Is Recognition of Speech in Noise Related to Memory Disruption Caused by Irrelevant Sound?

Abstract

Listeners with normal audiometric thresholds show substantial variability in their ability to understand speech in noise (SiN). These individual differences have been reported to be associated with a range of auditory and cognitive abilities. The present study addresses the association between SiN processing and the individual susceptibility of short-term memory to auditory distraction (i.e., the irrelevant sound effect [ISE]). In a sample of 67 young adult participants with normal audiometric thresholds, we measured speech recognition performance in a spatial listening task with two interfering talkers (speech-in-speech identification), audiometric thresholds, binaural sensitivity to the temporal fine structure (interaural phase differences [IPD]), serial memory with and without interfering talkers, and self-reported noise sensitivity. Speech-in-speech processing was not significantly associated with the ISE. The most important predictors of high speech-in-speech recognition performance were a large short-term memory span, low IPD thresholds, bilaterally symmetrical audiometric thresholds, and low individual noise sensitivity. Surprisingly, the susceptibility of short-term memory to irrelevant sound accounted for a substantially smaller amount of variance in speech-in-speech processing than the nondisrupted short-term memory capacity. The data confirm the role of binaural sensitivity to the temporal fine structure, although its association to SiN recognition was weaker than in some previous studies. The inverse association between self-reported noise sensitivity and SiN processing deserves further investigation.

Keywords: binaural temporal fine structure sensitivity; irrelevant sound effect; noise sensitivity; speech perception in noise; variable importance measures; working memory.

Figure 1.

Distribution of individual better-ear pure tone average thresholds between 250 Hz and 4 kHz (PTA_BE, left panel), and the average bilateral asymmetry of the thresholds at the same octave frequencies (HL_diff, right panel).

Figure 2.

Mean audiometric thresholds at octave frequencies between 125 and 4000 Hz. Squares: left ear. Triangles: right ear. Error bars show plus and minus one standard error of the mean (SEM) of the 67 individual values.

Figure 3.

Distribution of individual speech-in-speech recognition scores (SRS; average proportion of correctly identified words in the OLSA matrix test) in the simulated cocktail-party listening task with two competing talkers.

Figure 4.

Mean proportion correct in the serial recall task, as a function of item type (auditory or visual), Block (1 or 2), and sound condition (open squares: in quiet. Triangles: in the presence of two competing talkers). Error bars show ± 1 SEM.

Figure 5.

Distribution of the irrelevant sound effect for auditory (ISE_Aud, upper panel) and visual (ISE_Vis, lower panel) item lists, expressed as the reduction in the proportion of correct responses in the serial recall task under irrelevant speech, relative to the in-quiet condition.

Figure 6.

Distribution of short-term memory performance (PC_STM; average proportion of correctly recalled digits in the condition without interfering speech), averaged across auditorily and visually presented item lists.

Figure 7.

Distribution of the thresholds for detecting interaural phase delays (IPD_th).

Figure 8.

Distribution of the noise sensitivity score (NSS; total item score on the questionnaire by Zimmer and Ellermeier, 1998).

Figure 9.

Best subset selection. Histogram of the mean squared prediction error (MSPE) across the 256 subset models, averaged for each subset model across the 100 runs of four-fold cross validation. Upper panel: models not containing the predictor PC_STM. Lower panel: models containing PC_STM. Blue and orange histograms represent models not containing or containing the predictor NSS, respectively. The vertical blue line marks the fifth percentile. Models with an MSPE below this value are displayed in Table 2.

Figure 10.

Comparison of dominance analysis and best subset selection results. Black circles: GDW computed for the full regression model (Table 1). Blue squares: p_BSS from cross-validated best subset selection (Table 2). Same predictor names as in Table 1.