糖心vlog

Key Points

Question听 Is imaging-navigated repetitive transcranial magnetic stimulation (rTMS) an effective treatment for auditory verbal hallucinations (AVH) in patients with schizophrenia?

Findings听 In this randomized clinical trial of 62 patients with schizophrenia and AVH, active rTMS was superior to sham treatment in alleviating AVH symptoms, and the effects were maintained throughout the follow-up period. Furthermore, the strength of the TMS-induced electric field was independently associated with AVH symptom improvements.

Meaning听 The findings of this study indicate that imaging-navigated rTMS is an effective and safe treatment for AVH in patients with schizophrenia.

Importance听 Auditory verbal hallucinations (AVH) are a common symptom of schizophrenia, increasing the patient鈥檚 risks of suicide and violence. Repetitive transcranial magnetic stimulation (rTMS) is a potential treatment for AVH.

Objective听 To investigate the effect of imaging-navigated rTMS on AVH in patients with schizophrenia.

Design, Setting, and Participants听 This 6-week, double-blind, sham-controlled, randomized clinical trial was performed at the Anhui Mental Health Center, Hefei, China, from September 1, 2016, to August 31, 2021. Participants included 66 patients with AVH and schizophrenia. Data were analyzed from May 1, 2022, to March 31, 2023.

Interventions听 Participants were randomly assigned 1:1 to either imaging-navigated active or sham rTMS over the left temporoparietal junction for 2 weeks.

Main Outcomes and Measures听 The primary outcome measured improvements in AVH from baseline to week 2 and week 6 using the Auditory Hallucination Rating Scale (AHRS) scores. In addition, the TMS-induced electric field strength was used to estimate improvements in AVH as a secondary outcome.

Results听 A total of 62 participants (33 women [53%]; mean [SD] age, 27.4 [9.2] years) completed the 2-week treatments. Of these, 32 were randomized to the active rTMS group (18 women [56%]; mean [SD] age, 26.9 [9.2] years) and 30 to the sham treatment group (15 women [50%]; mean [SD] age, 27.8 [9.4] years). In the intention-to-treat analyses, patients receiving active rTMS showed a significantly greater reduction in AHRS scores compared with those receiving sham treatment at week 2 (difference, 5.96 [95% CI, 3.42-8.50]; t鈥�=鈥�4.61; P鈥�&濒迟;鈥�.001; Cohen d,鈥�1.17 [95% CI, 0.62-1.71]). These clinical effects were sustained at week 6. Additionally, a stronger TMS-induced electric field within a predefined AVH brain network was associated with greater reductions in AHRS scores (B鈥�=鈥�3.12; t鈥�=鈥�3.58; P鈥�=鈥�.002). No serious adverse event was observed.

Conclusions and Relevance听 The findings of this randomized clinical trial suggest that imaging-navigated rTMS may effectively and safely alleviate AVH in patients with schizophrenia. Findings also suggest that the electric field strength in the individualized AVH network is a vital parameter for optimizing the efficacy of the rTMS protocol.

Trial Registration听 ClinicalTrials.gov Identifier:

Auditory verbal hallucinations (AVH), generally defined as hearing nonexistent spoken voices, are a typical symptom of schizophrenia.¹ Approximately 70% to 80% of patients with schizophrenia have AVH at initial presentation,^2,3 and of these, 25% to 30% of AVH are nonresponsive to antipsychotic medications.⁴ The persistence of AVH can significantly diminish the quality of life⁵ and elevate the risk of suicide⁶ and violence.⁷ Consequently, effective alternative therapies are urgently needed to treat AVH in patients with schizophrenia. To this end, numerous neuroimaging studies have been tested to identify the neural correlates of AVH.^8,9

