Neural Networks to Understand the Neurobiological Mechanisms of General Intelligence

Human intelligence is a key concept in psychology, reflecting overall cognitive ability and predicting life outcomes like educational success, socioeconomic status, health, and longevity. Individuals with higher general intelligence scores tend to perform better across a variety of cognitive tasks, an observation known as the positive manifold of general intelligence, or g. While general intelligence likely has a biological basis in brain structure and function, the specific cognitive processes and neural mechanisms underlying individual differences remain unclear. This thesis explores, in three studies, how variations in brain function—assessed by functional magnetic resonance imaging (fMRI) and electroencephalography (EEG)—relate to individual differences in intelligence. The first study tested the relationship between adaptation in functional brain connectivity (FC) to different task demands and individual intelligence. Findings showed that less FC adaptation across tasks was associated with higher intelligence, supporting theories of neural efficiency. The second study examined resting-state brain dynamics using EEG, focusing on brain signal complexity at different temporal scales. Lower complexity at lower scales—likely reflecting short-range neural processes—correlated with higher intelligence. The third study applied predictive modeling to identify which functional brain connections best predict intelligence. Results showed that task type influences observable brain-behavior relationships, and that intelligence scores can be predicted from diverse brain-wide connectivity patterns, complementing existing theories. Together, the results suggest that individual differences in intelligence emerge from properties of a global, adaptive functional brain network that operates across multiple temporal and spatial scales. This highlights the need for integrative, multiscale neural recordings to better understand individual differences in intelligence at the neural level.