The N1 event-related potential component as an index of speech sound encoding for multiple phonetic contrasts

Abstract

Listeners rely on many phonetic cues to perceive speech, but it is not clear how acoustic and phonological differences are encoded at early stages of perception. Previous work has begun to address this using the event-related potential (ERP) technique, demonstrating that the amplitude of the auditory N1 ERP component varies linearly with differences along VOT continua and suggesting that it can serve as an index of cue encoding. However, it is not clear how the N1 varies more generally for other phonetic distinctions. We present ERP data for a large set of naturally-produced word-initial minimal pairs spanning 18 consonants (/b,d,tʃ,f,g,dʒ,k,l,m,n,p,ɹ,s,ʃ,t,v,w,z/), as well as stimuli varying along voice onset time (voicing) and burst frequency (place of articulation) continua for all six stop consonants. The results reveal widespread differences in N1 amplitude for place of articulation and voicing. The N1 is larger for voiced consonants than voiceless consonants. Moreover, N1 amplitude as a function of place of articulation patterns differently for fricatives (larger N1 for alveolar and post-alveolar phonemes) than for stops (smaller N1 for these place of articulation categories). Overall, these results suggest that the N1 serves as a useful index of speech sound encoding across a range of phonetic contrasts.