Sound-offset encoding is related to speech-in-noise perception at sentence level in older adults

Sound-onset and sound-offset encoding are critical aspects of auditory temporal processing with important roles in speech-in-noise perception. We investigated brain correlates of active-offset encoding in older adults and their relationship to real-world listening. A total of 34 adults aged 60-81 years underwent pure-tone audiometry and assessments of speech-in-noise ability at both word and sentence levels. EEG data were recorded while participants performed active duration discrimination of noise segments and during passive listening. Onset and offset responses were extracted from EEG data, and amplitude differences were analysed. We found robust onset and offset responses at a single-subject level. Onset responses were significantly reduced in the active condition compared to passive listening, whereas offset responses were significantly larger. This pattern suggests that offset encoding is more susceptible to attentional modulation, whereas onset responses may be less affected. We found that active sound-offset amplitudes were significantly correlated with sentence-in-noise performance, whereas no correlation was observed for word-in-noise performance. We suggest a role for offset analysis in parsing of the speech stream needed for speech segregation. The active EEG measure we use allows single-subject inference and is a potential clinical measure of auditory cognition relevant to speech-in-noise listening. KEY POINTS: Sound-offset responses can be robustly identified at the single-subject level using EEG. Sound-offset responses were larger in the active task than in passive listening, suggesting task dependency and indicating that attention may be involved in sound-offset encoding. Active sound-offset responses correlate with sentence-level speech-in-noise performance.