Among the abundant brain regions or networks that may be involved in the generation of AVH in patients with schizophrenia,¹⁰ the left temporoparietal junction (TPJ) has attracted particular attention because of its potential for reducing AVH in clinical treatments such as repetitive transcranial magnetic stimulation (rTMS).^11-14 Most of the rTMS studies treating AVH in patients with schizophrenia^15-19 localized the left TPJ according to the international 10-20 system of electroencephalography. However, the TPJ is a heteromodal association cortex consisting of heterogeneous subregions^20-23 with high interindividual variability.²⁴ The 10-20 system is not sufficiently accurate to guide the TMS coil to a predefined TPJ subregion.²⁵ As a result, the stimulation may be delivered to functionally distinct TPJ subregions on different treatment days.²⁶ This imprecise stimulation may fail to accomplish an accumulated effect to alleviate AVH symptoms in patients with schizophrenia.¹⁵ To overcome this shortcoming, our group precisely guided the coil and monitored its online position using individual magnetic resonance imaging (MRI) of the brain in an open-label rTMS study, which significantly decreased AVH in patients with schizophrenia.¹³

Despite the success of the imaging-navigated rTMS treatment at the group level, our group also noted high outcome variability among individuals.¹³ This variability could be attributed to the heterogeneous neuroanatomy (eg, cortical folding pattern) effect on the TMS-induced electric field (e-field). The imaging-navigated procedure ensured the spatial alignment between coil and target, but the e-field shape and strength were differentially distorted by the neuroanatomy of different individuals.^27-29 E-field variability in local areas and brain networks has been used to explain the treatment effects. For example, the e-field on the stimulation site³⁰ or within the depression-related network³¹ was associated with improvements in depressive symptoms. These local and network explanations are not mutually exclusive, and both may be related to the AVH outcome variability in patients with schizophrenia.

In this study, we designed a randomized, double-blind, sham-controlled clinical trial to test the efficacy of imaging-navigated rTMS in reducing AVH in patients with schizophrenia. To estimate the outcome variability, we correlated individual AVH improvements with the simulated e-field strength in the stimulation site and a predefined AVH network. We defined the AVH network as a circuit composed of brain regions that were associated with AVH.^32,33

This single-site, randomized, double-blind, sham-controlled clinical trial was performed at the Anhui Mental Health Center, Hefei, China, from September 1, 2016, to August 31, 2021. The Institutional Ethics Committee of Anhui Medical University approved the protocol, which is found in Supplement 1. This study followed the Consolidated Standards of Reporting Trials () reporting guideline. All participants provided written, informed consent to participate.

Patients with schizophrenia were recruited through posted flyers and physician referrals at Anhui Mental Health Center. The inclusion criteria for patients consisted of the following: (1) diagnosis by independent psychiatrists using the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition); (2) age of 18 to 60 years; (3) medication-resistant AVH (defined as an insufficient response to antipsychotic agents for AVH, administered at adequate dosages for at least 6 weeks^34,35); and (4) a stable dosage of antipsychotic medication for at least 2 weeks before inclusion and retention of this stable dose for the duration of the study. Exclusion criteria consisted of the following: (1) other comorbid mental illnesses or histories; (2) pregnancy; (3) a history of severe head trauma or neurological disease; (4) focal brain lesions on T1- or T2-weighted fluid-attenuated inversion-recovery MRI; (5) a history of rTMS or electroconvulsive therapy; (6) current severe suicidal thoughts or behavior; and (7) metal objects in the head or any other contraindication to MRI.

Participants were randomly assigned to an active treatment or a sham treatment group at a 1:1 ratio using a random number generator by an unblinded investigator (G.-J. J.) who was not involved with the study ratings or analysis. Patients, clinical raters, and all personnel responsible for the clinical care of patients were blinded to the treatment assignments. To measure blinding integrity, the participants were asked at week 6 to guess the group to which they were randomly assigned and were subsequently informed of their respective treatment conditions.

Structural and resting-state functional MRI (rs-fMRI) data were acquired for each patient at baseline (before the first session of rTMS treatment) using a 3.0-T scanner (Discovery MR750; GE Healthcare) at the University of Science and Technology of China in Hefei. The specific MRI parameters and scanning requirements are presented in eMethods 1 in Supplement 2.

Treatment was performed using a transcranial magnetic stimulator (Rapid²; Magstim) with a 70-mm air-cooled figure-of-8 coil. Participants received 3 daily sessions of rTMS treatment for 2 weeks (14 consecutive days). The specific rTMS protocol followed that of the previous open-label study from our group¹³ (eMethods 2 in Supplement 2).

