Authors: Saranya Mundayoor¹, Edward Lobarinas¹
¹The University of Texas at Dallas
Background: Changes in a talker’s voice can negatively impact speech perception, a phenomenon known as “multitalker processing cost.” While this has been studied using isolated words or digits, real-world communication involves full sentences. This study aimed to determine if this processing cost persists under more ecologically valid conditions (sentences) and to investigate the cognitive factors that might mediate this effect in normal-hearing adults.
Method: Twenty-five normal-hearing adults were tested on speech-in-noise tasks under two conditions: a single talker (low variability) and multiple different talkers presented in succession (high variability). Accuracy was measured by sentence repetition, while speed was assessed by the reaction time to identify the first word of each sentence. Participants also completed the NIH Toolbox Cognition Battery to assess crystallized and fluid cognitive domains.
Results: Both speech recognition accuracy and processing speed were significantly reduced in the high-variability condition compared to the low-variability condition. Correlation analysis revealed that crystallized cognition was moderately associated with higher accuracy across both conditions. Conversely, fluid cognition showed a significant moderate negative correlation with reaction time (indicating that higher fluid cognitive ability relates to faster processing) for both single- and multitalker scenarios.
Conclusion: Even in individuals with normal hearing, adapting to a constantly changing target talker significantly hinders sentence perception. These results suggest that clinical speech testing should incorporate greater talker variability to better reflect the communication challenges faced in everyday life.
