MULTIMODHAL, a multicenter phase 3 randomized controlled trial: fMRI-based symptom capture for guiding rTMS treatment of drug-resistant hallucinations in patients with schizophrenia

Abstract Background Auditory–verbal hallucinations (AVHs) often persist despite optimized antipsychotic regimens. While conventional low-frequency repetitive transcranial magnetic stimulation (rTMS) at the T3P3 scalp site shows limited efficacy due to interindividual variability in AVH-related brain networks, the MULTIMODHAL trial assessed whether fMRI-guided rTMS is superior to conventional targeting for drug-resistant AVHs. Methods This multicenter, double-blind, controlled trial enrolled 73 adults with drug-resistant AVHs, randomized to receive 10 sessions of active 1-Hz rTMS via either individualized fMRI-guided neuronavigation or conventional T3P3 localization. The primary outcome was the change in the Auditory Hallucination Rating Scale (AHRS) at one month post-treatment. Secondary outcomes included: patient-reported intensity and frequency, alternative severity scales, and AHRS follow-up at 3, 6, and 12 months. Mechanistic analyses assessed rTMS-induced electrical fields modeling, and baseline resting-state functional connectivity profiles. Results fMRI-guided rTMS yielded a greater reduction in AHRS scores at one month (mean difference, -5.43; 95% CI, -8.92 to -1.94), with sustained benefits at three and six months. The number-needed-to-treat for neuro-guided rTMS was 3.5. Clinical response was associated with greater E-field overlap with AVH-related networks rather than energy deposition or baseline functional connectivity. Both interventions were well-tolerated. Conclusions fMRI-guided neuronavigation significantly enhances rTMS efficacy for drug-resistant AVHs, with sustained benefits up to six months and a favorable safety profile. These findings emphasize that precise network engagement is more critical for neuromodulation success than stimulation intensity or resting-state connectivity. Study limitations include sample-size and progressive dropout after three months, warranting research on response predictors and maintenance strategies.