Mismatch Negativity: a tool for objective assessment of auditory discrimination

Abstract

This study assessed the reliability of the mismatch negativity (MMN) in individuals and investigated the effect of arousal on the MMN. A within-groups design was used by recording the MMN twice in 26 individuals across two arousal conditions (high and low arousal). Each participant was stimulated monaurally and responses were gathered using the Interacoustics Eclipse EP25 2-channel device. A standard stimulus of 1000 Hz and a deviant stimulus of 1500 Hz was used, and presented via Radioear IP30 insert earphones. Each test session was continued until 2000 trials of the stimuli had been obtained for both arousal conditions, with a deviant probability of 20%. For each participant, residual noise levels were calculated for the standard, deviant and difference waveforms in both arousal conditions. The Hotelling’s T² statistical analysis was applied to the trials comprising the difference waveforms of each participant alongside three criteria – (1) significant Hotelling’s T² p-value in the standard waveform (2) significant Hotelling’s T² p-value in the deviant waveform, and (3) residual noise <1µV in the difference waveform – in order to objectively identify the presence of a MMN. The subsequent analyses regarding an effect of arousal was only conducted using data from the participants where the Hotelling’s T² detected a response in both arousal conditions. For these participants, the amplitude and latency of the MMN and P300 responses were analysed. The results indicate that upon meeting a-priori residual noise criteria the MMN can be reliably detected in individuals. The mean MMN amplitude was not statistically significantly different between high arousal (1.71 µV ± 0.98 µV) and low arousal (1.03 µV ± 0.99 µV). Similarly, the mean MMN latency was not statistically significantly different between high arousal (136.22 ms ± 26.03 ms) and low arousal (145.78 ms ± 24.97 ms). The mean P300 amplitude in high arousal (1.90 µV ± 0.68 µV) was statistically significantly lower ( p = 0.025) than that in low arousal (3.25 µV ± 1.68 µV). However, the mean P300 latency in high arousal (227.56 ms ± 23.89 ms) was not statistically significantly different from that in low arousal (237.78 ms ± 28.65 ms).