Decoded fMRI neurofeedback for auditory verbal hallucinations in schizophrenia

Auditory verbal hallucinations (AVH) are among the most distressing and common symptoms of psychotic disorders, including schizophrenia. Persistent AVH predicts poor outcomes including suicide. Antipsychotic drugs, the primary treatment for AVH, are ineffective in ~1/3 of patients and often abandoned due to severe side effects related to off-target drug effects. Thus, there is an urgent need to develop new treatments that are more selective and better tolerated—and ideally personalized. Decoded neurofeedback (DecNef) is a novel, well-tolerated fMRI neurofeedback technique that allows individuals to modify their brain activation patterns in real time, through implicit trial-and-error learning, based solely on feedback indicating how similar their current activation pattern is to a target pattern. DecNef could thus provide a novel therapeutic avenue for selectively reshaping local brain-activation patterns associated with AVH, thereby decreasing AVH symptoms. The intermittent nature of AVH enabled “symptom-capture” functional-neuroimaging studies by our group and others that consistently showed increased activation in speech-selective regions of auditory association cortex concurrent with AVH events. These studies motivated neuromodulation interventions—using tDCS, rTMS, and conventional real-time fMRI neurofeedback—aiming to reduce activation or excitability at a coarse regional level, which had limited success. Our recent data shows that more granular activation patterns within speech- selective auditory cortex distinguish AVH from non-AVH silent events, and from speech-evoked responses, and do so in a subject-specific manner—with activation patterns being more informative than overall activation. Among existing neuromodulation tools, DecNef is uniquely suited to selectively target these granular within- region activation patterns relevant to AVH, and to do so using personalized targets. Given this, the proposed two-phase project aims to test personalized DecNef training for AVH, first evaluating target engagement and tolerability (R61) and then clinical benefit for AVH amelioration (R33). Individuals with schizophrenia with treatment-resistant AVH will undergo fMRI DecNef training based on subject-specific active (non-AVH) or control (speech location) activation patterns. Go/No-Go criteria for the R61 will include significant within-subject engagement of the target pattern with active DecNef training (with greater expression than in the control group), preliminary evidence for AVH improvement, and a low discontinuation rate (<25%). In the R33, patients will undergo a double-blind, controlled (active DecNef vs. control DecNef) parallel randomized clinical trial evaluating efficacy for AVH amelioration. Optimal dosing (number of sessions) will be informed by R61 data. We will also evaluate the relationship between target engagement and AVH improvement (target validation). We expect this work to provide an initial validation of personalized DecNef as a novel, safe, and selective circuit intervention for AVH in schizophrenia, and to pave the way for a broader use of this technique for psychosis across neuropsychiatric disorders—and for severe mental illness more generally. Project Number: 1R61MH143195-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Mental Health (NIMH) | Principal Investigator: Guillermo Horga (+2 co-PIs) | Institution: COLUMBIA UNIVERSITY HEALTH SCIENCES, NEW YORK, NY | Award Amount: $1,542,242 | Activity Code: R61 | Study Section: Special Emphasis Panel[ZRG1 CN-U (82)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11293263