Project 3: Efficiency of Language Processing

点击翻译按钮获取中文摘要

PROJECT 3: PROJECT SUMMARY In work on hearing loss, variability in outcomes is a persistent puzzle. Listeners with comparable hearing do not always perform similarly in noise or the real world. Project 3 proposes this derives from two linked issues. First, clinical tests focus on accuracy. But accuracy is not sufficient. To cope with real-world demands, speech perception must be efficient and flexible. Second, not all variation is due to the auditory periphery. Cognitive ability, which declines with age, could exacerbate the effects of hearing loss. However, research has largely focused on general cognitive skills like working memory. Mechanisms specific to language have not received much attention. Project 3 links these issues, testing the hypothesis that the cognitive mechanisms of language processing are how listeners achieve efficiency and flexibility. We address this in word recognition, a hub in the language system linking sound and meaning. Cognitive models suggest listeners match the speech signal to thousands of candidates in the lexicon via a competition process, and that variation in efficiency and flexibility are emergent from this competition. This is tested with eye-movements techniques that yield a temporally rich picture of the dynamics of competition. Prior work suggests two profiles of competition for CI users. Successful CI users show a small delay and keep competitors available to facilitate revisions after a mistake. In contrast, listeners who struggle wait to access candidates until most of the word is heard and show reduced competition. Project 3 asks if these dimensions of efficiency are uniquely important for outcomes over and above the auditory periphery. Data from the last funding cycle support this idea, linking efficiency to aging, listening effort, the listening environment, and cortical structure. Project 3 builds on this with four aims. Aim 1 conducts a large multi-modal cross-sectional study of experienced CI users and a longitudinal study of new recipients. These studies ask what factors drive differences in efficiency, both in the moment and during adaptation to the CI. They investigate the role of real-world practice (using ecological momentary assessment, social network analysis and device logs), language skills (using neuropsychological tests and a new reading/eye-movement task), cortical integrity (using pre-implantation MRI), and listening effort (using pupillometry). They also ask whether differences in efficiency predict differences in outcomes including the ability to keep up with fast speech. Aim 2 asks how the dimensions of efficiency adapt to challenging listening, using a novel experimental paradigm in which sentence contexts create expectations for either noise or rapid speech. Aim 3 asks whether efficiency is related to the way CI users combine information across ears. It uses a dichotic listening paradigm to ascertain whether information is combined into a composite object prior to speech recognition or integrated later. This measure is related to auditory encoding in each ear, and to eye-tracking measure of efficiency. Together, this work will reveal a new standard for assessment outcomes (efficiency), and other modifiable factors (real-world practice, tuning to support binaural integration) to enhance outcomes.