Benson Cheng Lin Hsu promoveerde op 3 juni 2019 aan de KULeuven (B) met
het proefschrift: "Effort to reduce listening effort in hard-of-hearing children: a novel assessment paradigm"
De Engelstalige samenvatting van het proefschrift van Benso Hsu geven wij hier even voor jullie weer:
This PhD thesis describes the design, development, validation, experimentation, and initial application of a tool, a novel behavioural paradigm assessing the level of effortful listening in school-age children.
The initial inspiration of this novel assessment came from Picou and Ricketts (2014b), who first demonstrated that the sensitivity of a dual-task paradigm toward changes in listening effort could be increased if the secondary task was performed at a deeper level of cognitive processing (Craik & Lockhart 1972). In other words, listening effort, when defined as a type of cognitive load, can be most readily measured when the auditory-linguistic task itself involves more intense and/or complex interactions between working memory, central executive functions, and long-term semantic memory stores. In clinical terms, our proposed tool may help identify hard-of-hearing children who may be suffering from sustained, high levels of listening effort in noisy mainstream classrooms. Naturally, a timely identification of this condition allows the timely provision of intervention strategies aiming to reduce listening effort.
This is the ultimate goal of the current PhD. To achieve this goal, three experimental studies were conducted to investigate the parameters affecting this paradigm of measuring listening effort.
Study I began with the adaptation of the methods from Picou and Ricketts (2014b) to a school-age population and validated the listening effort paradigm in 31 normal-hearing (NH), native Dutch-speaking children aged from 7 to 12 years old. The paradigm consisted of a classic word recognition task presented in quiet or two levels of background noise coupled to one of three categorization tasks designed to tap into varying levels of cognitive processing when a participant listened to a target word. For instance, the categorization condition “Is this something dangerous?” (simply “dangerous” from now on) was expected to invoke a greater depth of processing than the categorization condition “Is this an animal?” (simply “animal” from now on). Depending on how well a participant could repeat lists of target words, A percentage-correct word recognition score (WRS) was calculated for each condition. Moreover, the response time (RT) a participant took to answer a given categorization question was automatically recorded by the computer and served as the paradigm’s primary indicator of listening effort. The general hypotheses for the PhD project could thus be formulated as 1) increasing level of competing noise would significantly affect the performance on WRS as well as lengthen categorization RT, and 2) increasing the depth of semantic processing would significantly increase RT while having negligible effects on WRS.
These same hypotheses were carried over to Study II, where a similar behavioural listening effort paradigm was tested in a larger group (n = 90) of NH participants in order to investigate the effect of age on WRS and RT. The NH participants recruited for Study II ranged from school-age children as young as 6 years old to young adults in their early to mid 20’s. The speculation was that better WRS performance and faster categorization RT would be linked to the maturation of a participant’s central cognitive system as represented by his/her chronological age. Results from analyses of outcome data obtained from Studies I and II provided support for the aforementioned hypotheses regarding the effects of noise, processing depth, and age. WRS performance indeed increased with a participant’s age but decreased with increasing background noise levels. Listening effort, as measured by RT, tended to decrease with a participant’s age until approximately the end of the teenage period (12 to 18 years old). Listening effort also tended to increase with increasing levels of competing noise as well as increasing depth of semantic processing. Additional cognition/language test results from Study II suggested that listening effort, as measured by RT, was significantly correlated with long-term memory retrieval speed but not with working memory capacity in our group of NH participants.
The final stage of the PhD trajectory, Study III, aimed to be a sort of pilot “clinical trial” of our behavioural listening effort paradigm in a small group of school-age children who were active users of cochlear implant (CI) devices. Specifically, a shortened version of the assessment paradigm from Study II was administered to 14 CI recipients between 5 and 14 years of age during their regular follow-up appointment at a university audiology clinic. All but two of our CI participants successfully completed the listening effort assessment without any issue. To allow meaningful comparison of outcome measures between this CI group and their age-matched NH counterparts from Study II, a generalized linear mixed-effect model (GLMM) was utilized to construct predictive norms based on the RT data set from Study II. Data collected under Study III suggested that older (aged 9½ years or older) school-age CI recipients tended to achieve better WRS and shorter RT than younger (less than 9½ years old) school-age CI recipients when target words were presented in quiet. Curiously, this same age advantage among our CI participants was not as obvious when target words were presented in noise.
In summary, the three PhD studies were built around a novel dual-task assessment paradigm and the three crucial factors affecting word recognition and listening effort outcomes: the chronological age of a participant, the level of competing background noise, and the depth of semantic processing demanded by the paradigm’s secondary task. The current assessment’s clinical applicability could be enhanced by the further reduction of testing time without sacrificing test sensitivity and reliability. The ultimate validity of the assessment paradigm as a measure of listening effort could also be increased by comparing outcomes with other frequently cited tools such as pupillometry and structured questionnaires. The eventual deployment of a viable listening effort assessment in daily clinical practice would be another important step toward helping children with hearing impairment live and learn better in this noisy world of ours.