Sham treatments were delivered with the same rTMS protocol using a sham coil (Magstim) that was identical in appearance to the real one. The only difference was that the sham coil generated only sound and sensations on the scalp similar to those of the real coil but no current, which prevented patients from identifying the rTMS group.

We calculated realistic estimates of the e-fields induced by the TMS coil in our intervention based on the finite elements method using the SimNIBS software package, version 4.0.0 (Danish Research Center for Magnetic Resonance and the Technical University of Denmark).³⁶ We computed the head mesh for each patient using their individual T1- and T2-weighted images.³⁷ Brain images were segmented into 5 tissue types (white matter, gray matter, cerebrospinal fluid, bone, and skin). We then estimated the e-field strength using default tissue conductivities and coil-to-scalp distances.³⁸ The position and orientation of the TMS coil for each patient were exported from the neuronavigation system. The stimulation intensity (the derivative of the current flowing across the inductor [di/dt]鈥�=鈥�165 A/渭s) was provided by the transcranial magnetic stimulator manufacturer.

We calculated e-field strength for 2 regions of interest, target and AVH network regions. The e-field strength of the TMS target region was the mean strength within the sphere target of TPJ. The e-field strength in each patient鈥檚 AVH network was computed in 3 steps (eFigure 1 in Supplement 2). First, we computed the group-level AVH network defined as the connectivity map of a sphere seed in the cerebellum (Montreal Neurological Institute coordinates, 1, 鈭�50, 鈭�28). This seed in the cerebellum was the hub region of the AVH network identified by a prior lesion network mapping study.³² Specifically, we computed cerebellum seed鈥搕o鈥搘hole brain functional connectivity on the rs-fMRI data of 652 healthy individuals as part of a previous protocol (see the demographic information and MRI parameters in eMethods 3 in Supplement 2), then performed a 1-sample t test on 652 connectivity maps to obtain the group-level network. Second, we individualized this group-level AVH network by using each patient鈥檚 rs-fMRI data acquired at baseline. We used the positive part of the group-level AVH network as a seed map and computed its functional connectivity (Pearson correlation) with each voxel in the patient鈥檚 brain. Voxels positively correlated with the seed map (uncorrected P鈥�<鈥�.05) constituted the individualized AVH network of each patient. Finally, we calculated the mean effective e-field strength (>10 V/m)³⁹ within individualized AVH networks.

All assessments were performed by a trained psychiatrist (L.W.) blinded to the patients鈥� rTMS conditions. Clinical symptom assessments were performed at baseline, week 2 (after treatment), and week 6 (follow-up). Any adverse events related to rTMS (eg, significant discomfort, pain, or harm caused by the intervention) were recorded throughout the study from the first stimulation session to the last follow-up visit at week 6.

The primary outcome measure was the change in the Auditory Hallucination Rating Scale (AHRS) scores from baseline to weeks 2 and 6. Secondary outcomes included the Positive and Negative Syndrome Scale (PANSS), Hamilton Anxiety Rating Scale with 14 items, Hamilton Depression Rating Scale with 17 items, and response rates at weeks 2 and 6. Response to rTMS treatment was defined as a reduction of 25% or more from baseline in the AHRS score.¹³ We estimated the TMS-induced e-field strength using baseline structural brain images and tested whether the e-field strength could estimate the improvement of AVH symptoms after treatment.

Data were analyzed from May 1, 2022, to March 31, 2023. The sample size was estimated for a power of 80% and a 2-tailed 伪 of .05 for assessing AVH improvements. We calculated the sample size according to the effect size of 0.76 in low-frequency rTMS in the treatment of resistant AVH, the effect size reported in a previous meta-analysis.⁴⁰ In addition, considering a 10% attrition rate, each group was required to include 33 participants.

All patients who underwent randomization and received at least 1 active or sham treatment were classified as the intention-to-treat (ITT) analysis set. Notably, 4 participants discontinued trials before baseline assessments and first treatments, so their data were unavailable for analysis. Consequently, the main analyses were conducted on the ITT analysis set, excluding these 4 participants.

