Infant’s perception of speech in noise: Effect of the number of background talkers

Abstract

A number of recent studies have investigated infants’ abilities to hear speech in noise [Newman & Jusczyk, ‘‘The cocktail party effect in infants,’’ Percep. Psychophs. 58, 1145–1156 (1996); Newman, R. S., ‘‘The cocktail party effect in infants revisited: Listening to one’s name in noise,’’ Develop. Psychol. 41, 352–362 (2005); Barker & Newman, ‘‘Listen to your mother! The role of talker familiarity in infant streaming,’’ Cognition, 94, B45–B53 (2004); Hollich, et al., ‘‘Infants’ use of synchronized visual information to seperate streams of speech,’’ Child Develop. 76, 598–613 (2005)]. In the current study, we examine how the type of background noise influences infants’ ability to understand speech. Infants aged 5 months heard a talker repeat either their own name or another infant’s name in the presence of one of 3 types of background noise: multitalker babble, a single background talker, or a single talker reversed in time. With multitalker babble, infants could recognize their name (shown by longer listening to own-name than other-name test trials) only when the target speech was at least 10 dB more intense than the background. However, even at this high SNR, infants fail to recognize their name when the background consists of a single talker. This difficulty is not the result of semantic information or speechlike properties from the single talker; infants likewise fail when the single talker is temporally reversed. Rather, infants appear to be distracted by signals containing the time-varying amplitude patterns of a human voice, and do not appear to benefit from the dips in noise amplitude that occur with a single voice.