The effect of clear speech on dual-task measures of listening effort in varied noise levels

How Temporal Overlap in Task Processing and Acoustic Load Affect Listening Effort in Adults

Listening effort reflects the cognitive resources allocated to overcome obstacles in auditory processing. While acoustic degradation (e.g., noise, reverberation) is known to affect listening effort, little is known about how temporal overlap of task processing influences effortful listening in adults. This study examined behavioral and subjective listening effort in 26 adults using a sequential dual-task paradigm manipulating stimulus onset asynchrony (SOA; 0.05 s, 0.25 s, 1.5 s, single task), noise condition (silence, multi-talker babble at 0 dB SNR), and acoustic environment (anechoic, room effects of T30 = 0.49 s). Task difficulty via decreasing stimulus onset asynchrony (SOA) significantly increased word recognition response times and subjective effort ratings, even when speech intelligibility remained high (∼90%) . Noise effects emerged only at longer SOAs, but vanished at very short SOAs where capacity was exhausted, indicating that SOA is a stronger predictor of listening effort than noise exposure. Interestingly, mild reverberation reduced behavioral listening effort compared to anechoic conditions in silence.

Both fast and slow speech can increase listening effort and impair speech comprehension in young listeners.

Speech rate can vary under different communicative circumstances (e.g., listener demands, speaking style), and such variation can affect the cognitive effort required to process speech. The present study examined how different speech rates, including both fast and slow rates, affect listening effort and speech comprehension in normal-hearing young listeners using a dual-task paradigm. The results showed that the recognition accuracy of key content words decreased under fast conditions for both the single-task and dual-task groups. In contrast, the number of comprehension errors increased under both the fast and slow speech conditions in the dual-task group. This was found using an additional method of evaluating speech comprehension, demonstrating that linguistic and mnemonic processes required for successful speech comprehension can be impaired when listening to slow speech, especially under high cognitive load. The subjective effort ratings similarly increased for fast and slow speech, but this was shown only in the control group, whereas the dual-task group showed no changes in the ratings by speech rates. This suggests that the subjective awareness of listening difficulty was reduced in dual-tasking situations. This study highlights the importance of multi-level assessment of speech perception, especially in cognitively challenging situations.

Unique contributions of acoustic and cognitive factors to spatial release from masking

Spatial Release from Masking (SRM) refers to the improvement in speech intelligibility that occurs when target and masker sounds are spatially separated. While traditionally attributed to acoustic cues, such as interaural time and level differences, growing evidence indicates that non-acoustic factors also shape the magnitude of SRM. This study quantifies the relative contributions of acoustic cue sensitivity and cognitive abilities on SRM using behavioral speech-in-noise tasks paired with a comprehensive cognitive battery. Cognitive factors were assessed using the Trail Making Task (attention shifting and processing speed), Flanker Task (inhibitory control), Stroop Task (interference resolution), and a working memory task. Results show that acoustic cues robustly support SRM in predominantly energetic masking conditions; however, under informational masking, individual differences in cognitive performance significantly modulate SRM magnitude. Individuals with stronger inhibitory control, faster attentional switching, and higher selective auditory attentional capacity demonstrated greater SRM, even with identical acoustic cues. These findings highlight SRM as a joint outcome of peripheral auditory processing and central cognitive mechanisms. Understanding how these factors interact has implications for hearing-aid processing strategies, assessment of hearing impairment, and the design of spatial audio systems that better support real-world communication.

Combined Effects of Dysphonic Voice and Classroom Noise on Cognitive Load in Children's Word Recognition and Sentence Comprehension.

BACKGROUND Classroom listening often occurs under acoustically adverse conditions, where background noise and variability in teacher voice quality place substantial demands on children's speech processing and learning. AIM This study examined how classroom noise and dysphonic teacher voice affect children's word recognition and sentence comprehension, with attention to cognitive factors supporting performance under adverse conditions. METHODS Twenty-six children aged 8-12 years with normal hearing completed word recognition (WIPI) and sentence comprehension (TROG-2) tasks in a soundproof booth. Speech stimuli were presented in either a normal or dysphonic voice with classroom noise at -8 and -12 dB SNRs. Accuracy, self-rated listening effort, and response time were measured. Working memory (WM) and inhibitory control (IC) were assessed once per child. RESULTS Dysphonic voice and lower SNR reduced accuracy and increased effort and response times. Recognition was more vulnerable to poor SNR, whereas comprehension was more affected by dysphonic voice. Higher WM supported recognition in degraded voice conditions, while stronger IC was associated with increased effort. Older children showed faster comprehension responses. CONCLUSIONS Classroom noise and dysphonic voice increase cognitive demands that hinder listening. Differential contributions of WM and IC highlight the importance of addressing both classroom acoustics and teacher vocal health to support equitable learning.

Intentional encoding does not alter the trade-off between speech-in-noise perception and memory.

Degraded speech increases listening effort at the expense of memory encoding, but the flexibility of this resource trade-off under varying motivational conditions remains poorly understood. The present study investigated how encoding intention influences cognitive resource allocation during a speech-in-noise task. Participants' sentence comprehension was tested using a sentence verification task (SVT) under varying signal-to-noise ratios (SNRs). Memory of the sentences was then assessed by a recognition task about which they were either naïve or forewarned. Accuracy in the SVT and memory task represent comprehension and recognition, respectively. Reaction time in the SVT serves as a proxy for listening effort. Increasing noise impaired both comprehension and recognition and required greater listening effort. Forewarning imposed an additional listening effort cost to comprehension, consistent with encoding intention functioning as an upfront cognitive investment. However, noise and forewarning each imposed independent, additive penalties in both comprehension and listening effort rather than a compounding interaction, and gains in recognition were minimal. These findings suggest that intentional encoding does not fundamentally alter the comprehension-memory trade-off during effortful listening. Implications for finite resource models and their assumptions about how motivational and perceptual demands compete for cognitive resources are discussed.

Effect of Spatial Release from Masking on Listening Effort in Different Semantic Contexts