We used SPSS statistical software for Windows, version 23.0 (IBM Corporation) for statistical analysis. Baseline demographic and clinical characteristics were compared between the 2 groups using 蠂² tests for categorical variables. Continuous variables were compared using independent-sample t tests and Mann-Whitney tests, depending on the normality of the data. The normality of continuous variables was tested using the Shapiro-Wilks test, and if normality was violated, we used a Mann-Whitney test. For all continuous outcome measures, we used linear mixed-effects models to test the treatment efficacies. Time, group, and time-by-group interaction were included in the model as fixed effects. The individual participant intercept was included in the model as a random effect. For categorical outcome measures, we used 蠂² tests (or Fisher exact tests if the expected frequencies in any cell of the contingency table were less than 1) to compare the differences between groups. Two-sided P鈥�<鈥�.05 was considered statistically significant. Effect sizes were calculated as Cohen d and odds ratios (ORs) with 95% CIs for continuous and binary outcomes, respectively. We also provided the number needed to treat (NNT) to assess treatment effectiveness.⁴¹

To assess whether the TMS-induced e-field strength was associated with clinical efficacy, hierarchical multiple regression analysis was conducted.⁴² Hierarchical multiple regression analysis included at least 2 models. Model 1 represented a control model with confounders. Model 2 included both confounders and estimation factors. In this study, model 1 included the reduction in AHRS scores as the dependent variable and baseline data as confounders (ie, sex, age, education years, illness duration, olanzapine equivalent, and baseline AHRS scores). In model 2, the estimation factors were TMS-induced e-field strength within the TPJ target region or AVH network. The estimation abilities of the 2 models were compared using an F test. We then used a 1-sample t test to assess the contribution of the factors in estimating the dependent variable. Notably, the sham coil in the sham treatment group did not generate an e-field in the cortex. Thus, it was not appropriate to include both active and sham treatment data in 1 regression model, although we could simulate the e-field for the sham treatment group.

Of 104 participants, 38 were excluded for several reasons and 66 (31 men [47%] and 35 women [53%]; mean [SD] age, 27.4 [9.0] years) were randomized to the active group (n鈥�=鈥�33) or sham treatment group (n鈥�=鈥�33). A total of 62 participants (29 men [47%] and 33 women [53%]; mean [SD] age, 27.4 [9.2] years) completed the 2-week treatments. Of these, 32 participants were randomized to the active rTMS group (14 men [44%] and 18 women [56%]; mean [SD] age, 26.9 [9.2] years) and 30 to the sham treatment group (15 men [50%] and 15 women [50%]; mean [SD] age, 27.8 [9.4] years). There was no significant difference in baseline demographic and clinical variables between the active rTMS and sham treatment groups (Table 1). There was a significant difference in the dropout rates between the active rTMS and sham treatment groups during the follow-up phase (4 in the active rTMS group and 12 in the sham treatment group; 蠂²鈥�=鈥�6.12; P鈥�=鈥�.02). Participant flow is illustrated in Figure 1.

The primary outcome consisted of changes in AHRS scores at weeks 2 and 6. In the ITT analysis set, linear mixed-effects model analysis revealed significant time鈥壝椻�塯roup interactions for the AHRS at week 2 (F_1,60鈥�=鈥�21.20; P鈥�<鈥�.001) and week 6 (F_1,48.62鈥�=鈥�24.52; P鈥�<鈥�.001). A post hoc test revealed that patients receiving active rTMS showed a greater reduction in the AHRS scores than the sham treatment group at week 2 (between-group difference, 5.96 [95% CI, 3.42-8.50]; P鈥�&濒迟;鈥�.001; t = 4.61; Cohen d, 1.17 [95% CI, 0.63-1.71]) and at week 6 (group difference, 7.89 [95% CI, 4.77-11.01]; P鈥�&濒迟;鈥�.001; Cohen d,鈥�1.49 [95% CI, 0.82-2.16]) (Figure 2A and Table 2).

