Repetition-dependent adaptation and Prediction error signalling in Schizophrenia patients with auditory hallucinations: A Roving Mismatch Negativity Study

ABSTRACT BackgroundPrediction error is the surprise that is elicited when the sensory expectations are first established and then violated. Positive symptoms of schizophrenia, like auditory hallucinations (AH), are thought to arise from dysregulated prediction error-signalling. Roving mismatch negativity (rMMN) is a unique event-related potential (ERP) based assessment that allows examination of repetition-dependent adaptation and deviance detection—complimentary processes that are integral to prediction-error signalling. In the rMMN paradigm, the deviant tone becomes the new standard with several repetitions. Also, the number of repetitions of the standard stimuli keeps changing throughout the experiment; longer repetitions yield a more positive ERP response; this phenomenon is Repetition Positivity (RP). Longer repetitions of standard stimuli elicit stronger deviance detection when interrupted, and this is called deviant negativity (DN). A difference waveform between RP and DN reflects the strength of prediction error signalling, the mismatch negativity (MMN). MethodsTwenty-three schizophrenia patients with auditory hallucinations (SZ-AH) and twenty-three healthy controls (HC) underwent rMMN assessment. Standard stimuli were repeated in sets of 3, 8 and 33. The first tone of the succeeding set became the deviant for the preceding set, yielding three components for RP (RP3, RP8, RP33), DN (DN3, DN8, DN33), and MMN (MMN3, MMN8, MMN33). Amplitudes and latencies of these components were compared between SZ-AH and HC. We also looked for potential associations between rMMN indices (amplitudes and latencies) and clinical scores in SZ-AVH. We examined the correlation between the RP-DN pair for all three repetition sets (3, 8, 33). ResultsSZ-AH had suppressed DN (DN3, DN8, DN33) and MMN33 amplitudes in comparison to HC. However, none of the RP amplitudes were diminished. Only MMN33 latency was significantly longer in SZ-AH than in HC. Amplitudes and latencies associated with repetition set of 8 showed a significant correlation with the frequency and loudness of AH. HC showed a strong positive correlation between RP-DN pairs; SZ-AH did not, except for the RP33-DN33 pair. DiscussionThe link between repetition-dependent sensory adaptation and deviance detection is compromised in SZ-AH patients. Though RP profile (RP3, RP8, RP33) is unimpaired in SZ-AVH, it is potentially due to AH interfering with auditory information processing and not because of intact short-term plasticity of the echoic memory trace.