71(11):833-40. Yeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106(3):1125-65. Chen P, Yao H, Tijms BM, Wang P, Wang D, Song C, et al. Four Distinct Subtypes of Alzheimer's Disease Based on Resting-State Connectivity Biomarkers. Biol Psychiatry. 2023;93(9):759-69. Fox MD, Buckner RL, White MP, Greicius MD, Pascual-Leone A. Efficacy of transcranial magnetic stimulation targets for depression is related to intrinsic functional connectivity with the subgenual cingulate. Biol Psychiatry. 2012;72(7):595-603. Cash RFH, Cocchi L, Lv J, Fitzgerald PB, Zalesky A. Functional Magnetic Resonance Imaging-Guided Personalization of Transcranial Magnetic Stimulation Treatment for Depression. JAMA Psychiatry. 2021;78(3):337-9. Park HJ, Friston K. Structural and functional brain networks: from connections to cognition. Science. 2013;342(6158):1238411. Di Lazzaro V, Rothwell JC. Corticospinal activity evoked and modulated by non-invasive stimulation of the intact human motor cortex. J Physiol. 2014;592(19):4115-28. Cerins A, Corp D, Opie G, Do M, Speranza B, He J, et al. Assessment of cortical inhibition depends on inter individual differences in the excitatory neural populations activated by transcranial magnetic stimulation. Sci Rep. 2022;12(1):9923. Opie GM, Semmler JG. Preferential Activation of Unique Motor Cortical Networks With Transcranial Magnetic Stimulation: A Review of the Physiological, Functional, and Clinical Evidence. Neuromodulation. 2021;24(5):813-28. Ziemann U. I-waves in motor cortex revisited. Exp Brain Res. 2020;238(7-8):1601-10. Additional Declarations No competing interests reported. Supplementary Files Supplementary.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7319128","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":508243229,"identity":"21d38675-479a-4048-912c-4b805d5f1aff","order_by":0,"name":"Nianshuang Wu","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Nianshuang","middleName":"","lastName":"Wu","suffix":""},{"id":508243230,"identity":"92397d7e-207b-4237-935f-6ce1dc24c94f","order_by":1,"name":"Ziyan Zhu","email":"","orcid":"","institution":"Xuanwu Hospital of Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ziyan","middleName":"","lastName":"Zhu","suffix":""},{"id":508243231,"identity":"df28ae60-435a-49bb-95f2-de566f4e21fa","order_by":2,"name":"Zhen Wu","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Zhen","middleName":"","lastName":"Wu","suffix":""},{"id":508243232,"identity":"ed98dee4-4916-4b85-863f-5c13a6f64664","order_by":3,"name":"Shuxiang Zhu","email":"","orcid":"","institution":"Xuanwu Hospital of Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Shuxiang","middleName":"","lastName":"Zhu","suffix":""},{"id":508243233,"identity":"9993010a-b95d-4f2d-b73e-59d70d660b06","order_by":4,"name":"Penghao Wang","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Penghao","middleName":"","lastName":"Wang","suffix":""},{"id":508243236,"identity":"af89e892-85c4-4a01-ba56-6eb98356909b","order_by":5,"name":"Yuxuan Shao","email":"","orcid":"","institution":"Xuanwu Hospital of Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuxuan","middleName":"","lastName":"Shao","suffix":""},{"id":508243238,"identity":"ed2a38b5-a2de-4ccd-a189-8309a19b0ea5","order_by":6,"name":"Cheng Zhang","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Cheng","middleName":"","lastName":"Zhang","suffix":""},{"id":508243239,"identity":"0ca9300c-9b68-4997-85de-3509a48c10a2","order_by":7,"name":"Changzhe Wu","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Changzhe","middleName":"","lastName":"Wu","suffix":""},{"id":508243241,"identity":"8be51e81-e141-4f12-b468-3246d155de82","order_by":8,"name":"Xiaolin Huo","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Xiaolin","middleName":"","lastName":"Huo","suffix":""},{"id":508243243,"identity":"399d7e55-e8a8-4333-aa57-c98f57346a8d","order_by":9,"name":"Hua Lin","email":"","orcid":"","institution":"Xuanwu Hospital of Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hua","middleName":"","lastName":"Lin","suffix":""},{"id":508243244,"identity":"26aef4ca-a5ef-424f-abe9-875da905f1f8","order_by":10,"name":"Guanghao Zhang","email":"data:image/png;base64,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","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":true,"prefix":"","firstName":"Guanghao","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-08-07 13:08:32","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7319128/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7319128/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90691641,"identity":"b03ea418-988a-44ac-947c-6e9abb2f5f85","added_by":"auto","created_at":"2025-09-05 18:18:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":852533,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverall study design.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFour ROIs (M1, DLPFC, AG, PC) were selected from 45 clinically recruited patients, with coil orientations adjusted accordingly. The MNI coordinates for these ROIs were as follows: left M1 (-34.19, -14.33, 66.83), left DLPFC (-38, 44, 26), left AG (-46, -64, 38)[41], and PC (0, -65, 45)[18]. The clinical treatment orientation was designated as the initial orientation (defined as 0°). The E-field components analyzed include E\u003csub\u003e⊥\u003c/sub\u003e, E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e, E\u003csub\u003eWM\u003c/sub\u003e, E\u003csub\u003eGM\u003c/sub\u003e, and E\u003csub\u003eROI\u003c/sub\u003e. The relationship between coil orientation and E-field components was evaluated to determine the optimal coil orientation, which was then validated against clinical outcomes.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/a11f9c191c52c6e26e5d60b4.png"},{"id":90691643,"identity":"9d0150ea-eda8-4d80-9369-2fecbb0d3cac","added_by":"auto","created_at":"2025-09-05 18:18:05","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":977023,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVariability of GM and WM voxels and E-field components at the initial orientation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA. Variability in GM and WM voxels across the four ROIs, normalized by the mean value of 45 patients, with significant differences observed in each brain region (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001). Horizontal lines (—) indicate means, cross symbols (×) denote medians, and error bars show the 95% confidence intervals. The ranking of the CVs for the WM voxels across the four brain regions is as follows: DLPFC, M1, AG, and PC. CV comparisons between WM and GM across regions are: M1 (75.42% vs. 33.09%), DLPFC (75.89% vs. 36.76%), AG (69.16% vs. 33.50%), and PC (63.05% vs. 31.94%).\u003c/p\u003e\n\u003cp\u003eB-E. Normalized E-fields distribution at the initial orientation, illustrated using data from patient1 at the d\u003cem\u003ei/\u003c/em\u003ed\u003cem\u003et\u003c/em\u003e value of 1A/μs. Region-specific CVs for E\u003csub\u003e⊥\u003c/sub\u003e vs. E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e were as follows: M1 (12.31% vs. 17.08%, \u003cem\u003eP\u003c/em\u003e=0.001), DLPFC (18.99% vs. 17.50%, \u003cem\u003eP \u003c/em\u003e= 0.057), AG (20.39% vs. 16.41%, \u003cem\u003eP\u003c/em\u003e=0.80), and PC (26.11% vs. 22.85%,\u003cem\u003e P \u003c/em\u003e= 0.48). Region-specific CVs for E\u003csub\u003eWM\u003c/sub\u003e vs. E\u003csub\u003eGM\u003c/sub\u003e were as follows: M1 (10.55% vs. 10.98%, \u003cem\u003eP\u003c/em\u003e=0.14), DLPFC (16.29% vs. 13.55%, \u003cem\u003eP \u003c/em\u003e= 0.01), AG (17.23% vs. 15.20%, \u003cem\u003eP\u003c/em\u003e=0.016), and PC (23.66% vs. 23.37%,\u003cem\u003e P \u003c/em\u003e= 0.020).\u003c/p\u003e\n\u003cp\u003eF-H. Correlations between the E-field components and voxel counts within the DLPFC ROI.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/6614e71fc86a7e0882365f00.png"},{"id":90691642,"identity":"197f3dc7-a622-4661-b051-3de86044b2a8","added_by":"auto","created_at":"2025-09-05 18:18:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1049064,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEffect of coil orientation on E-field components.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA-H: Changes in E-field components under different coil orientations, illustrated using patient1 as an example.\u003c/p\u003e\n\u003cp\u003eA. M1: E\u003csub\u003e⊥\u003c/sub\u003e peaked at 0° and reached its minimum at 90°, with a maximum to minimum ratio (RMmV) of 2.71. E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e peaked at 90° (RMmV=1.12) with minimum at -30°.\u003c/p\u003e\n\u003cp\u003eB. M1: E\u003csub\u003eWM\u003c/sub\u003e (peak 0°, RMmV=1.92), E\u003csub\u003eGM\u003c/sub\u003e (peak 0°, RMmV=1.29), and E\u003csub\u003eROI\u003c/sub\u003e (peak 0°, RMmV=1.50) all reached minima at 90°.\u003c/p\u003e\n\u003cp\u003eC. DLPFC: E\u003csub\u003e⊥\u003c/sub\u003e peaked at 30° (RMmV=2.83) with minimum at -60°; E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e peaked at -60° (RMmV=1.18) with minimum at 30°.\u003c/p\u003e\n\u003cp\u003eD. DLPFC: E\u003csub\u003eWM\u003c/sub\u003e (peak 30°, RMmV=1.67), E\u003csub\u003eGM\u003c/sub\u003e (peak 30°, RMmV=1.25), and E\u003csub\u003eROI\u003c/sub\u003e (peak 30°, RMmV=1.36) all reached minima at -60°.