Neural encoding of native and non-native speech sounds in quiet and in background noise

Background The neural encoding of voicing in speech sounds has been relatively well studied in monolingual, native speakers. Little research, however, has examined neural encoding of aspiration feature in speech sounds or focused on encoding of non-native phonetic features or the effect of noise on processing these features. Purpose This study examined Auditory Evoked Potentials (AEPs) to bilabial stops with English versus Hindi phonetic properties of aspiration and voicing Hindi, English and Tamil listeners, in quiet and in noise. Method A total of 48 participants (16 Hindi, 16 American English, and 16 native Tamil native speakers) between 20 and 45 years of age participated. Natural digitized speech sounds including Hindi /ba/, /pa/, and /p h a/ and American English /ba/, and /pa/ were presented at 70 dB SPL using insert earphones in quiet, and in background noise at signal-to-noise ratio of 0. AEP peaks P1, N1, P2, and N2 were measured at the central electrode site (FCz). Results The P1 and P2 peak amplitudes were significantly larger for Hindi CV stimuli in Hindi participants relative to American English participants and Tamil participants. The morphology to Hindi /p h a/ was similar to English /pa/, and the morphology for Hindi /pa/ was similar to English /ba/. P1 amplitudes were larger and P2 amplitudes were smaller in noise relative to quiet. N2 peak latency in response to Hindi /pa/ was slightly longer relative to Hindi /p h a/ in American English listeners. Conclusion The findings add evidence to the Automatic Speech Perception model by observing cross-linguistic differences in P1 and P2. The results contribute to a better understanding of neural encoding in the cortex across native and non-native listeners, and how noise modulates early stages of processing.