11th Speech in Noise Workshop, 10-11 January 2019, Ghent, BE

Quantifying listening effort: who tells what?

Annelies Bockstael(a)
Université de Montréal, Canada

Thomas Koelewijn(b)
VU University Medical Center, The Netherlands

Aarlenne Khan
Université de Montréal, Canada

Albert Guillemette
Institut Universitair de gériatrie de Montréal, Canada

Jean-Pierre Gagné
Université de Montréal, Canada

(a) Presenting
(b) Attending

Listening effort becomes more integrated in the assessment of speech comprehension in noise. Measuring listening effort assesses the cognitive resources and exertion related to achieve a certain level of speech intelligibility, and is as such a complementary parameter to speech intelligibility scores (i.e. the number of syllables/words/sentences identified correctly) to quantify speech comprehension.
Listening effort is measured indirectly and therefore challenging, and different measurement paradigms exist. In general, listening effort can be assessed behaviorally, physiologically and through self-report. The different paradigms (to a certain extent) assess different aspects of listening effort, and each of them has its proper strengths and advantages. To-date it is unclear how the outcome from different paradigms should be combined into an overall view on listening effort.

This project investigates how listening effort measured by a dual task paradigm (behavioral measure) and by pupillometry (physiological measure) correlate, and how these two paradigms can be combined to quantify listening effort efficiently and effectively. Outcomes of the dual-task paradigm and pupillometry are both well-established measures for listening effort, however they have never been measured in parallel. In the dual task paradigm, participants are asked to perform two task simultaneously, a primary speech-in-noise task and a secondary tactile task. In this project, speech-in-noise has been assessed for young normal-hearing participants for speech (sentences) presented in multitalker background noise at three signal-to-noise ratios: -5 dB, 0 dB and 5 dB. Speech intelligibility in noise has been carried out alone (single task) and in combination with the tactile task (dual task). Pupillometry has been carried out for all signal-to-noise ratios, and for both single and dual task conditions.

Pilot data from 9 participants show very promising results. For pupillometry, especially the slope of the pupil diameter varying over time appears to be sensitive to the listening conditions. Even with the small sample size, a significant increase in pupil diameter slope is seen during the listening phase when the sentences are presented, compared to the baseline conditions before and after stimulus presentation. In addition, this significant increase appears to be more pronounced in the dual task conditions when listening task and tactile task have to be carried out simultaneously, compared to the single task paradigm where only the speech intelligibility task is done.

Last modified 2018-12-08 00:23:30