| Speech contains not only linguistic information but also rich non-linguistic information,such as the talker’s emotions and identity,with linguistic and non-linguistic information interacting with each other.Accurately recognizing the talker’s identity through their speech is an important social skill in interpersonal interactions.It has been found that listeners can more accurately recognize voices of their native language than that of a foreign language,a phenomenon known as the Language Familiarity Effect(LFE).Although the LFE in speech perception has been supported by behavioral evidence,its neural mechanisms remain unclear.Therefore,this study aims to explore in depth how the language familiarity effect is achieved at the neural level by combining behavioral and functional near-infrared imaging techniques.Thirty-eight Chinese-English bilinguals were recruited for our voice recognition task.The task was divided into two phases: learning and testing.During the learning phase,participants learned to associate the voice characteristics of five male speakers with their identities.In the testing phase,participants judged the speaker’s identity based on their spoken voice.The study manipulated the level of language familiarity in the recognition task: highly familiar(native language: Chinese),moderately familiar(English: secondlanguage),and unknown(a foreign language: Ewe).Behaviorally,the study collected accuracy for speaker recognition under the three language conditions in two stages.At the neural level,functional near-infrared spectroscopy was used to collect bilateral temporal lobe activation signals and analyze their activation patterns.The behavioral results showed that talker identification was most accurate when recognizing speakers in their native language,followed by the second language condition and then the unknown language condition.Identification was more accurate during the learning phase than the testing phase,and this difference was due to non-native language familiarity.There was no significant difference in accuracy between the second language and unfamiliar language conditions during the learning phase,but during the testing phase,identification of second language speakers was more accurate.Regarding brain activation,recognizing non-native speakers’ voices elicited stronger activation than recognizing native speakers’ voices,mainly in the left supramarginal gyrus,insula and angular gyrus,which are associated with language processing,as well as the right dorsolateral prefrontal cortex and angular gyrus,which are associated with working memory and information integration.Overall,speaker recognition in all three languages showed a left hemisphere dominance.However,by stage,recognizing native speakers’ voices showed right hemisphere dominance during the learning phase and left hemisphere dominance during the testing phase.Recognizing second language speakers’ voices showed left hemisphere dominance in both stages,while there was no hemisphere advantage for recognizing the unknown language speakers’ voices.During the testing phase of recognizing second language speakers’ voices,left hemisphere activation was stronger than during the learning phase.In summary,language familiarity has an impact on speaker recognition,and this effect produces corresponding changes in brain activation.Bilateral temporal lobe activation is reduced for the recognition of native voices,and language familiarity does not act directly on the brain area where the voice is recognized,but rather indirectly.Recognition of non-native voices requires greater activation of non-language related brain areas such as working memory and attentional selection to facilitate voice recognition in that language.This study reveals the neural basis for the modulation of speaker recognition by language familiarity,providing new insights and evidence for the impact of language ability on speaker recognition. |