\u003c/p\u003e\n\u003cp\u003eE. AG: E\u003csub\u003e⊥\u003c/sub\u003e peaked at 60° (RMmV=1.81) with minimum at -60°; E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e peaked at -30° (RMmV=1.54) with minimum at 60°.\u003c/p\u003e\n\u003cp\u003eF. AG: E\u003csub\u003eWM\u003c/sub\u003e peaked at 30° (RMmV=1.29, minimum -60°); E\u003csub\u003eGM\u003c/sub\u003e peaked at 0° (RMmV=1.05, minimum 90°); E\u003csub\u003eROI\u003c/sub\u003e peaked at 30° (RMmV=1.13, minimum 120°).\u003c/p\u003e\n\u003cp\u003eG. PC: E\u003csub\u003e⊥\u003c/sub\u003e peaked at 60° (RMmV=1.50) with minimum at -30°; E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e peaked at 30° (RMmV=1.18) with minimum at -60°.\u003c/p\u003e\n\u003cp\u003eH. PC: E\u003csub\u003eWM\u003c/sub\u003e peaked at 30° (RMmV=1.33, minimum -60°); E\u003csub\u003eGM\u003c/sub\u003e peaked at 30° (RMmV=1.42, minimum -30°); E\u003csub\u003eROI\u003c/sub\u003e peaked at 30° (RMmV=1.38, minimum 120°).\u003c/p\u003e\n\u003cp\u003eI. Variability of E\u003csub\u003e⊥\u003c/sub\u003e and E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e across the four brain regions. Region-specific CV distributions were as follows: M1 (19.49% vs. 7.29%, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001), DLPFC (14.51% vs. 6.69%, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001), AG (14.78% vs. 15.09%, \u003cem\u003eP\u003c/em\u003e=0.016) and PC (14.55% vs. 13.81%, \u003cem\u003eP\u003c/em\u003e=0.039).\u003c/p\u003e\n\u003cp\u003eJ. Variability of E\u003csub\u003eWM\u003c/sub\u003e, E\u003csub\u003eGM\u003c/sub\u003e, and E\u003csub\u003eROI\u003c/sub\u003e across the four brain regions. Region-specific CV distributions for E\u003csub\u003eWM\u003c/sub\u003e and E\u003csub\u003eGM\u003c/sub\u003e were as follows: M1 (11.83% vs. 5.06%, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), DLPFC (11.32% vs. 4.36%, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), AG (7.25% vs. 3.64%, \u003cem\u003eP\u003c/em\u003e = 0.036) and PC (11.26% vs. 16.05%, \u003cem\u003eP\u003c/em\u003e = 0.010).\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/3778d8fad3e2b23176cd9ea4.png"},{"id":90692263,"identity":"b7e246b0-5e94-4d31-9af4-285f9f5103a3","added_by":"auto","created_at":"2025-09-05 18:26:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2022840,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSelection of the mOCOs based on E\u003c/strong\u003e\u003csub\u003e⊥\u003c/sub\u003e\u003cstrong\u003e and E\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003eWM\u003c/strong\u003e\u003c/sub\u003e\u003c/p\u003e\n\u003cp\u003eA-H: E-field distribution across 45 patients at different coil orientations. Distributions of E\u003csub\u003e⊥\u003c/sub\u003e (A-D) and E\u003csub\u003eWM\u003c/sub\u003e (E-H) across four brain regions: M1 (A, E), DLPFC (B, F), AG (C, G), and PC (D, H). ANOVA tests revealed that a significant variation in E\u003csub\u003e⊥\u003c/sub\u003e values across all coil orientations for M1 (\u003cem\u003eF\u003c/em\u003e = 25.87, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), AG (\u003cem\u003eF\u003c/em\u003e = 16.90, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), and PC (\u003cem\u003eF\u003c/em\u003e = 10.89, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001). Significant variation in E\u003csub\u003eWM\u003c/sub\u003e values was observed across the coil orientations for M1 (\u003cem\u003eF\u003c/em\u003e = 17.22, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), DLPFC (\u003cem\u003eF\u003c/em\u003e = 2.73, \u003cem\u003eP\u003c/em\u003e = 0.020), AG (\u003cem\u003eF\u003c/em\u003e = 3.32, \u003cem\u003eP\u003c/em\u003e = 0.006) and PC (\u003cem\u003eF\u003c/em\u003e = 7.41, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001).\u003c/p\u003e\n\u003cp\u003eI-J. Proportion of patients showing maximum E\u003csub\u003e⊥\u003c/sub\u003e (I) and E\u003csub\u003eWM\u003c/sub\u003e (J) at different coil orientations.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/7e7545b0cf02c3548fab8442.png"},{"id":90692264,"identity":"b51d2cd1-caab-4539-98fe-42e24e51fc20","added_by":"auto","created_at":"2025-09-05 18:26:05","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":387200,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelationship between the E-field components of the ROIs and clinical outcomes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA-D. Significant associations were observed between DLPFC E-field components (E\u003csub\u003e⊥\u003c/sub\u003e, E\u003csub\u003eWM\u003c/sub\u003e) and clinical outcomes: AVLT_all (A-B) and AVLT3 (C-D). E-F. The Δscore for patients\u0026nbsp; in subgroup-1(\u003cem\u003en\u003c/em\u003e=16) and subgroup-3 (\u003cem\u003en\u003c/em\u003e=17) was higher than that in subgroup-2 (\u003cem\u003en\u003c/em\u003e=14) and subgroup-4 (\u003cem\u003en\u003c/em\u003e=13), respectively, for the AVLT_all (E⊥: 6.50±7.21 vs. 3.93±4.34, \u003cem\u003eP\u003c/em\u003e = 0.26; EWM: 7.00±7.04 vs. 3.08±3.75, \u003cem\u003eP\u003c/em\u003e = 0.080; E), and AVLT3 (E⊥: 2.19±2.69 vs. 0.93±1.14, \u003cem\u003eP\u003c/em\u003e = 0.12; EWM: 2.29±2.64 vs. 0.69±0.75, \u003cem\u003eP\u003c/em\u003e = 0.043; F).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/d382cada2dd295c26c2602e1.png"},{"id":96708543,"identity":"d31c3e05-cc48-41fe-8c24-2134a9c46b25","added_by":"auto","created_at":"2025-11-25 10:04:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6238189,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/d46044a2-6fac-40b4-963a-a75ef1e1ff1c.pdf"},{"id":90691644,"identity":"6e0e57f0-bfdd-46bf-9f4f-55ef451b6116","added_by":"auto","created_at":"2025-09-05 18:18:05","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":297973,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementary.docx","url":"https://assets-eu.researchsquare.com/files/rs-7319128/v1/428fb3ff04393b7ba4768f0d.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"White Matter Electric Field Maximization Guided Coil Orientation Optimization for rTMS in Alzheimer's Disease","fulltext":[{"header":"1. Background","content":"\u003cp\u003eRepetitive transcranial magnetic stimulation (rTMS) has been widely used to treat neuropsychiatric disorders, particularly showing significant potential in the clinical management of Alzheimer's disease (AD)[1, 2]. However, current research has not integrated the anatomical features of AD with electric field (E-field) components induced by transcranial magnetic stimulation (TMS), an approach that could optimize stimulation parameters and enhance treatment efficacy.\u003c/p\u003e\u003cp\u003eMounting evidence establishes the E-field magnitude as a critical mediator of neuromodulation outcomes across diverse brain stimulation modalities. In rTMS protocols, cortical E-field magnitude was correlated with depressive symptom relief, while individualized auditory-verbal hallucination network E-field strength predicts therapeutic efficacy in schizophrenia interventions[3\u0026ndash;5]. Parallel findings in electroconvulsive therapy reveal hippocampal E-field intensity as a determinant of executive function changes, and transcranial alternating current stimulation (tACS) studies demonstrate age-dependent correlations between modeled E-field strength and cognitive performance[6, 7]. Crucially, the vector decomposition of TMS-generated E-fields into normal (perpendicular to cortical surface) and tangential components reveals distinct neurophysiological effects. Clinical data from treatment-resistant depression indicate that normal component magnitude, not tangential, mediates antidepressant responses during bilateral sequential rTMS, underscoring the therapeutic relevance of field orientation analysis[8].\u003c/p\u003e\u003cp\u003eTraditional E-field analyses have predominantly focused on gray matter (GM) territories, given their high neuronal soma density. However, emerging connectomic evidence positions white matter (WM) integrity as equally critical for maintaining functional network dynamics impaired in AD[9, 10]. The structural decoupling between GM nodes and WM edges, exacerbated by AD-related axonal degeneration, necessitates dual consideration of both tissue compartments in field modeling. Advanced multimodal approaches integrating diffusion tractography with E-field modeling demonstrate enhanced spatial specificity when accounting for WM anisotropy[11], while image-guided TMS paradigms show potential for transcending conventional depth limitations through network-targeted stimulation[12]. These developments highlight the imperative for combined GM-WM E-field analyses in AD therapeutic optimization.