Poster #: 127
Topic: Speech and language (incl. deficits)
Wednesday, Sep 9, 2015
Omission responses in speech are differentially modulated by speaking rate and attention
1Biological Psychology, University of Leipzig, Leipzig, Germany
2Neuroscience, Max Planck Institute for Empirical Aesthetics, Frankfurt, Germany
Human brains closely monitor the surrounding acoustic scene: Omissions of anticipated sounds result in an automatic change detection response, the “omission mismatch negativity” (OMMN). Previous research has provided evidence for temporal constraints on the OMMN, with a maximum stimulus-onset-asynchrony (SOA) of 150 ms between actual (preceding) and anticipated (omitted) sound (cf. window of integration, TWI). However, given the rather flexible temporal structure of speech afforded to less discrete sound boundaries and differences in speaking rate, this TWI may actual differ between non-speech and speech stimuli. In this EEG-study (N=21, 64 electrodes), we compared the OMMN in response to word-final consonant omissions (La– from German Lachs ‘salmon’) across two levels of SOAs, 150 and 300 ms (corresponding to normal and slow speaking rate). Additionally, we contrasted a passive with an active condition. Results showed discernible OMMNs peaking between 120 and 170 post deviance onset. Mixed-effect models on mean amplitudes in this window stemming from central midline electrodes revealed an interaction of stimulus type (standard, deviant) and SOA, with greater OMMN amplitudes for the short than for the long SOA. This interaction further depended on presentation mode (active, passive), and was only visible for the passive presentation. Thus, both speaking rate and attention affect the OMMN, albeit in a differential way: Pre-attentively, an extension of the TWI seems to indeed cause a reduction in the OMMN response, similar to non-speech findings. Under attention, however, no fixed TWI appears to exist, suggesting that attentive listening is beneficial for dealing with slow speakers.