This project, funded by the Spanish Ministry of Science and Innovation (PID2022-141787NB-I00), aims to enhance our understanding of cognitive diversity in the general population and its neural underpinnings. Assessing general cognitive function has long been considered valuable for academic, occupational, and health-related applications. Traditionally, a single numerical score (e.g., IQ) has been used to summarize complex cognitive performance profiles, based on the observation that performance across different tests tends to correlate. While this approach has empirical support, it oversimplifies the picture for many individuals, who often exhibit heterogeneous cognitive profiles. Such variability challenges the fairness and accuracy of representing cognitive ability through a single score and may partly explain inconsistencies in identifying the neural correlates (e.g., cortical structure, functional connectivity) of general cognitive function.
We propose that cognitive structure itself may vary between individuals, reflecting distinct neural patterns. To investigate this, we will employ a novel combination of statistical methods-person-fit indices from Item Response Theory, Gaussian mixture models, and t-SNE-which our simulations suggest are sensitive to how well individual cognitive profiles align with a general model. We hypothesize that:
-Person-fit indices will modulate the relationship between general cognitive function and brain structure in prefrontal and parietal cortices, with better model fit predicting stronger associations.
-Accounting for person-fit will improve cross-validation in independent samples.
-Irrespective of their cognitive performance, individuals with general cognitive profiles will exhibit greater cortical thickness in domain-general regions linked to executive functions and behavior control (e.g., dorsolateral prefrontal and anterior cingulate cortex), along with more homogeneous cortical structure and functional connectivity