Neuroimaging evidence for a network sampling theory of individual differences in human intelligence test performance

Area of Science:

• Neuroscience
• Cognitive Psychology
• Network Science

Background:

• The structure of human intelligence (single vs. multiple abilities) and its brain basis remain debated.
• Understanding cognitive abilities' relationship to brain organization is crucial.

Purpose of the Study:

• To investigate the relationship between cognitive task performance, intelligence factor structure, and dynamic brain network connectomics.
• To determine if brain network states can predict cognitive task performance and individual differences in intelligence.

Main Methods:

• Combined psychometric testing with machine learning analysis of functional magnetic resonance imaging (fMRI) data.
• Examined dynamic-network connectomics during the performance of 12 cognitive sub-tasks.
• Correlated behavioral-psychometric task structures with similarities in evoked brain network states.

Main Results:

• Brain network states accurately classified 12 cognitive sub-tasks (74% accuracy).
• Similarities in task network states mirrored their behavioral-psychometric relationships.
• Network states were more precisely classified in higher-performing individuals.

Conclusions:

• The human brain employs a high-dimensional network-sampling mechanism for flexible cognitive coding.
• Individual differences in intelligence relate to the fidelity of expressing task-optimized brain network states.