S43. ASSESSING THE VARIABILITY OF RESPONSE TO NON-INVASIVE BRAIN STIMULATION IN PSYCHOSIS

BACKGROUND: Non-invasive brain stimulation has been introduced as add-on treatment for psychotic symptoms, not least because transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS) have a low risk profile. Yet, the initial excitement is wearing off and there is a lack of consistent evidence from randomized controlled trials (RCT) for the efficacy of brain stimulation. It is often claimed that this is because patients respond very differently to brain stimulation, with some benefiting much more than others. However, is there really strong evidence from RCTs that patients do respond differently? This question can be assessed by comparing the overall variability under active stimulation with the variability under sham stimulation across studies. METHODS: We included all double-blinded, sham-controlled RCTs and crossover studies of adults with a diagnosis of a schizophrenia spectrum disorder that used TMS or tDCS for the treatment of psychotic symptoms. We extracted a measure of variability (standard deviation, standard error, or confidence interval) of the primary outcome for active and sham stimulation, computed variance-weighted variability ratios for each study, and entered them into a random-effects model. In the case of individual differences in response to TMS or tDCS, we expected that the overall variability under active stimulation would be increased compared to sham stimulation (as evidenced by an overall variability ratio significantly larger than 1). RESULTS: A total of 39 RCTs and crossover trials with 1 352 patients were included. We found that the variability under active stimulation was not significantly larger than under sham stimulation (variability ratio = 1.07; 95% CI: 0.99, 1.16; P = 0.098). DISCUSSION: These results do not provide strong evidence for the presence of individual differences in response to non-invasive brain stimulation in patients with schizophrenia spectrum disorders. The search for subgroups and prognostic biomarkers may require more complex study designs including N-of-1 trials.