THE EFFECTS OF VOICE FAMILIARITY ON PERCEPTION AND NEURAL RESPONSE

When masking noise is present, familiar voices are more intelligible than strangers’ voices. However, the mechanisms underlying this familiarity benefit remain unclear. Notably, familiarity can aid not only in attending to a familiar target voice, but also in ignoring a familiar distractor. This has been observed as both improved intelligibility and reduced working memory load when the masking voice is familiar. Such findings suggest that voice familiarity may enhance perceptual organization of the auditory scene, for example by improving segregation of competing voices, rather than solely facilitating speech recognition via a familiar-voice template.

To date, this work has been largely behavioural, leaving open the question of whether neural tracking of speech differs for familiar versus unfamiliar voices. Here, we used EEG and decoding methods to examine neural coupling to two-talker mixtures when either the attended or ignored talker was naturally familiar.

We recruited eight long-term heterosexual couples (>1 year) who recorded speech stimuli in the laboratory. Participants later listened to mixtures of familiar and unfamiliar voices while EEG was recorded. In two-talker blocks, participants alternated attention between voices every two minutes; in single-talker blocks, they listened to individual voices, providing a baseline for comparisons. Comprehension questions followed each trial.

We find reliable decoding of selective attention across conditions. While decoding accuracy did not differ for attended voices as a function of familiarity, encoding models suggest differences in the processing of familiar and unfamiliar speech when unattended.