\u003c/p\u003e\u003cp\u003eThe biophysical foundation of TMS-induced E-fields comprises rotational (directly magnetic flux-derived) and irrotational (surface charge-mediated) components, computationally separable via contemporary simulation frameworks[13, 14]. For the irrotational E-field component, while existing models typically incorporate standard conductivity profiles, they frequently neglect AD-specific neuroanatomical perturbations. Progressive cerebral atrophy, a hallmark AD pathology exhibiting marked interindividual variability[15], fundamentally alters the cortical geometry of key rTMS targets including dorsolateral prefrontal cortex (DLPFC), angular gyrus (AG), and precuneus (PC)[16\u0026ndash;18]. Concurrently, amyloid-β/tau-mediated WM degeneration disrupts structural connectivity through myelin oxidative damage and neuroinflammation[19\u0026ndash;21], creating complex interactions between pathological tissue remodeling and E-field distribution. Though direct evidence remains sparse, transcranial direct current stimulation (tDCS) studies confirm cranial anatomy's significant impact on E-field patterns[22], while finite element analyses reveal tissue-specific E-field gradients dependent on coil positioning and WM anisotropy[23]. Therefore, further research is needed to explore the impact of brain atrophy on TMS induced E-fields in different stimulation targets.\u003c/p\u003e\u003cp\u003eFor the rotational E-field component, while neuronavigation systems achieve millimeter-level spatial precision in target localization[24, 25], rotational freedom around the coil-to-target axis introduces substantial angular variability[26]. Current clinical protocols often arbitrarily adopt 45\u0026deg; medial-sagittal angles[27, 28] or midline-parallel orientations[29, 30] without empirical justification, a stark contrast to M1 stimulation where motor evoked potentials (MEPs) enable real-time optimization guided by cortical column cosine models[31\u0026ndash;34]. The absence of analogous physiological feedback in cognitive targets necessitates computational approaches to orientation optimization. Though engineering studies propose generalized workflows incorporating finite element analysis and deep learning[24, 35\u0026ndash;38], their clinical translation remains impeded by technical complexity. This disparity between motor and cognitive stimulation paradigms creates critical knowledge gaps in rTMS parameter selection for AD.\u003c/p\u003e\u003cp\u003eIn this study, we hypothesize that orientation-dependent maxima in critical E-field components serve as robust predictors of enhanced therapeutic efficacy. Initially, we quantified the interindividual variability in GM and WM volumetric distributions and characterized their impact on E-field component. Then we evaluated orientation-dependent E-field modulation profiles to identify critical E-field components for optimization. This enabled the derivation of region-specific optimal coil orientations through E-field maximization across regions of interest (ROIs). The clinical relevance and therapeutic potential of these optimized orientations were subsequently validated through two clinical trials. Collectively, this work establishes a novel region-specific coil orientation optimization strategy to maximize the therapeutic potential of rTMS in AD treatment.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Patients and neuropsychological assessment\u003c/h2\u003e\u003cp\u003eWe recruited 45 patients diagnosed with probable AD from the outpatient clinic service of Xuanwu Hospital, Capital Medical University, Beijing, China. All patients provided written informed consent prior to participation. Diagnoses for all patients were made by two trained senior neurologists following a detailed consultation. In this study, 30 patients underwent rTMS targeting the left DLPFC from August 2022 to August 2023, while 15 patients received bilateral AG stimulation between November 2020 and September 2021. This study is registered in the Chinese Clinical Trial Registry (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.chictr.org.cn/index.html\u003c/span\u003e\u003cspan address=\"https://www.chictr.org.cn/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), number ChiCTR2200062564 and ChiCTR1900025045. The inclusion criteria for the DLPFC and AG groups, as well as the cognitive assessments, are detailed in eMethods in Supplement 1.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Image acquisition\u003c/h2\u003e\u003cp\u003eAll T1-weighted data were collected by a Siemens 3.0 T MRI system (Siemens, Erlangen, Germany). The specific MRI parameters and scanning requirements are presented in eMethods in Supplement 2.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 rTMS treatment\u003c/h2\u003e\u003cp\u003eThe stimulation of the left DLPFC and bilateral AG was performed using a Magstim Rapid\u003csup\u003e2\u003c/sup\u003e transcranial magnetic stimulator (Magstim, Co. Ltd, UK), equipped with an air-cooled figure-of-8 coil (70 mm diameter). The patient's head was co-registered with structural MRI using the BrainSight TMS navigation system (Rogue Research, Montreal, QC) to accurately target the intracerebral markers. In the DLPFC group, the stimulation target was the left DLPFC, whereas in the AG group, it was the bilateral AG. The protocol for rTMS therapy is detailed in eMethods in Supplement 3.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 E-field modeling\u003c/h2\u003e\u003cdiv id=\"Sec7\" class=\"Section3\"\u003e\u003ch2\u003e2.4.1 Coil positions and orientations\u003c/h2\u003e\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the overall workflow of this study. Simulations were performed for each patient at four target regions: left M1[39], left DLPFC[40], left AG[41], and PC[18]. For each target, the coil was positioned tangentially to the skull, with the handle pointing posterolaterally at a 45\u0026deg; angle to the midsagittal plane. The junction point of the figure-8 coil was positioned at the scalp point closest to the intracerebral target[27]. This position and orientation were defined as the initial orientation (0\u0026deg;) for this study. The line connecting the coil center and the intracranial target was defined as the rotation axis. Starting from 0\u0026deg;, the coil was rotated counterclockwise around the rotation axis in 30\u0026deg; increments up to 150\u0026deg;. Due to the symmetry of the stimulation coil, E-field values for coil orientations from \u0026minus;\u0026thinsp;180\u0026deg; to -30\u0026deg; were calculated by mirroring the corresponding orientations from the first half of the rotation (0\u0026deg; to 150\u0026deg;).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u003cem\u003ehere.\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\u003ch2\u003e2.4.2 Stimulation strength\u003c/h2\u003e\u003cp\u003eThe stimulation strength is determined by the d\u003cem\u003ei\u003c/em\u003e/d\u003cem\u003et\u003c/em\u003e value in SimNIBS software. For the M1 target, the d\u003cem\u003ei\u003c/em\u003e/d\u003cem\u003et\u003c/em\u003e value is set to 1 A/\u0026micro;s, while for the DLPFC, AG, and PC targets, the d\u003cem\u003ei\u003c/em\u003e/d\u003cem\u003et\u003c/em\u003e values are set to 0.8 A/\u0026micro;s, corresponding to 80% rMT. M1 E-field strength (E\u003csub\u003eM1\u003c/sub\u003e) of each patient was calculated as the top 99.9% of E-field magnitude from all voxels, serving as an estimate of the peak-induced field strength. Other simulation parameters were set to the default values.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section3\"\u003e\u003ch2\u003e2.4.3 E-field analyses\u003c/h2\u003e\u003cp\u003e Stimulation targets, defined in Montreal Neurological Institute (MNI) coordinates, were mapped to individual subject spaces and projected onto the nearest gyral crown surfaces relative to the stimulation coil. The gyral crown surface point was designated as the center of the region of interest (ROI), with a 10-mm radius. Within each ROI, the number of GM and WM voxels was calculated. The MRI T1 data processing is detailed in eMethods in Supplement 4. The E-fields in each voxel within the ROI were divided by E\u003csub\u003eM1\u003c/sub\u003e at the initial orientation to normalize the E-field strength. Then E-field strength in ROI (E\u003csub\u003eROI\u003c/sub\u003e) of each patient was calculated by averaging E-field strength in all voxels within ROI. Moreover, we divided the voxels in ROI into GM and WM in the simulation model. The GM and WM E-fields (E\u003csub\u003eGM\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e) were also calculated by averaging E-fields in their respective voxels. In addition, the E-field can be divided into two components: the normal component (E\u003csub\u003e\u0026perp;\u003c/sub\u003e), which is perpendicular to the cortical surface, and the tangential component (E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e), which runs parallel to it.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Statistical analyses\u003c/h2\u003e\u003cp\u003eDifferences between GM and WM voxels were assessed using T-tests. The values of GM and WM voxels, along with the E-fields, were standardized by dividing the data by the mean of the all patients. Data variability was expressed as the coefficient of variation (CV), and F Test and Levene's Test is used to evaluate variance homogeneity. The changes in clinical score (Δscore) were calculated as the difference between the post-treatment and pre-treatment scores. T-tests and Wilcoxon tests were performed on clinical scales before and after rTMS treatment to assess differences.