The Effects of Task Difficulty Predictability and Noise Reduction on Recall Performance and Pupil Dilation Responses

Abstract

Objectives: Communication requires cognitive processes which are not captured by traditional speech understanding tests. Under challenging listening situations, more working memory resources are needed to process speech, leaving fewer resources available for storage. The aim of the current study was to investigate the effect of task difficulty predictability, that is, knowing versus not knowing task difficulty in advance, and the effect of noise reduction on working memory resource allocation to processing and storage of speech heard in background noise. For this purpose, an “offline” behavioral measure, the Sentence-Final Word Identification and Recall (SWIR) test, and an “online” physiological measure, pupillometry, were combined. Moreover, the outcomes of the two measures were compared to investigate whether they reflect the same processes related to resource allocation. Design: Twenty-four experienced hearing aid users with moderate to moderately severe hearing loss participated in this study. The SWIR test and pupillometry were measured simultaneously with noise reduction in the test hearing aids activated and deactivated in a background noise composed of four-talker babble. The task of the SWIR test is to listen to lists of sentences, repeat the last word immediately after each sentence and recall the repeated words when the list is finished. The sentence baseline dilation, which is defined as the mean pupil dilation before each sentence, and task-evoked peak pupil dilation (PPD) were analyzed over the course of the lists. The task difficulty predictability was manipulated by including lists of three, five, and seven sentences. The test was conducted over two sessions, one during which the participants were informed about list length before each list (predictable task difficulty) and one during which they were not (unpredictable task difficulty). Results: The sentence baseline dilation was higher when task difficulty was unpredictable compared to predictable, except at the start of the list, where there was no difference. The PPD tended to be higher at the beginning of the list, this pattern being more prominent when task difficulty was unpredictable. Recall performance was better and sentence baseline dilation was higher when noise reduction was on, especially toward the end of longer lists. There was no effect of noise reduction on PPD. Conclusions: Task difficulty predictability did not have an effect on resource allocation, since recall performance was similar independently of whether task difficulty was predictable or unpredictable. The higher sentence baseline dilation when task difficulty was unpredictable likely reflected a difference in the recall strategy or higher degree of task engagement/alertness or arousal. Hence, pupillometry captured processes which the SWIR test does not capture. Noise reduction frees up resources to be used for storage of speech, which was reflected in the better recal...

A decrease in physiological arousal accompanied by stable behavioral performance reflects task habituation

“…The sentence baseline pupillary response was defined as the mean pupil dilation during a time interval of 1 s prior to the beginning of a sentence (Figure 1). This way of calculating the sentence baseline pupillary responses has been used in previous studies combining the SWIR test and pupillometry (Bönitz et al, 2021;Micula et al, 2021). These studies have also shown that the sentence baseline pupillary responses tend to increase over the course of several SWIR test sentences, reflecting the resources allocated to storing an increasing number of items for later recall.…”

“…A combination of behavioral and physiological measures has been used in order to investigate listening effort and cognitive resource allocation during listening tasks. Specifically, such studies often combine pupillometry with speech recognition or auditory recall test in order to investigate the effects of various manipulations, such as background noise and hearing aid signal processing (Wendt et al, 2017;Ohlenforst et al, 2017Ohlenforst et al, , 2018Zekveld et al, 2018;Bönitz et al, 2021;Micula et al, 2021). Additionally, pupillometry in combination with behavioral measures has also been used to track levels of arousal, in order to investigate the effects of time-on-task on engagement and fatigue (Hopstaken et al, 2015;Alhanbali et al, 2020;Ayasse and Wingfield, 2020).…”

“…This level is chosen in order to ensure that differences in SWIR test recall performance are due to differences in cognitive resource allocation rather than audibility in various conditions. In recent studies, the SWIR test has been combined with pupillometry in order to obtain behavioral and physiological indices of how various manipulations, such as noise reduction in hearing aids or task difficulty, affect cognitive resource allocation and listening effort (Bönitz et al, 2021;Micula et al, 2021). The SWIR test is a lengthy test, and the recall task and background noise make it cognitively demanding.…”

“…The SWIR test is a lengthy test, and the recall task and background noise make it cognitively demanding. Listeners with hearing loss, who may be more prone to experiencing increased listening effort (Pichora-Fuller et al, 2016), are often included in studies using the SWIR test (Ng et al, 2013(Ng et al, , 2015Micula et al, 2020Micula et al, , 2021. It could be speculated that at some point during a test session arousal may decrease due to the factors illustrated in the FUEL, such as displeasure with the task.…”

Listen Carefully protocol: an exploratory case–control study of the association between listening effort and cognitive function

“…Based on the framework for understanding efffortful listening, it is possible that fatigue, motivation and/or the evaluation of demands on available capacity may influence the allocation of effort during recall and thereby the pupil dilation response 17. Given previous research using older adults both with and without hearing loss, we expect a speech reception threshold at 80% to load working memory without overloading its capacity, making interindividual differences in effort apparent while maintaining attentional engagement 49 55–58…”