This Title All WIREs
How to cite this WIREs title:
WIREs Cogn Sci
Impact Factor: 3.476

Listening effort: Are we measuring cognition or affect, or both?

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

Abstract Listening effort is increasingly recognized as a factor in communication, particularly for and with nonnative speakers, for the elderly, for individuals with hearing impairment and/or for those working in noise. However, as highlighted by McGarrigle et al., International Journal of Audiology, 2014, 53, 433–445, the term “listening effort” encompasses a wide variety of concepts, including the engagement and control of multiple possibly distinct neural systems for information processing, and the affective response to the expenditure of those resources in a given context. Thus, experimental or clinical methods intended to objectively quantify listening effort may ultimately reflect a complex interaction between the operations of one or more of those information processing systems, and/or the affective and motivational response to the demand on those systems. Here we examine theoretical, behavioral, and psychophysiological factors related to resolving the question of what we are measuring, and why, when we measure “listening effort.” This article is categorized under: Linguistics > Language in Mind and Brain Psychology > Theory and Methods Psychology > Attention Psychology > Emotion and Motivation
A schematic illustration of the proposed clear distinction between demanded effort (associated with task properties) and the two less easily distinguished (hence fuzzily bounded) behaviorally or physiologically measurable properties of exerted and (self‐)assessed effort. Demanded effort is quantifiable in terms of task properties such as SNR, number of competing talkers, memory load, degree of distortion of target speech, and so forth. Exerted and assessed effort are typically quantified using the behavioral and physiological measures discussed in the text. We also indicate a proposed distinction between externally and internally directed cognitive resources and associated these both mainly with exerted effort because, while they are often distinguished theoretically (see text), they may or may not be distinguishable in terms of how listeners feel about (assess) the effort they are exerting. SNR, signal‐to‐noise ratios
[ Normal View | Magnified View ]
Three models of speech perception illustrating the use of structural elements associated with listening effort. (a) Active processing incorporating aspects of both top‐down and bottom‐up components based on Heald and Nusbaum () in this model, a mental representation of the signal (i) is compared to a linguistic pattern representation (n) generated on the basis of expectations and prior knowledge. The comparison generates feedback to modify either/both the structure of the perceptual representation and the linguistic interpretation. For example, the contribution or weighting of particular acoustic features in the perceptual representation may be emphasized or de‐emphasized depending on their relative diagnosticity in a given context, while the hypothesized linguistic interpretation of the signal may be adjusted in response to changing awareness of the validity of expectations and prior knowledge. Demand on processing resources is incurred through the complexity of the comparison process and the possible need for multiple iterations of either or both feedback cycles when the initial perceptual representation is insufficiently determinative. Note, however, that this model makes no claims in itself as to the nature of the resources committed to either or both feedback loops, and it is entirely possible that both depend on the same capacity or capacities. (b) The original ELU model (J. Rönnberg et al., ). In this model, the incoming signal (which is explicitly multimodal) is initially converted into a phonological representation through automatic (i.e., not capacity‐demanding) processes in the rapid, automatic, and multimodally bound PHOnological (RAMBPHO) episodic buffer. If this representation is sufficiently well‐specified to match a linguistic representation (assumed to be syllabic in this case), then speech perception has occurred without the commitment of cognitive resources, that is, effortlessly. If, however, the phonological representation does not adequately match a single mental representation, then explicitly resource‐demanding mechanisms are engaged. These mechanisms include both those we might consider to be more signal‐oriented (such as shifting attentional focus and inhibition of irrelevant information), as well as those typically considered to involve higher levels of processing (such as inference and semantic integration), but the focus is primarily on the latter (J. Rönnberg et al., ). (c) The modified ELU model (based on an older representation) incorporating explicit top‐down control of early sensory processing (Edwards, ). Here, the model described in (b) is augmented with a second, explicit locus toward which effort may be applied, namely a perceptual level of “feature extraction, object formation, and selective attention.” Thus, in principle, the application of cognitive processing resources could alter the mental representation of the signal (in RAMBPHO) and, like the model shown in (a), there are two potentially distinct kinds of processes, one perceptual/signal‐oriented and one more cognitive/semantic. ELU, Ease of Language Understanding
[ Normal View | Magnified View ]

Browse by Topic

Psychology > Emotion and Motivation
Psychology > Attention
Psychology > Theory and Methods
Linguistics > Language in Mind and Brain

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts