Time-efficient Methodology for Robustly Assessing Speech-related Mismatch Responses in Adults and Infants.

Abstract

The mismatch response (MMR) is a critical neural indicator of discrimination of speech contrasts. Using MMRs, previous research has demonstrated that language experience can affect MMRs, such that MMRs to native speech contrasts are different from ones to nonnative speech contrasts. This effect is observed as early as 11-12 months, but not at 6-7 months of age, indicating early learning of speech sounds. Yet, many challenges remain to use MMR to advance our understanding of speech learning especially in infants, including prolonged recording time, inefficient use of data, and a lack of reconciliation of MMR recorded using different technologies (i.e., EEG vs. magnetoencephalography [MEG]). Using an improved recording paradigm and analysis approaches, the current study addressed these challenges by examining (1) whether MEG-MMR is linked to well-established EEG-MMR in the same adults and (2) whether our methods capture the difference of MEG-MMR between native and nonnative speech contrasts in adults and (3) in older infants. Results from 18 adults with simultaneous M/EEG demonstrated a high correlation between the MEG-MMR and the EEG-MMR. Additionally, MEG-MMRs to native speech contrasts were different from ones to nonnative speech contrasts, replicating spatiotemporal patterns documented in existing literature. Finally, we replicated this effect in the MEG-MMR in 14 infants aged between 9 and 14 months using the same methods. These findings validate our new methodologies (less than 15 min) for acquiring and analyzing speech-related MMR across ages, paving the way for studying early language development, and improving early detection of language-related disorders.