Abstract:
Long-term hearing loss in post-lingually deaf adults may lead to progressive structural changes in the cerebral cortices where auditory and language functions are process...Show MoreMetadata
Abstract:
Long-term hearing loss in post-lingually deaf adults may lead to progressive structural changes in the cerebral cortices where auditory and language functions are processed. These alterations may affect the outcome of the cochlear implant (CI) surgery. In our study, we aim to predict the surgery outcome using imaging features that characterize such cortical structural changes. We used voxel-based morphometry approach to calculate the brain GM density. To reconstruct a smaller feature-set while collapsing the number of voxel-wise GM density values, we applied an anatomically hypothesized ROI-based method and a data-driven cluster-based method. We fed the reconstructed features to a Random-Forest Regression model combined with clinical features. We observed that the cluster-based method outperformed the ROI-based method and proved the competence of the image features in CI outcome prediction. Our data-driven approach found that the most accurate prediction was made with the clusters of GM density changes in the middle temporal cortex, a critical network node of language processing, and in the thalamus, a structure (dis-)engaging and (de-)coupling cortical language operations.
Date of Conference: 08-11 April 2019
Date Added to IEEE Xplore: 11 July 2019
ISBN Information: