Slide Slam R10
Cognitive Control Associated with Noise-Induced Changes in Spoken Language Complexity
Catherine Pham1, Elisabeth Karuza1; 1The Pennsylvania State University
To produce fluent and grammatically correct speech, speakers have been argued to engage in a monitoring process that involves some degree of cognitive control. The demand on cognitive control is likely to be even greater when speakers are subject to the additional burden of ignoring distracting auditory input in noisy environments. Moreover, under these circumstances, they may contend with the knowledge that listeners are also engaging in more effortful processing. Despite the prevalence of noise in everyday life, psycholinguistic research has largely focused on examining language production under optimal, relatively silent conditions. Those studies that have examined the effect of noise have largely done so at the acoustic level, centering on speakers’ tendency to increase the loudness of their speech in noisy environments. To address this gap, the present study (1) measures how noise affects higher-level, non-acoustic properties of language production; and (2) investigates the extent to which these noise-induced changes are modulated by individual differences in cognitive control. Remotely through videoconferencing software, participants completed a picture description task during which they described a set of images under noise and silence conditions. Crucially, both the speaker (the participant) and their listener (the experimenter) were simultaneously exposed (or not exposed in the case of the silence condition) to the background noise, which consisted of multi-taker babble. Following the picture description task, participants completed a series of cognitive control tasks, each intended to tap into different aspects of cognitive control: the AX-Continuous Performance Task (AX-CPT), Flanker, and counting Stroop. Overall, background noise strongly affected language production, with speakers reducing the number of clauses, words, unfilled pauses, filled pauses, and mazes produced in noise relative to silence. Although performance on the AX-CPT and Flanker tasks was not significantly associated with noise-induced changes, performance on the Stroop task was correlated with the reduction of clauses, words, and unfilled pauses. Individuals who exhibited a greater reaction time interference effect on the Stroop (i.e., weaker cognitive control) produced fewer clauses, words, and unfilled pauses in noise. We consider this specific set of reductions to be speaker-oriented modifications implemented as a means of alleviating the cognitive burden of speaking in noise. However, we suggest that the reduction of filled pauses and mazes in noise, neither of which was correlated with cognitive control, to be listener-oriented modifications implemented as a means of facilitating comprehension for the sake of the listener.