Participants in the active rTMS group had higher response rates than the sham treatment group at week 2 (15 of 32 [47%] vs 4 of 30 [13%]; 蠂²鈥�=鈥�8.20; P鈥�=鈥�.004; OR,鈥�5.74 [95% CI, 1.63-20.24]; NNT,鈥�2.98 [95% CI, 1.86-12.04]) (Table 2 and Figure 2B). The same pattern was observed at week 6 (14 of 28 [50%] vs 1 of 18 [6%]; 蠂²鈥�=鈥�9.85; P鈥�=鈥�.002; OR,鈥�17.00 [95% CI, 1.98-145.73]; NNT,鈥�2.25 [95% CI, 1.56-7.13]) (Table 2 and Figure 2B).

At weeks 2 and 6, we observed significant time鈥壝椻�塯roup interactions for the PANSS total score (F_1,60鈥�=鈥�19.51 and F_1,47.13鈥�=鈥�18.80, respectively [P鈥�<鈥�.001]) (eFigure 2 in Supplement 2), PANSS positive score (F_1,60鈥�=鈥�11.30 and F_1,48.53鈥�=鈥�11.63, respectively [P鈥�=鈥�.001]), PANSS negative score (F_1,60鈥�=鈥�12.00 and F_1,46.47鈥�=鈥�13.98, respectively [P鈥�<鈥�.001]), PANSS general score (F_1,60鈥�=鈥�14.13 [P鈥�<鈥�.001] and F_1,47.74鈥�=鈥�9.54 [P鈥�=鈥�.003], respectively), Hamilton Anxiety Rating Scale (F_1,60鈥�=鈥�5.09 and F_1,46.24鈥�=鈥�5.21, respectively [P鈥�=鈥�.03]), and Hamilton Depression Rating Scale (F_1,60鈥�=鈥�5.97 [P鈥�=鈥�.02] and F_1,42.15鈥�=鈥�7.39 [P鈥�=鈥�.009], respectively) (Table 2). Post hoc tests revealed that patients receiving active rTMS showed greater reductions in these scale scores than the sham treatment group at week 2 and week 6 (Table 2). We additionally defined the response to rTMS treatment as a reduction of 25% or more from baseline in the PANSS total score (for detailed results, see eResults 1 and eFigure 2 in Supplement 2). The missing data analysis and sensitivity analysis were detailed in eMethods 4, eResults 2, and eTable 1 in Supplement 2.

Model 1 did not exhibit a significant ability to estimate outcome (R²鈥�=鈥�0.34; F鈥�=鈥�2.10; P鈥�=鈥�.09). The reduction in AHRS scores was positively correlated with e-field strength within the AVH network (r鈥�=鈥�0.54; P鈥�=鈥�.001) (Figure 3) but not with e-field strength within the target region (r鈥�=鈥�0.14; P鈥�=鈥�.44). Therefore, the e-field strength within the AVH network was added into model 2 as a variable of estimation, while the e-field strength within the target region was not included. Model 2 exhibited a significant ability to estimate outcomes (R²鈥�=鈥�0.57; F鈥�=鈥�4.48; P鈥�=鈥�.003). The e-field strength within the AVH network significantly enhanced the ability of the model to estimate outcomes (change in R²鈥�=鈥�0.23; F鈥�=鈥�12.80; P鈥�=鈥�.002; B鈥�=鈥�3.12; t鈥�=鈥�3.58; P鈥�=鈥�.002). Detailed regression coefficients of hierarchical multiple regression are given in eTable 2 in Supplement 2. Additionally, analysis of the ability of the e-field strength to estimate outcomes in the sham treatment group is detailed in the eResults 3 and eFigure 3 in Supplement 2. To test the robustness of the estimation findings, we reconducted the hierarchical multiple regression analysis using different confounders (details in eResults 4 and eTables 3 and 4 in Supplement 2).

There was no significant difference in the rates of adverse events between the active rTMS and sham treatment groups throughout the study (eTable 5 in Supplement 2). The most common adverse event was sleepiness. All adverse effects were tolerable and gradually disappeared on cessation of treatment. No serious adverse events were reported.

Eighteen of 32 participants in the active rTMS group and 11 of 30 in the sham treatment group correctly guessed their actual group. There was no statistically significant difference between groups (蠂²鈥�=鈥�2.39; P鈥�=鈥�.12).

