Cortical tracking of speech in quiet and noise in infants and children

Abstract

Past research has shown that in multi-talker listening situations, cortical tracking of the speech envelope is observed for the target speech stream in adults. While children are known to perform worse than adults under noisy conditions, many aspects of how the cortical processing of speech develops is unknown. Here, we recorded EEG responses to continuous speech in infants and 7-to-18-year-olds in Quiet, Co-located Noise, and Segregated Noise. The target speech stream consisted of infant directed speech or an audiobook for children. The noise consisted of four-talker babble and was constructed from four audiobooks read by 2 males and 2 females. Thirty-two channels of EEG data were recorded with the stimuli presented at an overall level of 65 dB SPL for children and 70 dB SPL for infants with a +5 dB target-to-noise SNR via speakers at 0°, +90°, and −90° azimuth. EEG signals were analyzed using the decoding model of the Multivariate Temporal Response Function toolbox to assess how well the stimulus envelope could be reconstructed from the recorded neural responses. Preliminary results show evidence of cortical tracking of the speech envelope for both infants and children in quiet and that cortical tracking degrades in both Co-located and Segregated Noise.