Axis 2. Artificial intelligence for cognitive and language profiling?

Healthy adults possess distinctive, heterogeneous patterns of relative strengths and weaknesses across various cognitive domains such as attention, processing speed, working memory, and long-term memory (cognitive profile). The gold standard for obtaining an individual’s cognitive profile is administering an in-depth neuropsychological evaluation. However, a complete neuropsychological evaluation is expensive and time-consuming. Identifying accurate biomarkers of cognitive profiles of healthy individuals based on short and easy-to-implement data collection represents an interesting and poorly explored challenge in fundamental neuroscience. Here, we hypothesized that the analysis of discourse could represent a valid biomarker of individuals’ cognitive and structural brain heterogeneity.

Axis #2 tests whether discourse features predict cognitive and structural brain profiles in healthy adults. The specific aims of Axis 2 were:

• to use recent artificial intelligence developments to optimize the pipelines of automatic discourse analyses in Canadian English and Quebec French
• to use the pipelines of automatic discourse analysis to extract discourse features from image descriptions in healthy adults from available databases and relate them to the individuals’ cognitive profile, and pattern of cortical thickness and surface area.