44. Parsing elements of auditory predictive coding in schizophrenia with roving standard mismatch negativity

Mismatch negativity is an ERP reflecting auditory predictive coding. MMN is comprised of two components: 1) response to the standard stimulus, reflecting expectations about incoming stimuli, and 2) response to the deviant stimulus, reflecting deviance detection and the prediction error signal. One or both of these components may contribute to abnormal MMN in schizophrenia (Sz). Roving standard paradigms are optimized to evaluate these components of MMN. Here, we evaluate components of auditory predictive coding in 36 people with Sz and 29 healthy controls using a roving standard paradigm. We examined ERPs to the 3rd, 8th, and 33rd standard and subsequent deviants with RM ANOVA. Change in ERP response across repetition conditions for standards and deviants were assessed with planned orthogonal contrasts. For deviants, overall ERP response was significantly reduced in Sz (p=0.04), and a marginal reduction in the change in ERP amplitude to deviants over repetition conditions was also evident (p=0.06). For standards, the groups did not significantly differ in terms of overall ERP response (p=0.14) or change over repetition conditions (p's>0.70). These data suggest abnormalities in predictive coding in Sz may be attributable to reduced deviance detection rather than inability to establish expectations about incoming stimuli.