In this randomized clinical trial, we determined the effects of imaging-navigated rTMS for treating AVH in patients with schizophrenia. The treatment protocol was tolerable to patients, and no serious adverse events were reported. The primary outcome, hallucination symptoms, showed a significantly higher decrease in the active treatment group than in the sham treatment group after 2 weeks of treatment, which persisted to week 6. The e-field strength on the AVH network was independently associated with AVH outcomes and was specific to active stimulations compared with sham stimulations.

Although meta-analysis has suggested that rTMS is a promising treatment for AVH in patients with schizophrenia,^11,43 there has been considerable variation in the stimulation parameters across different studies. We did not know how to establish an effective parameter. Clinical guidelines also defined the efficacy at a 鈥減ossible level.鈥�⁴⁴ In a pilot open-label study,¹³ our group showed the possible efficiency of an MRI-navigated protocol for treating AVH symptoms. This efficiency was further validated by the current sham-control clinical trial. Similarly to the pilot study (response ratio鈥塷f鈥�37.5%),¹³ we found that 47% (15 of 32) of patients were responsive to active stimulations. This high response ratio may be attributed to precise guiding and online monitoring of the coil position on the left TPJ. This is consistent with the positive findings in fMRI-guided studies when stimulating areas around the left TPJ.^14,45 However, in studies where the target was not restricted around the TPJ, despite precise localizations, no significant efficacy was found compared with a placebo effect.^34,46 This is consistent with the recently identified schizophrenia network synthesized from 143 structural and functional MRI studies.⁴⁷ This network highlighted 2 superficial cortical areas within the stimulation depth of rTMS. One was the TPJ. The other was the dorsal prefrontal cortex (DLPFC), which was confirmed by another clinical trial.⁴⁸ Future studies may adopt the precise locations of the TPJ and DLPFC in these studies^13,48 and perform dual-site trials to enhance clinical efficacy.

The beneficial effects of treatment were not limited to AVH. We also found significant active rTMS effects on PANSS total and subscale scores. This is consistent with the findings of the previous open-label study.¹³ This improvement can be attributed to 2 factors. First, AVH is one of the symptoms evaluated in the PANSS. For some patients, AVH is even the dominant symptom. Thus, the PANSS scores naturally decreased with the alleviation of AVH. This phenomenon has also been found in a transcranial direct-current stimulation study.⁴⁹ Second, this stimulation protocol designed for AVH also modulated the neural correlates of other schizophrenia symptoms because the left TPJ is a hub region of AVH and also a part of the schizophrenia network.⁴⁷ The high connectivity between the TPJ and prefrontal cortex may also mediate the modulation effect to the DLPFC hub regions of negative symptoms.^50,51 However, the reduction in PANSS scores in the active group was below the minimum clinically meaningful difference.⁵²

In contrast to the significant findings at the group level, we noticed high outcome variability between individuals in the active group. This variability was probably related to the precise yet impersonalized stimulation setting with a one-fits-all target. We selected the TPJ as the stimulation target based on previous group-level findings^13,47 but did not consider the function variability of the TPJ among individuals.^24,53 The same neuroanatomical site (eg, TPJ) of different persons may be involved in different functional networks.⁵⁴ A unified TPJ site cannot represent AVH-related regions in all patients. Another parameter that needs to be controlled for each patient is the e-field pattern.⁵⁵ TMS manipulates neural excitability by changing the local e-field. In the brain, the e-field is reshaped by individual tissue distributions, such as the coil-to-cortex distances and cortical gyrifications. As a result, TMS may lose its focality and modulate cortices outside the predefined target.²⁸ In hierarchical multiple regression analysis, we considered these 2 factors individually to identify factors associated with outcome variability. We found that the individual e-field strength of the individualized AVH network, rather than the TPJ target, was associated with AVH improvement. This is consistent with other transcranial electric stimulation studies, which reported that e-field distributions within different brain networks were associated with particular symptoms and cognitive outcomes.^31,56

