A Multimodal Approach to Measuring Listening Effort: A Systematic Review on the Effects of Auditory Task Demand on Physiological Measures and Their Relationship.

Only 28 studies were included and the review could not pool them statistically, so the findings are a description of patterns rather than an estimate of any effect. Most studies tested people with normal hearing, and only a few included hard-of-hearing participants, so the picture for the patients who struggle with effortful listening is thin. Many studies used easy listening conditions where people performed at 70% or better, and these were the ones that most often showed no physiological change, so the measures were never really pushed. More than half of the studies did not justify their sample size with a power analysis and had small samples, so some 'no effect' results may simply be studies too small to see anything. Several studies did not report enough detail about their methods, analyses or significance thresholds to be repeated. Not all measures were recorded at the same moment: saliva samples are taken once before, during and after a task, which is hard to line up with a continuous signal like pupil size. The summary of effects does not account for the possibility that each measure peaks at a different level of difficulty. The search was limited to English articles, and no review protocol was registered before the search.

The authors declared no conflicts of interest. The work was funded by ZonMw, a non-U.S. government funder.

Do not read the consistency of pupil dilation or skin conductance as evidence that any of these measures has been validated as a marker of listening effort in patients. The review found no best measure, correlations between measures were weak at best, and most studies tested young people with normal hearing on tasks they could already do well, so nothing here supports using a physiological reading to judge how effortful listening is for an individual patient.