\u003c/p\u003e\u003cp\u003eAn ANOVA was performed to assess whether statistically significant differences existed in E-field values across various coil orientations. Frequency statistics were used to quantify the occurrence of maximum E-field values at various coil orientations. The coil orientation that corresponded to the maximum values of E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e for each patient was defined as the patient's optimal coil orientation (pOCO) for E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e, respectively. A Chi-square test was applied to examine the frequency of occurrence of pOCO across different coil orientations for the 45 patients. The coil orientations with the highest frequency of occurrence were defined as the modal optimal coil orientation (mOCO). Pearson correlation was employed to assess the relationship between E-field components and clinical outcomes.\u003c/p\u003e\u003cp\u003eThe coil orientation deviation for each patient group was assessed by calculating the absolute difference between each patient's optimal and actual treatment orientations, with the standard deviation (SD) representing the deviation. T-test performed to compare the coil orientation deviations between the two groups.\u003c/p\u003e\u003cp\u003eThe patients in the DLPFC group were divided into four subgroups: Subgroup-1 and Subgroup-3 represented the optimal coil orientation groups (treatment orientation matched the pOCO) for the E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e, respectively, while Subgroup-2 and Subgroup-4 represented the non-optimal coil orientation groups (treatment orientation did not match the pOCO). The impact of optimal coil positioning on clinical outcomes was analyzed using a T-test to determine whether precise coil placement enhances treatment efficacy.\u003c/p\u003e\u003cp\u003eStatistical significance for all analyses was set at a two-sided \u003cem\u003eP\u003c/em\u003e value of less than 0.05. Statistical analyses were performed using SPSS Statistics V.26.0 (SPSS) and MATLAB (v.8.5, R2015a).\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Interindividual variability in GM/WM volumes and E-field components\u003c/h2\u003e\u003cp\u003eWe calculated the variability of WM and GM voxels within each ROI, as well as the E-field components at the initial coil orientation, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The number of GM voxels significantly exceeded that of WM voxels in all ROIs (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), with detailed voxel counts and CVs for four brain regions provided in Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. To facilitate comparison of the CVs of GM and WM voxel counts across different brain regions, the normalized results are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA. Quantitative analysis revealed significantly higher CVs in WM than GM across all target regions (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe TMS-induced E-fields exhibited distinct region-specific distribution patterns, with spatial heterogeneity observed in the E-field components within each ROI (Table S2). Normalized comparisons of E-field component CVs across different ROIs are systematically illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB-E, with the intracerebral E-field distribution of patient1 shown as a representative example. Cortical surface E-field component analysis revealed no significant difference in CVs between E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e, except M1 (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Voxel-based analysis of E-field components within ROIs revealed that E\u003csub\u003eWM\u003c/sub\u003e had significantly greater variability than E\u003csub\u003eGM\u003c/sub\u003e across all regions except M1. Notably, correlation analysis demonstrated a significant association between the E-field and regional brain volume only in the DLPFC (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), while no such association was observed in other regions (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05), as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF-H.\u003c/p\u003e\u003cp\u003e\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Orientation-dependent modulation of E-field components\u003c/h2\u003e\u003cp\u003eWe further examine how coil orientation variations impact E-fields. For each patient, we calculated the E-field components at various coil orientations within each ROI. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e shows the variability of E-field components across coil orientations. Representative data from patient 1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-H) demonstrate characteristic elliptical modulation patterns for all E-field components, where the major and minor axes of the ellipses correspond to the orientation-dependent maximum and minimum E-field magnitudes, respectively. Notably, the coil orientations yielding peak values for E\u003csub\u003eROI\u003c/sub\u003e, E\u003csub\u003eGM\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e showed strong spatial consistency, with angular deviations limited to less than 30\u0026deg; across these parameters.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe E-field components (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD) and their CVs for 45 patients across six coil orientations are shown in Table S3. Cortical surface E-field components analysis revealed significantly higher CVs for E\u003csub\u003e\u0026perp;\u003c/sub\u003e compared to E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e in all regions except AG (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI). Given the orthogonal maxima of E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003e\u003cem\u003e∥\u003c/em\u003e\u003c/sub\u003e across ROIs, coupled with CV difference less than 0.5% for both components in the AG, we prioritized E\u003csub\u003e\u0026perp;\u003c/sub\u003e for orientation-specific optimization. Voxel-based analysis of E-field components within ROIs revealed that E\u003csub\u003eWM\u003c/sub\u003e exhibited significantly higher CV than E\u003csub\u003eGM\u003c/sub\u003e in all regions except PC (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eJ). Intriguingly, while the PC exhibited maximal CV in GM, we still chose E\u003csub\u003eWM\u003c/sub\u003e for voxel-based optimization to maintain consistency in parameter selection across regions.\u003c/p\u003e\u003cp\u003e\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Optimal coil orientation determination\u003c/h2\u003e\u003cp\u003eTo determine the mOCO, E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e were selected as the key E-field components on the cortical surface and within the voxels of the ROI, respectively. We analyzed the distribution of maximum E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e across six coil orientations for each patient, as well as the frequency of occurrence of these maximum values as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA-H show the distributions of E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e at different coil orientations for all patients. ANOVA tests revealed significant variation in E\u003csub\u003e\u0026perp;\u003c/sub\u003e across all coil orientations (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), except for DLPFC (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.69), while E\u003csub\u003eWM\u003c/sub\u003e showed significant orientation dependence in all regions (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). These findings confirm orientation-specific E-field variations across brain regions, with DLPFC exhibiting unique insensitivity to E\u003csub\u003e\u0026perp;\u003c/sub\u003e modulation.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eI and J, along with Table S4, show the occurrence of frequency of pOCO for E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e across different coil orientations. In the M1 region, significant differences were found in the occurrence of pOCO for E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e (E\u003csub\u003e\u0026perp;\u003c/sub\u003e: \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; E\u003csub\u003eWM\u003c/sub\u003e: \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). For both E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e, about 50% of patients achieved maxima at 0\u0026deg;, with 86.7% of maxima clustered within \u0026minus;\u0026thinsp;30\u0026deg;~30\u0026deg;. This convergence motivated defining M1 mOCO as -30\u0026deg;~30\u0026deg;. In the DLPFC region, E\u003csub\u003e\u0026perp;\u003c/sub\u003e maxima exhibited non-significant dispersion across orientations (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.12), whereas E\u003csub\u003eWM\u003c/sub\u003e demonstrated significant orientation preference (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008), peaking at -30\u0026deg; (33.3% occurrence) with 84.4% of maxima distributed between \u0026minus;\u0026thinsp;60\u0026deg;~30\u0026deg;. The discordant E\u003csub\u003e\u0026perp;\u003c/sub\u003e profiles and broad E\u003csub\u003eWM\u003c/sub\u003e distribution precluded definitive mOCO determination. In the AG region, both E\u003csub\u003e\u0026perp;\u003c/sub\u003e and E\u003csub\u003eWM\u003c/sub\u003e showed strong orientation tuning: 88.9% of E\u003csub\u003e\u0026perp;\u003c/sub\u003e maxima concentrated at 30\u0026deg;~60\u0026deg; (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while 86.7% of E\u003csub\u003eWM\u003c/sub\u003e