Electrophysiological responses reveal a dedicated learning mechanism to process salient consonant sounds in human newborns

Isolating relevant sounds in the auditory stream is a crucial feature accomplished by human infants and a pivotal ability for language acquisition. Therefore, it is reasonable to postulate the existence of early mechanisms reorienting attention toward salient acoustic stimuli. Previous studies suggest that infants consider consonant sounds as more salient than dissonant ones, because the former resemble human vocalizations. However, systematic evidence investigating the neural processes underlying consonance tuning in newborns is still scarce. Here, we investigate newborns ability to recognize and learn salient auditory stimuli by collecting Mismatch Responses (MMRs) to consonant and dissonant sounds and by computing the trial-by-trial correlation of the neural signal with Bayesian Surprise (a theoretical measure of learning). We present 22 healthy newborns (40.4 {+/-} 15.8 hours) with a pseudo-random sequence of deviant and standard auditory events, while we record their electroencephalogram. Our results show that newborns exhibit a neural encoding of auditory regularities for all sound types (consonant and dissonant), as demonstrated by the presence of MMRs and significant correlation of the neural signal with Bayesian Surprise. Furthermore, consonant and dissonant sounds elicited MMRs and correlations with Bayesian Surprise of opposite polarities, with consonant auditory stimulation evoking negative responses, reminiscent of an adult-like MMR. Overall, our findings suggest that newborns display a dedicated perceptual learning mechanism for salient consonant sounds. We speculate that this mechanism might represent an evolutionary-achieved neural tuning to detect and learn salient auditory stimuli with acoustic features resembling human vocalizations. SIGNIFICANCE STATEMENTDiscriminating salient sounds in noisy sensory streams is a fundamental ability displayed by human infants, pivotal for acquiring crucial skills including language. Our study shed light on this ability by: (1) investigating perceptual learning mechanisms in newborns with a neurocomputational approach; (2) exploring the role of salient consonant sounds in modulating such mechanisms. Since human vocalizations are often consonant, the presence of a mechanism dedicated to enhance the processing of consonant sounds in newborns would confer evolutionary advantages. Our findings, indicating that newborns possess a dedicated and more refined perceptual learning mechanism to process consonance, corroborates this hypothesis. We speculate that this neural mechanism might facilitate the identification of salient acoustic input and support language acquisition in early infancy.