Poster #: 54
Topic: Error signals
Friday, Sep 11, 2015
Does primacy bias in mismatch negativity (MMN) diminish with repeated exposure to sound sequences?
1Psychology, University of Newcastle, Newcastle, Australia
The ability of the auditory system to detect patterning in sound sequences enables the system to predict the most likely characteristics of sound in the environment whilst remaining sensitive to sounds that deviate from predictions. If a sound deviates from an active ‘prediction model’, an evoked-potential component called mismatch-negativity (MMN) is elicited. MMN reflects a prediction-error when a discrepancy between inferred and actual sound properties occurs and is thought to be confidence-weighted where the higher the confidence (i.e. when patterns are stable), the larger the MMN following pattern violation. Using the ‘multi-timescale’ paradigm, we have consistently shown that MMN is susceptible to order-driven bias dependent on initial tone roles. In this study we show that the bias remains even with repeated exposure to sound sequences. Participants heard four occurrences of either stable or unstable sequences over headphones. Both sequences contained 60ms and 30ms tones that alternated the role of standard (p = .875) and deviant (p = .125). In stable sequences, tone roles alternated every 2.4min (480 tones per block; 420 standard tones, 60 deviant tones). In unstable sequences, tone roles alternated every 0.8min (160 tones per block; 140 standard tones, 20 deviant tones). Results were consistent with primacy-bias: More confident predictions for stimulus configurations matching the one first encountered than the reversed one. Remarkably, first-deviant MMN remained large while second-deviant MMN reduced with repeated presentation. Rather than diminishing, the primacy bias pattern appears to intensify with repeated sequence encounters.