maxima spanned 0\u0026deg;~60\u0026deg; (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). This partial overlap justified defining AG mOCO as 0\u0026deg;~60\u0026deg;. In the PC region, significant orientation effects emerged for both E-field components (E\u003csub\u003e\u0026perp;\u003c/sub\u003e: \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; E\u003csub\u003eWM\u003c/sub\u003e: \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001). Remarkably, 88.9% of E\u003csub\u003e\u0026perp;\u003c/sub\u003e and 91.1% of E\u003csub\u003eWM\u003c/sub\u003e maxima co-localized at 30\u0026deg;~60\u0026deg;, establishing 30\u0026deg;~60\u0026deg; as the PC mOCO (The E\u003csub\u003eGM\u003c/sub\u003e optimization result remained consistent, as detailed in eResults, Supplement 1). A comprehensive summary of regional mOCO ranges is provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eOptimal coil orientation distribution\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBrain regions\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eE\u003csub\u003e\u0026perp;\u003c/sub\u003e range\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eE\u003csub\u003eWM\u003c/sub\u003e range\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eE\u003csub\u003e\u0026perp;\u003c/sub\u003e range\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eE\u003csub\u003eWM\u003c/sub\u003e range\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eM1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0\u0026deg; (-30\u0026deg;~30\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0\u0026deg; (-30\u0026deg;~30\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e45\u0026deg; (15\u0026deg;~75\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e45\u0026deg; (15\u0026deg;~75\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDLPFC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\\\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\\\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\\\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\\\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAG\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e45\u0026deg; (30\u0026deg;~60\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e30\u0026deg; (0\u0026deg;~60\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0\u0026deg; (-15\u0026deg;~15\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-15\u0026deg; (-45\u0026deg;~15\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e45\u0026deg; (30\u0026deg;~60\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e45\u0026deg; (30\u0026deg;~60\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0\u0026deg; (-15\u0026deg;~15\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0\u0026deg; (-15\u0026deg;~15\u0026deg;)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003eNote: a. The optimal orientation range refers to counterclockwise rotation angle from the initial coil handle orientation, which points posterolaterally at a 45-degree angle to the midsagittal plane. b. The optimal orientation range refers to the angle between the coil handle, which points posterolaterally, and the midsagittal plane.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Clinical correlates of E-field optimization\u003c/h2\u003e\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\u003ch2\u003e3.4.1 Coil orientation governs neuromodulation outcome uniformity\u003c/h2\u003e\u003cp\u003ePatient demographics are detailed in eResults in Supplement 2. Despite the uniform 0\u0026deg; coil orientation applied to all patients in the clinical trials, orientation discrepancies were observed between the actual coil orientations and the computationally derived optimal orientations in both the DLPFC and AG groups. The DLPFC group demonstrated pronounced variability of orientation discrepancies under both E\u003csub\u003e\u0026perp;\u003c/sub\u003e (42\u0026deg;\u0026plusmn;29.05\u0026deg;) and E\u003csub\u003eWM\u003c/sub\u003e (42\u0026deg;\u0026plusmn;25.65\u0026deg;) paradigms, whereas AG group exhibited markedly reduced deviations for E\u003csub\u003eWM\u003c/sub\u003e (22\u0026deg;\u0026plusmn;21.11\u0026deg;, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.013) despite comparable fluctuations (46\u0026deg;\u0026plusmn;19.20\u0026deg;, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.63) for E\u003csub\u003e\u0026perp;\u003c/sub\u003e. This refined directional specificity in AG group corresponded to superior clinical stabilization, manifesting as 39.19% reduced CV in improvement of MoCA scores relative to DLPFC group (101.03% vs 166.14%, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.006, Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e), with more detailed information provided in Table S5. The robust association between optimized field orientation fidelity and reduced outcome variability underscores precise coil orientation as a critical determinant of response predictability in rTMS targeting distinct cerebral regions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section3\"\u003e\u003ch2\u003e3.4.2 E-field strength and clinical outcomes\u003c/h2\u003e\u003cp\u003eGiven the absence of definitive mOCO in DLPFC, we investigated E-field strength correlations with cognitive changes. Significant associations emerged between specific E-field components and neuropsychological task performance in the DLPFC group (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-D), contrasting with null findings in the AG group (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Complete correlation matrices are provided in Table S6.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section3\"\u003e\u003ch2\u003e3.4.3 Relationship between individualized optimal coil orientation and clinical outcomes\u003c/h2\u003e\u003cp\u003eSubgroup stratification by E\u003csub\u003eWM\u003c/sub\u003e-optimized and E\u003csub\u003e\u0026perp;\u003c/sub\u003e-optimized coil orientation revealed critical orientation-dependent cognitive effects in DLPFC neuromodulation. Patients in Subgroup-3 demonstrated significantly greater improvements in AVLT3 scores (2.29\u0026thinsp;\u0026plusmn;\u0026thinsp;2.64 vs. 0.69\u0026thinsp;\u0026plusmn;\u0026thinsp;0.75, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.043) compared to Subgroup-4. Although the mean of AVLT_all score in Subgroup-3 (7.00\u0026thinsp;\u0026plusmn;\u0026thinsp;7.04) was greater than that in subgroup-4 (3.08\u0026thinsp;\u0026plusmn;\u0026thinsp;3.75), this difference did not reach statistical significance (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Patients in Subgroup-1 demonstrated greater improvements in AVLT_all (6.50\u0026thinsp;\u0026plusmn;\u0026thinsp;7.21 vs. 3.93\u0026thinsp;\u0026plusmn;\u0026thinsp;4.34) and AVLT3 (2.19\u0026thinsp;\u0026plusmn;\u0026thinsp;2.69 vs. 0.93\u0026thinsp;\u0026plusmn;\u0026thinsp;1.14) compared to Subgroup-2, however, these differences not statistically significant (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The aggregate cognitive profile, as visualized in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE-F, further supports the selective efficacy of E\u003csub\u003eWM\u003c/sub\u003e-guided orientation.\u003c/p\u003e\u003cp\u003e\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study establishes a neurobiologically grounded framework for optimizing TMS coil orientation in AD, demonstrating that target-specific maximization E\u003csub\u003eWM\u003c/sub\u003e significantly enhances therapeutic outcomes. Our findings extend prior motor cortex research to cognitive networks through three key advances: (1) identification of WM anatomical variations as primary contributors to TMS-induced E-field variability, (2) demonstration of superior angular modulation sensitivity of E\u003csub\u003eWM\u003c/sub\u003e compared to that of E\u003csub\u003eGM\u003c/sub\u003e, and (3) validation that precise alignment between computational optimization and clinical coil orientation reduces outcome variability while enhancing efficacy. These insights provide critical theoretical support for personalized neuromodulation in AD.