Covert imitation acts as an adaptive reflexive gain control system in non-native speech perception

Current accounts posit that heard speech is covertly imitated to optimise speech perception. Behaviourally, covert imitation is measured using the stimulus-response compatibility (SRC) paradigm. In a speech SRC task, participants produce a prompted speech syllable whilst perceiving a speech distractor that either matches the target response (compatible condition) or does not (incompatible condition). The degree to which the distractor is covertly imitated is captured by the automatic imitation effect, computed as the difference in response times (RTs) between incompatible and compatible trials. Two theoretical accounts disagree on whether whether covert imitation is enhanced when speech perception is challenging or instead when the speech signal is familiar to the listener. To dissociate between these two accounts, we conducted two experiments utilising the speech SRC task (Experiments 1 and 2) and Transcranial Magnetic Stimulation (TMS) (Experiment 2) to evaluate covert imitation of native speech sounds (perceptually easy and familiar) and of non-native speech sounds (perceptually challenging and unfamiliar). In Experiment 1, participants performed an SRC task with native and non-native sounds, before and after receiving perceptual identification training with the non-native sounds. Automatic imitation was enhanced for non-native sounds compared to native sounds and decreased to a greater extent for non-native sounds than for native sounds following perceptual training. In Experiment 2, participants passively perceived native and non-native speech sounds whilst receiving single pulses of TMS to the tongue area of the primary motor cortex (M1). Tongue M1 showed higher excitability during passive perception of non-native sounds. Our results highlight that covert imitation is enhanced during the perception of perceptually challenging, unfamiliar speech, suggesting that covert imitation acts as an adaptive reflexive gain control mechanism to optimise speech perception.