Age-related Changes in Hearing, Speech Recognition, and Decision-Making

The ability to follow conversation becomes more difficult as we get older. Our laboratory uses brain imaging techniques to examine structural and functional reasons for speech recognition difficulty and the consequences of declines in the auditory periphery. We consistently observe a pattern of elevated activity in cingulo-opercular brain regions when listeners work to recognize speech in difficult listening conditions.  These brain regions appear to support speech recognition by detecting when tasks are challenging so that behavior can be adjusted to optimize performance.  This idea is supported by our findings that activity is often elevated in these regions before someone correctly recognizes a word. These types of findings have guided our work in behavioral and neuroeconomic directions to consider the value that people experience from communication to better understand when and how well older adults will  recognize speech and benefit from interventions.

Figure 1. Cingulo-opercular regions that are responsive to task difficulty and predict next trial performance. Here we are showing results from a study demonstrating that people don’t remember hearing words that they had correctly recognized in noise ~20 minutes earlier when cingulo-opercular activity was relatively low when they recognized the words.

Dyslexia Data Consortium: Methods for Multi-site Studies and Missingness

As part of a multi-site study to develop methods for retrospective studies, we have collected a large sample of data with the generous contribution of investigators who study reading development and disability. We have demonstrated how to deal with: 1) missing data when one research site did not collect a measure of interest; 2) unusual cases in large imaging datasets that could obscure results and provide mechanistic insight into reading disability; 3) the behavioral heterogeneity in reading disability samples; and 4) multi-site imaging data that can have varied contrast by using a case-control design that provides increased efficiency for observing voxel-based effects.


Figure 2. Case-control results demonstrating: A) regions where reading disability cases exhibited lower volume than controls; B) increased efficiency of the case-control design or matched-pair brain morphometry compared to a typical group difference approach for comparisons across the brain; and C) that there is increased efficiency for a case-control approach by controlling for within and between-site variance.