Reception Thresholds for Sentences in Quiet and Noise for Monolingual English and Bilingual Mandarin-English Listeners

Abstract

Bilingual (BL) listeners' difficulties in adverse noise conditions are exacerbated when perceiving their second language (L2) relative to their first language (L1). Perception of L2 is also significantly poorer by BL listeners compared to native monolingual (ML) listeners. The purpose of the study was to examine the effect of stationary and nonstationary energetic noise maskers on L1 and L2 speech perception in native and nonnative listeners. A mixed multivariate quasi-experimental design was employed. Two groups of 12 ML English-speaking and BL Mandarin-English-speaking normal-hearing young adult female volunteers participated. An adaptive technique was employed to determine reception thresholds for sentences (RTSs) in quiet and in backgrounds of competing continuous and interrupted noise. The noises differed only in their temporal continuity. The sentence stimuli employed consisted of the Hearing in Noise Test (HINT) and the Mandarin Hearing in Noise Test (MHINT). ML participants received the HINT stimuli while the BL participants received both HINT and MHINT stimuli. Between-group differences in RTSs were examined for the same stimuli (i.e., HINT) and for L1 stimuli (i.e., HINT vs. MHINT). Within-group differences in RTSs were examined with the BL participants' perception of L1 and L2 stimuli (i.e., MHINT vs. HINT). The amount of "release from masking" (i.e., the difference of RTS signal-to-noise ratios [SNRs] in interrupted and continuous noise) was also examined between and within groups. In quiet there was no significant difference in mean RTSs between the BL and ML participants with their respective L1 stimuli; MLs had significantly lower mean RTSs in English compared to the BLs; and mean RTSs for the BLs were significantly lower for L1 versus L2 stimulus. In noise, a significantly higher RTS SNR was found for the MLs in continuous noise but not interrupted noise for L1 stimuli compared to the BLs; BLs had a significantly higher mean RTSs in English compared to the MLs; and BLs had significantly higher mean RTSs for L2 versus L1 stimuli. The release from masking was significantly greater for MLs compared to BLs with their respective L1 stimuli and with the same English stimuli. There was no significant difference for the BLs' release from masking with L1 versus L2 stimulus. BL listeners display significantly poorer performance when perceiving nonnative L2 sentences in quiet and in continuous and interrupted noise relative ML listeners. When listening to their respective native L1 sentences, only a difference in continuous noise was found. This difference was attributed to differential masking effect on the English stimuli. Similar performance in the interrupted noise between the ML and BL participants with L1 stimuli and the equivalent release from masking with the BL participants for both L1 and L2 stimuli suggest comparable basic auditory temporal resolving capacities between these ethnic groups.