\u003c/p\u003e\u003cp\u003eUnlike conventional E\u003csub\u003eROI\u003c/sub\u003e/E\u003csub\u003e\u0026perp;\u003c/sub\u003e maximization approaches[24, 35], our E\u003csub\u003eWM\u003c/sub\u003e-targeted optimization considers AD-specific neuropathology. While prior coil orientation studies predominantly focused on cortical geometry\u0026rsquo;s influence on E\u003csub\u003eROI\u003c/sub\u003e and E\u003csub\u003e\u0026perp;\u003c/sub\u003e[8, 26, 35, 42\u0026ndash;45], emerging computational evidence underscores the critical role of orientation-dependent brain network engagement[46, 47]. These findings collectively emphasize the necessity of addressing individual WM variability for optimizing cognitive modulation and therapeutic outcomes[10]. By integrating E\u003csub\u003eWM\u003c/sub\u003e quantification with E\u003csub\u003eROI\u003c/sub\u003e/E\u003csub\u003e\u0026perp;\u003c/sub\u003e metrics within ROIs, our approach fills a crucial gap in TMS research, offering a more comprehensive framework. While the methodological differences between E\u003csub\u003eWM\u003c/sub\u003e optimization and conventional E\u003csub\u003e\u0026perp;\u003c/sub\u003e maximization seem minor, their biophysical implications are fundamentally different. The E\u003csub\u003eWM\u003c/sub\u003e averages the E-field across WM voxels within ROIs, whereas E\u003csub\u003e\u0026perp;\u003c/sub\u003e is limited to CSF-GM interfaces, making E\u003csub\u003eWM\u003c/sub\u003e a more comprehensive biomarker for axonal activation patterns rather than just cortical effects. Notably, our study focuses on atrophic AD brains, which differ from studies on healthy or depressed individuals[3, 26, 36]. These AD brains exhibit significant GM/WM volumetric alterations, with GM predominance in ROI[48], coupled with substantial inter-individual anatomical variability in cortical thickness and sulcal geometry[49]. These factors collectively enhance E\u003csub\u003eWM\u003c/sub\u003e fluctuations. Remarkably, despite WM being deeper, where conventional distance decay principles would predict attenuated E-field intensity[50], E\u003csub\u003eWM\u003c/sub\u003e consistently demonstrated higher field strength than E\u003csub\u003eGM\u003c/sub\u003e across all ROIs. This apparent paradox can be explained by tissue conductivity gradients, where under quasi-static conditions governed by \u0026nabla;\u0026middot;J\u0026thinsp;=\u0026thinsp;0 and J\u0026thinsp;=\u0026thinsp;σE, the lower conductivity (σ) of WM results in elevated E-field magnitude at GM-WM interfaces relative to adjacent GM regions[51].\u003c/p\u003e\u003cp\u003eAlthough external stimulation parameters (e.g., intensity, frequency, and pulse count) are typically guided by protocols, the selection of coil orientation has received less attention. We resolve this gap by combining computational modeling with data from two independent clinical trials, ensuring comprehensive validation. Our motor cortex findings demonstrate that the maximum E\u003csub\u003eWM\u003c/sub\u003e and E\u003csub\u003e\u0026perp;\u003c/sub\u003e in M1 are optimally clustered within \u0026minus;\u0026thinsp;30\u0026deg;~30\u0026deg; of the central sulcus perpendicular, aligning with both established neurophysiological principles of MEP optimization[26, 52] and clinical evidence showing that the posteroanterior coil orientation, perpendicular to the central sulcus, is effective for rTMS pain therapy, while the lateromedial orientation is not[53]. Crucially, we extend our coil orientation optimization principle to cognitive relevant stimulation targets, i.e. DLPFC, AG, and PC, where no feedback signal similar to MEP can be used to help optimize coil orientation. The analysis of angular deviations between actual and optimal coil orientations revealed group differences based on optimization criteria, with E\u003csub\u003e\u0026perp;\u003c/sub\u003e maximization showed no significant variation between AG and DLPFC groups, whereas E\u003csub\u003eWM\u003c/sub\u003e maximization demonstrated reduced variability in the AG group. These findings support the reliability of E\u003csub\u003eWM\u003c/sub\u003e-based optimization. Consistent alignment between the actual and optimal coil orientations reduces variability in treatment outcomes and enhances efficacy. Our E\u003csub\u003eWM\u003c/sub\u003e maximization premise aligns with established E-field/efficacy correlations[3\u0026ndash;5, 8], supported by significant relationships between E-field and clinical outcome. These insights provide critical theoretical support for coil orientation optimization in AD. Although our study does not include clinical trials targeting PC, recent studies have demonstrated promising therapeutic outcomes by stimulating the PC with a midline-parallel coil orientation, inducing posterior-anterior currents[18]. Notably, this coil orientation aligns with the mOCO identified in our computational modeling study. Contrary to prevailing individualization trends[24, 36], we demonstrate non-personalized approaches suffice for AG, M1, and PC stimulation, significantly streamlining clinical workflows. For DLPFC, despite significant E\u003csub\u003eWM\u003c/sub\u003e variation across orientations (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008) with peak frequencies at -60\u0026deg; (26.67%), -30\u0026deg; (33.33%), 0\u0026deg; (11.11%), and 30\u0026deg; (13.33%), no universal mOCO exists within the 90\u0026deg; range, necessitating individualization. The 30\u0026deg; sampling interval balances practicality with field stability, because smaller increments (10\u0026deg;) show negligible field changes[26]. In addition, our left-hemisphere optimizations can be mirrored for right-sided targets.\u003c/p\u003e\u003cp\u003eTMS targets specific regions in the brain, which are part of different brain networks[54]. Though the exact mechanisms behind rTMS's therapeutic effects are still unclear, the DLPFC is thought to be a node in a brain network linked to cognitive impairment in AD[55]. Ongoing research aims to optimize and personalize rTMS targets based on functional connectivity[56, 57]. Emerging connectomic evidence positions WM integrity as equally critical for maintaining functional network dynamics impaired in AD[9, 10]. Functional connectivity is established upon the basis of structural connectivity[58], which is primarily composed of WM fibers. Our findings indicate that targeting the DLPFC necessitates personalized optimization, which may underlie the differential activation of neural networks through distinct coil orientations. These results establish a framework for future investigations to systematically examine how E\u003csub\u003eWM\u003c/sub\u003e modulate TMS-induced network effects.\u003c/p\u003e\u003cp\u003eOur study has several limitations. First, we calculated E-fields only from 0\u0026deg; to 150\u0026deg; due to mirror symmetry beyond 150\u0026deg; (180\u0026deg;-330\u0026deg;). This approach is practical for studying intracerebral E-field distribution, but previous studies on M1 have shown that rotating the TMS coil by 180\u0026deg; can alter motor response, including amplitude and latency[59, 60]. This is likely because different neural populations are recruited depending on the direction of the E-field[61, 62]. Second, the \"mri2mesh\" command in SimNIBS was used for image segmentation based on templates from healthy individuals. Future research should explore differences between segmented structures and actual anatomy of patients with brain atrophy. Third, the accuracy of conductivity values affects the simulation results, so more work is needed to optimize these values. Since WM is anisotropic, future studies should include DTI-based conductivity models to improve the accuracy of the simulations. Moreover, current methods don\u0026rsquo;t consider the low-conductivity Aβ and tau proteins, which could impact the accuracy of the simulations. Fourth, increasing the sample size is important for validating and generalizing findings. Finally, the AG group received both rehabilitation training and rTMS, and without a control group, it\u0026rsquo;s hard to separate the effects of rehabilitation training on clinical improvements. Although this study focuses on AD, the optimized coil orientations could also be applied to rTMS therapy for other neuropsychiatric disorders with some adjustments.\u003c/p\u003e"},{"header":"5 Conclusion","content":"\u003cp\u003eIn conclusion, this study highlights the importance of optimizing coil orientation in TMS for AD. We identified optimal coil orientations with the coil handle pointing posterolaterally at a -15\u0026deg; and 0\u0026deg; angle to the midsagittal plane for AG and PC respectively, while the DLPFC requires a more individualized approach. We also demonstrated E\u003csub\u003eWM\u003c/sub\u003e as a reliable indicator for determining the optimal coil orientations. These findings provide strong theoretical support for coil orientation optimization in rTMS protocols for AD patients.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was approved by the Ethics Committee of Xuanwu Hospital of Capital Medical University, and written informed consent was obtained from each participant. All participants were informed of the study\u0026rsquo;s objectives and procedures.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConsent for the publication of these data was not obtained from the participants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors thank all the participants who participated in this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was funded by several projects. The National Key R\u0026amp;D Program of China supported this work with the project numbers 2022YFC2402200. The National Natural Science Foundation of China provided support through the project numbers 51977205, 82471498, and NSFC - AF 82211530041. Additionally, the Capital\u0026apos;s Funds for Health Improvement and Research also contributed to this work with the project number CFH2022 - 2 - 2014. Neither of the mentioned funders was involved in the conception and performance of the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNianshuang Wu: Conceptualization, Investigation, Methodology, Writing \u0026ndash; original draft, review and editing, Supervision and Visualization. Ziyan Zhu: Data curation, Investigation, Methodology, Supervision and Visualization. Zhen Wu: Conceptualization, Investigation, Methodology, Supervision and Visualization. Shuxiang Zhu: Data curation. Penghao Wang: Conceptualization. Yuxuan Shao: Data curation. Cheng Zhang: Conceptualization. Changzhe Wu: Conceptualization. Xiaolin Huo: Conceptualization. Hua Lin: Conceptualization, Data curation, Investigation , Methodology, Supervision and Visualization, Project administration and Funding acquisition. Guanghao Zhang: Conceptualization, Investigation, Methodology, Writing \u0026ndash; original draft, review and editing, Supervision, Visualization, Project administration and Funding acquisition.