The strengths of our study include using image navigation to keep stimulations at the same site on different treatment days and identifying the interaction of the e-field and AVH network as a factor associated with symptom improvements. There are also some limitations to consider. First, this trial was conducted at a single center with a small sample size. A larger multisite trial is required to replicate our findings and determine the effects of more clinical factors on treatment efficacy. Second, the rTMS effect on cognitive function was not tested. We did not include cognition measures in this study because the previous open-label study showed that the same stimulation protocol had no significant effect on cognitive function.¹³ Third, although we found that the therapeutic effects of the active group were still evident at the end of the study, a longer follow-up period is required to determine how long the treatment effects can be sustained. This information is vital for designing a continuous treatment schedule. Fourth, the resting motor threshold of each participant was measured only at baseline. Considering that the resting motor threshold varied significantly across days among participants receiving rTMS,⁵⁷ future studies should test the threshold frequently to adjust the intensity of rTMS to enhance treatment efficacy. Fifth, clinical raters were blinded to the treatment assignments, but we did not assess the blinding integrity. This absence may increase the risk of detection bias. Sixth, the dropout rate in the sham treatment group was higher than in the active group at follow-up. This discrepancy should be noted in interpreting the follow-up results, since the small sample size in the sham treatment group may reduce the statistical power.

In this randomized clinical trial, imaging-navigated rTMS on the TPJ significantly alleviated the AVH in patients with schizophrenia compared with sham treatment. The rTMS protocol was also effective for positive and negative symptoms. Future studies should consider the role of individual e-fields in individualized symptom networks to improve treatment efficacy.

Accepted for Publication: September 19, 2024.

Published: November 11, 2024. doi:10.1001/jamanetworkopen.2024.44215

Open Access: This is an open access article distributed under the terms of the CC-BY License. 漏 2024 Hua Q et al. 糖心vlog Open.

Corresponding Authors: Gong-Jun Ji, PhD, School of Mental Health and Psychological Sciences, Anhui Medical University, 81 Meishan Rd, Hefei 230032, China (jigongjun@163.com); Kai Wang, MD, Department of Neurology, The First Affiliated Hospital of Anhui Medical University, 218 Jixi Rd, Hefei 230022, China (wangkai1964@126.com).

Author Contributions: Dr Ji had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Hua, L. Wang, and He contributed equally to this work.

Concept and design: L. Wang, K. Wang, Ji.

Acquisition, analysis, or interpretation of data: Hua, L. Wang, He, Sun, Xu, Zhang, Tian, K. Wang.

Drafting of the manuscript: Hua, L. Wang.

Critical review of the manuscript for important intellectual content: Hua, He, Sun, Xu, Zhang, Tian, K. Wang, Ji.

Statistical analysis: Hua, Sun.

Obtained funding: He, Tian, K. Wang.

Administrative, technical, or material support: Xu.

Supervision: L. Wang, K. Wang, Ji.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grants 81971689 (Dr Ji), 82371507 (Dr Ji), 82090034 (Dr K. Wang), and 32071054 (Dr Tian) from the National Natural Science Foundation of China; grant GXXT-2022-028 from the Collaborative Innovation Project Between Universities and the Hefei Comprehensive National Science Center (Dr Ji); grant 1808085J23 from the Science Fund for Distinguished Young Scholars of Anhui Province (Dr Tian); grant SYS2023B10 from the Key Laboratory of Philosophy and Social Science of Anhui Province on Adolescent Mental Health and Crisis Intelligence Intervention (Dr He); the Hefei Comprehensive National Science Center Hefei Brain Project (Dr K. Wang.); grant 202104j07020033 from the 2021 Anhui Province Key R&D Project: Population Health Special Project (Dr K. Wang); grant 2020zhyx-A04 from the major project of the Research Fund of Anhui Institute of Translational Medicine in 2020 (Dr K. Wang); grant 202204295107020006 from the Anhui Province Clinical Medical Research Transformation Special Project (Dr K. Wang); and grant 1565 from the Foundation for the Cultivation of Doctoral Research Talents (Prof L. Wang).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We thank the participants for taking part in this study and the Information Science Laboratory Center of the University of Science and Technology of China for the magnetic resonance imaging measurement services.