Listening Effort Through Depth of Processing in School-Age Children

A reliable and practical measure of listening effort is crucial
in the aural rehabilitation of children with communication disorders. In
this article, we propose a novel behavioral paradigm designed to measure
listening effort in school-age children based on different depths
and levels of verbal processing. The paradigm consists of a classic word
recognition task performed in quiet and in noise coupled to one of three
additional tasks asking the children to judge the color of simple pictures
or a certain semantic category of the presented words. The response
time (RT) from the categorization tasks is considered the primary indicator
of listening effort.
Design: The listening effort paradigm was evaluated in a group of 31 normal-
hearing, normal-developing children 7 to 12 years of age. A total of
146 Dutch nouns were selected for the experiment after surveying 14 local
Dutch-speaking children. Windows-based custom software was developed
to administer the behavioral paradigm from a conventional laptop computer.
A separate touch screen was used as a response interface to gather
the RT data from the participants. Verbal repetition of each presented word
was scored by the tester and a percentage-correct word recognition score
(WRS) was calculated for each condition. Randomized lists of target words
were presented in one of three signal to noise ratios (SNR) to examine the
effect of background noise on the two outcome measures of WRS and RT.
Three novel categorization tasks, each corresponding to a different depth
or elaboration level of semantic processing, were developed to examine the
effect of processing level on either WRS or RT. It was hypothesized that,
while listening effort as measured by RT would be affected by both noise
and processing level, WRS performance would be affected by changes in
noise level only. The RT measure was also hypothesized to increase more
from an increase in noise level in categorization conditions demanding a
deeper or more elaborate form of semantic processing.
Results: There was a significant effect of SNR level on school-age children’s
WRS: their word recognition performance tended to decrease
with increasing background noise level. However, depth of processing
did not seem to affect WRS. Moreover, a repeated-measure analysis
of variance fitted to transformed RT data revealed that this measure of
listening effort in normal-hearing school-age children was significantly
affected by both SNR level and the depth of semantic processing. There
was no significant interaction between noise level and the type of categorization
task with regard to RT.
Conclusions: The observed patterns of WRS and RT supported the
hypotheses regarding the effects of background noise and depth of
processing on word recognition performance and a behavioral measure
of listening effort. The magnitude of noise-induced change in RT did
not differ between categorization tasks, however. Our findings point to
future research directions regarding the potential effects of age, working
memory capacity, and cross-modality interaction when measuring
listening effort in different levels of semantic processing.