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eLi X, Qi G, Yu C, Lian G, Zheng H, Wu S, et al. Cortical plasticity is correlated with cognitive improvement in Alzheimer\u0026apos;s disease patients after rTMS treatment. Brain Stimul. 2021;14(3):503-10.\u003c/li\u003e\n\u003cli\u003eSabbagh M, Sadowsky C, Tousi B, Agronin ME, Alva G, Armon C, et al. Effects of a combined transcranial magnetic stimulation (TMS) and cognitive training intervention in patients with Alzheimer\u0026apos;s disease. Alzheimers Dement. 2020;16(4):641-50.\u003c/li\u003e\n\u003cli\u003eDeng ZD, Robins PL, Dannhauer M, Haugen LM, Port JD, Croarkin PE. Optimizing TMS Coil Placement Approaches for Targeting the Dorsolateral Prefrontal Cortex in Depressed Adolescents: An Electric Field Modeling Study. Biomedicines. 2023;11(8).\u003c/li\u003e\n\u003cli\u003eQuinn DK, Upston J, Jones TR, Gibson BC, Olmstead TA, Yang J, et al. Electric field distribution predicts efficacy of accelerated intermittent theta burst stimulation for late-life depression. Front Psychiatry. 2023;14:1215093.\u003c/li\u003e\n\u003cli\u003eHua Q, Wang L, He K, Sun J, Xu W, Zhang L, et al. Repetitive Transcranial Magnetic Stimulation for Auditory Verbal Hallucinations in Schizophrenia: A Randomized Clinical Trial. JAMA Netw Open. 2024;7(11):e2444215.\u003c/li\u003e\n\u003cli\u003eDeng ZD, Argyelan M, Miller J, Quinn DK, Lloyd M, Jones TR, et al. Electroconvulsive therapy, electric field, neuroplasticity, and clinical outcomes. Mol Psychiatry. 2022;27(3):1676-82.\u003c/li\u003e\n\u003cli\u003eZanto TP, Jones KT, Ostrand AE, Hsu WY, Campusano R, Gazzaley A. Individual differences in neuroanatomy and neurophysiology predict effects of transcranial alternating current stimulation. Brain Stimul. 2021;14(5):1317-29.\u003c/li\u003e\n\u003cli\u003eZhang BBB, St\u0026ouml;hrmann P, Godbersen GM, Unterholzner J, Kasper S, Kranz GS, et al. Normal component of TMS-induced electric field is correlated with depressive symptom relief in treatment-resistant depression. Brain Stimul. 2022;15(5):1318-20.\u003c/li\u003e\n\u003cli\u003eQin T, Wang L, Xu H, Liu C, Shao Y, Li F, et al. rTMS concurrent with cognitive training rewires AD brain by enhancing GM-WM functional connectivity: a preliminary study. Cereb Cortex. 2024;34(1).\u003c/li\u003e\n\u003cli\u003eMart\u0026iacute;n-Signes M, Rodr\u0026iacute;guez-San Esteban P, Narganes-Pineda C, Caracuel A, Mata JL, Mart\u0026iacute;n-Ar\u0026eacute;valo E, et al. The role of white matter variability in TMS neuromodulatory effects. Brain Stimul. 2024;17(6):1265-76.\u003c/li\u003e\n\u003cli\u003eNummenmaa A, McNab JA, Savadjiev P, Okada Y, H\u0026auml;m\u0026auml;l\u0026auml;inen MS, Wang R, et al. Targeting of white matter tracts with transcranial magnetic stimulation. Brain Stimul. 2014;7(1):80-4.\u003c/li\u003e\n\u003cli\u003eLuber B, Davis SW, Deng ZD, Murphy D, Martella A, Peterchev AV, et al. Using diffusion tensor imaging to effectively target TMS to deep brain structures. Neuroimage. 2022;249:118863.\u003c/li\u003e\n\u003cli\u003eTofts PS, Branston NM. The measurement of electric field, and the influence of surface charge, in magnetic stimulation. Electroencephalogr Clin Neurophysiol. 1991;81(3):238-9.\u003c/li\u003e\n\u003cli\u003eThielscher A, Antunes A, Saturnino GB. Field modeling for transcranial magnetic stimulation: A useful tool to understand the physiological effects of TMS? Annu Int Conf IEEE Eng Med Biol Soc. 2015;2015:222-5.\u003c/li\u003e\n\u003cli\u003eMarshall GA, Lorius N, Locascio JJ, Hyman BT, Rentz DM, Johnson KA, et al. Regional cortical thinning and cerebrospinal biomarkers predict worsening daily functioning across the Alzheimer\u0026apos;s disease spectrum. J Alzheimers Dis. 2014;41(3):719-28.\u003c/li\u003e\n\u003cli\u003eChen H, Li M, Qin Z, Yang Z, Lv T, Yao W, et al. Functional network connectivity patterns predicting the efficacy of repetitive transcranial magnetic stimulation in the spectrum of Alzheimer\u0026apos;s disease. Eur Radiol Exp. 2023;7(1):63.\u003c/li\u003e\n\u003cli\u003eLiu C, Han T, Xu Z, Liu J, Zhang M, Du J, et al. Modulating Gamma Oscillations Promotes Brain Connectivity to Improve Cognitive Impairment. Cereb Cortex. 2022;32(12):2644-56.\u003c/li\u003e\n\u003cli\u003eKoch G, Casula EP, Bonn\u0026igrave; S, Borghi I, Assogna M, Minei M, et al. Precuneus magnetic stimulation for Alzheimer\u0026apos;s disease: a randomized, sham-controlled trial. Brain. 2022;145(11):3776-86.\u003c/li\u003e\n\u003cli\u003ePereira JB, Janelidze S, Ossenkoppele R, Kvartsberg H, Brinkmalm A, Mattsson-Carlgren N, et al. Untangling the association of amyloid-\u0026beta; and tau with synaptic and axonal loss in Alzheimer\u0026apos;s disease. Brain. 2021;144(1):310-24.\u003c/li\u003e\n\u003cli\u003eHuang Z, Jordan JD, Zhang Q. Myelin Pathology in Alzheimer\u0026apos;s Disease: Potential Therapeutic Opportunities. Aging Dis. 2024;15(2):698-713.\u003c/li\u003e\n\u003cli\u003eBartzokis G, Cummings JL, Sultzer D, Henderson VW, Nuechterlein KH, Mintz J. White matter structural integrity in healthy aging adults and patients with Alzheimer disease: a magnetic resonance imaging study. Arch Neurol. 2003;60(3):393-8.\u003c/li\u003e\n\u003cli\u003eLaakso I, Tanaka S, Mikkonen M, Koyama S, Sadato N, Hirata A. Electric fields of motor and frontal tDCS in a standard brain space: A computer simulation study. Neuroimage. 2016;137:140-51.\u003c/li\u003e\n\u003cli\u003eOpitz A, Windhoff M, Heidemann RM, Turner R, Thielscher A. How the brain tissue shapes the electric field induced by transcranial magnetic stimulation. Neuroimage. 2011;58(3):849-59.\u003c/li\u003e\n\u003cli\u003eBalderston NL, Roberts C, Beydler EM, Deng ZD, Radman T, Luber B, et al. A generalized workflow for conducting electric field-optimized, fMRI-guided, transcranial magnetic stimulation. Nat Protoc. 2020;15(11):3595-614.\u003c/li\u003e\n\u003cli\u003eWeise K, Numssen O, Kalloch B, Zier AL, Thielscher A, Haueisen J, et al. Precise motor mapping with transcranial magnetic stimulation. Nat Protoc. 2023;18(2):293-318.\u003c/li\u003e\n\u003cli\u003eJanssen AM, Oostendorp TF, Stegeman DF. The coil orientation dependency of the electric field induced by TMS for M1 and other brain areas. J Neuroeng Rehabil. 2015;12:47.\u003c/li\u003e\n\u003cli\u003eWu X, Ji GJ, Geng Z, Wang L, Yan Y, Wu Y, et al. Accelerated intermittent theta-burst stimulation broadly ameliorates symptoms and cognition in Alzheimer\u0026apos;s disease: A randomized controlled trial. Brain Stimul. 2022;15(1):35-45.\u003c/li\u003e\n\u003cli\u003eBagattini C, Zanni M, Barocco F, Caffarra P, Brignani D, Miniussi C, et al. Enhancing cognitive training effects in Alzheimer\u0026apos;s disease: rTMS as an add-on treatment. Brain Stimul. 2020;13(6):1655-64.\u003c/li\u003e\n\u003cli\u003eKoch G, Casula EP, Bonni S, Borghi I, Assogna M, Minei M, et al. Precuneus magnetic stimulation for Alzheimer\u0026apos;s disease: a randomized, sham-controlled trial. Brain. 2022;145(11):3776-86.\u003c/li\u003e\n\u003cli\u003eMencarelli L, Torso M, Borghi I, Assogna M, Pezzopane V, Bonni S, et al. Macro and micro structural preservation of grey matter integrity after 24 weeks of rTMS in Alzheimer\u0026apos;s disease patients: a pilot study. Alzheimers Res Ther. 2024;16(1):152.\u003c/li\u003e\n\u003cli\u003eBrasilneto JP, Cohen LG, Panizza M, Nilsson J, Roth BJ, Hallett M. Optimal Focal Transcranial Magnetic Activation of the Human Motor Cortex - Effects of Coil Orientation, Shape of the Induced Current Pulse, and Stimulus-Intensity. J Clin Neurophysiol. 1992;9(1):132-6.\u003c/li\u003e\n\u003cli\u003eMills KR, Boniface SJ, Schubert M. Magnetic brain stimulation with a double coil: the importance of coil orientation. Electroencephalogr Clin Neurophysiol. 1992;85(1):17-21.\u003c/li\u003e\n\u003cli\u003eBashir S, Perez JM, Horvath JC, Pascual-Leone A. Differentiation of motor cortical representation of hand muscles by navigated mapping of optimal TMS current directions in healthy subjects. J Clin Neurophysiol. 2013;30(4):390-5.\u003c/li\u003e\n\u003cli\u003eBalslev D, Braet W, McAllister C, Miall RC. Inter-individual variability in optimal current direction for transcranial magnetic stimulation of the motor cortex. J Neurosci Methods. 2007;162(1-2):309-13.\u003c/li\u003e\n\u003cli\u003eGomez-Tames J, Hamasaka A, Laakso I, Hirata A, Ugawa Y. Atlas of optimal coil orientation and position for TMS: A computational study. Brain Stimul. 2018;11(4):839-48.\u003c/li\u003e\n\u003cli\u003eCerins A, Thomas EHX, Barbour T, Taylor JJ, Siddiqi SH, Trapp N, et al. A New Angle on Transcranial Magnetic Stimulation Coil Orientation: A Targeted Narrative Review. Biol Psychiatry Cogn Neurosci Neuroimaging. 2024;9(8):744-53.\u003c/li\u003e\n\u003cli\u003eDannhauer M, Gomez LJ, Robins PL, Wang D, Hasan NI, Thielscher A, et al. Electric Field Modeling in Personalizing Transcranial Magnetic Stimulation Interventions. Biol Psychiatry. 2024;95(6):494-501.\u003c/li\u003e\n\u003cli\u003eMoser P, Reishofer G, Pr\u0026uuml;ckl R, Schaffelhofer S, Freigang S, Thumfart S, et al. Real-time estimation of the optimal coil placement in transcranial magnetic stimulation using multi-task deep learning. Sci Rep. 2024;14(1):19361.\u003c/li\u003e\n\u003cli\u003eNumssen O, Zier AL, Thielscher A, Hartwigsen G, Kn\u0026ouml;sche TR, Weise K. Efficient high-resolution TMS mapping of the human motor cortex by nonlinear regression. Neuroimage. 2021;245:118654.\u003c/li\u003e\n\u003cli\u003eMir-Moghtadaei A, Caballero R, Fried P, Fox MD, Lee K, Giacobbe P, et al. Concordance Between BeamF3 and MRI-neuronavigated Target Sites for Repetitive Transcranial Magnetic Stimulation of the Left Dorsolateral Prefrontal Cortex. Brain Stimul. 2015;8(5):965-73.\u003c/li\u003e\n\u003cli\u003eSchro\u0026euml;n JAM, Gunter TC, Numssen O, Kroczek LOH, Hartwigsen G, Friederici AD. Causal evidence for a coordinated temporal interplay within the language network. Proc Natl Acad Sci U S A. 2023;120(47):e2306279120.\u003c/li\u003e\n\u003cli\u003eThielscher A, Opitz A, Windhoff M. Impact of the gyral geometry on the electric field induced by transcranial magnetic stimulation. Neuroimage. 2011;54(1):234-43.\u003c/li\u003e\n\u003cli\u003eBijsterbosch JD, Barker AT, Lee KH, Woodruff PW. Where does transcranial magnetic stimulation (TMS) stimulate? Modelling of induced field maps for some common cortical and cerebellar targets. Med Biol Eng Comput. 2012;50(7):671-81.\u003c/li\u003e\n\u003cli\u003eLaakso I, Hirata A, Ugawa Y. Effects of coil orientation on the electric field induced by TMS over the hand motor area. Phys Med Biol. 2014;59(1):203-18.\u003c/li\u003e\n\u003cli\u003eGomez LJ, Dannhauer M, Peterchev AV. Fast computational optimization of TMS coil placement for individualized electric field targeting. Neuroimage. 2021;228:117696.\u003c/li\u003e\n\u003cli\u003eOpitz A, Fox MD, Craddock RC, Colcombe S, Milham MP. An integrated framework for targeting functional networks via transcranial magnetic stimulation. Neuroimage. 2016;127:86-96.\u003c/li\u003e\n\u003cli\u003eLynch CJ, Elbau IG, Ng TH, Wolk D, Zhu S, Ayaz A, et al. Automated optimization of TMS coil placement for personalized functional network engagement. Neuron. 2022;110(20):3263-77.e4.\u003c/li\u003e\n\u003cli\u003eGe Y, Grossman RI, Babb JS, Rabin ML, Mannon LJ, Kolson DL. Age-related total gray matter and white matter changes in normal adult brain. Part I: volumetric MR imaging analysis. AJNR Am J Neuroradiol. 2002;23(8):1327-33.\u003c/li\u003e\n\u003cli\u003eIm K, Lee JM, Lyttelton O, Kim SH, Evans AC, Kim SI. Brain size and cortical structure in the adult human brain. Cereb Cortex. 2008;18(9):2181-91.\u003c/li\u003e\n\u003cli\u003eThielscher A, Kammer T. Linking physics with physiology in TMS: a sphere field model to determine the cortical stimulation site in TMS. Neuroimage. 2002;17(3):1117-30.\u003c/li\u003e\n\u003cli\u003eHuang Y, Liu AA, Lafon B, Friedman D, Dayan M, Wang X, et al. Measurements and models of electric fields in the in vivo human brain during transcranial electric stimulation. Elife. 2017;6.\u003c/li\u003e\n\u003cli\u003eValero-Cabr\u0026eacute; A, Amengual JL, Stengel C, Pascual-Leone A, Coubard OA. Corrigendum to \u0026quot;Transcranial magnetic stimulation in basic and clinical neuroscience: A comprehensive review of fundamental principles and novel insights\u0026quot; [Neurosci. Biobehav. Rev. 83 (2017) 381-404]. Neurosci Biobehav Rev. 2019;96:414.\u003c/li\u003e\n\u003cli\u003eAndr\u0026eacute;-Obadia N, Mertens P, Gueguen A, Peyron R, Garcia-Larrea L. Pain relief by rTMS: differential effect of current flow but no specific action on pain subtypes. Neurology. 2008;71(11):833-40.\u003c/li\u003e\n\u003cli\u003eYeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106(3):1125-65.\u003c/li\u003e\n\u003cli\u003eChen P, Yao H, Tijms BM, Wang P, Wang D, Song C, et al. Four Distinct Subtypes of Alzheimer\u0026apos;s Disease Based on Resting-State Connectivity Biomarkers. Biol Psychiatry. 2023;93(9):759-69.\u003c/li\u003e\n\u003cli\u003eFox MD, Buckner RL, White MP, Greicius MD, Pascual-Leone A. Efficacy of transcranial magnetic stimulation targets for depression is related to intrinsic functional connectivity with the subgenual cingulate. Biol Psychiatry. 2012;72(7):595-603.\u003c/li\u003e\n\u003cli\u003eCash RFH, Cocchi L, Lv J, Fitzgerald PB, Zalesky A. Functional Magnetic Resonance Imaging-Guided Personalization of Transcranial Magnetic Stimulation Treatment for Depression. JAMA Psychiatry. 2021;78(3):337-9.\u003c/li\u003e\n\u003cli\u003ePark HJ, Friston K. Structural and functional brain networks: from connections to cognition. Science. 2013;342(6158):1238411.\u003c/li\u003e\n\u003cli\u003eDi Lazzaro V, Rothwell JC. Corticospinal activity evoked and modulated by non-invasive stimulation of the intact human motor cortex. J Physiol. 2014;592(19):4115-28.\u003c/li\u003e\n\u003cli\u003eCerins A, Corp D, Opie G, Do M, Speranza B, He J, et al. Assessment of cortical inhibition depends on inter individual differences in the excitatory neural populations activated by transcranial magnetic stimulation. Sci Rep. 2022;12(1):9923.\u003c/li\u003e\n\u003cli\u003eOpie GM, Semmler JG. Preferential Activation of Unique Motor Cortical Networks With Transcranial Magnetic Stimulation: A Review of the Physiological, Functional, and Clinical Evidence. Neuromodulation. 2021;24(5):813-28.\u003c/li\u003e\n\u003cli\u003eZiemann U. I-waves in motor cortex revisited. Exp Brain Res. 2020;238(7-8):1601-10.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Alzheimer's disease, Transcranial magnetic stimulation, Electric field, White matter, Coil orientations","lastPublishedDoi":"10.21203/rs.3.rs-7319128/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7319128/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eBackground\u003c/b\u003e\u003c/p\u003e\u003cp\u003eRepetitive transcranial magnetic stimulation (rTMS) has emerged as a promising intervention for Alzheimer's disease (AD), yet current protocols lack standardized methodologies for correlating coil orientation with electric field (E-field) characteristics. This study establishes coil orientation optimization protocols for AD-related rTMS targets by analyzing E-field distribution patterns.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMethods\u003c/b\u003e\u003c/p\u003e\u003cp\u003eIn a study population of 45 AD patients undergoing targeted stimulation of either the dorsolateral prefrontal cortex (DLPFC, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;30, M:F, 11:19) or angular gyrus (AG, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;15, M:F, 8:7), we performed E-field simulations across four critical regions: DLPFC, AG, precuneus (PC), and primary motor cortex (M1), determining optimal orientation through region-specific E-field maximization. Post hoc analysis of neuropsychological outcomes validated the proposed optimization strategy.\u003c/p\u003e\u003cp\u003e\u003cb\u003eResults\u003c/b\u003e\u003c/p\u003e\u003cp\u003eOur findings revealed that white matter E-field (E\u003csub\u003eWM\u003c/sub\u003e) and normal E-field component (E\u003csub\u003e\u0026perp;\u003c/sub\u003e) demonstrated higher orientation-dependent variability compared to gray matter E-field (E\u003csub\u003eGM\u003c/sub\u003e), despite larger GM volumes. Orientation optimization achieved more than 85% consistency rates for M1 (45\u0026deg;), AG (-15\u0026deg;) and PC (0\u0026deg;) through E\u003csub\u003eWM\u003c/sub\u003e maximization, whereas DLPFC required individualized coil orientation. Clinical validation demonstrated that AG-targeted stimulation maintained significantly lower orientation deviation variance compared to DLPFC protocols, corresponding to reduced coefficient of variation (CV) in the improvement of MoCA scores (55.65% vs. 100.76%, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.006). Notably, patients aligned with E\u003csub\u003eWM\u003c/sub\u003e-optimized orientation in DLPFC group showed superior improvement of auditory verbal learning test (AVLT) scores compared to non-optimized cases (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusions\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThese findings establish E\u003csub\u003eWM\u003c/sub\u003e as a critical determinant of rTMS efficacy and advocate for region-specific orientation protocols to enhance AD treatment outcomes.\u003c/p\u003e","manuscriptTitle":"White Matter Electric Field Maximization Guided Coil Orientation Optimization for rTMS in Alzheimer's Disease","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-05 18:18:00","doi":"10.21203/rs.3.rs-7319128/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e6e6f789-24b8-49f0-b68c-574e6e327b52","owner":[],"postedDate":"September 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-11-24T11:38:39+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-05 18:18:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7319128","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7319128","identity":"rs-7319128","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}