Brain Networks and Intelligence: A Graph Neural Network Based Approach to Resting State fMRI Data

TL;DR

This work tackles predicting individual intelligence from resting-state fMRI by leveraging static functional connectivity through a novel ROI-aware graph neural network, BrainRGIN. BrainRGIN combines a clustering-based, ROI-aware Graph Isomorphism Network (GIN) with edge-aware aggregation (RGIN), TopK pooling for interpretability, and attention-based readouts (GARO/SERO) to produce graph-level predictions of fluid, crystallized, and total intelligence. Across ABCD and HCP datasets, BrainRGIN achieves lower MSE and higher correlations than relevant baselines, with interpretable neuroanatomical findings such as the Middle Frontal Gyrus and Middle Temporal Gyrus contributing to fluid intelligence, and the Caudate contributing to crystallized intelligence. The model’s ROI-focused design and end-to-end training demonstrate a robust framework for linking intrinsic brain networks to cognitive abilities in rsfMRI data, with potential extensions to task-based fMRI analyses.

Abstract

Resting-state functional magnetic resonance imaging (rsfMRI) is a powerful tool for investigating the relationship between brain function and cognitive processes as it allows for the functional organization of the brain to be captured without relying on a specific task or stimuli. In this paper, we present a novel modeling architecture called BrainRGIN for predicting intelligence (fluid, crystallized, and total intelligence) using graph neural networks on rsfMRI derived static functional network connectivity matrices. Extending from the existing graph convolution networks, our approach incorporates a clustering-based embedding and graph isomorphism network in the graph convolutional layer to reflect the nature of the brain sub-network organization and efficient network expression, in combination with TopK pooling and attention-based readout functions. We evaluated our proposed architecture on a large dataset, specifically the Adolescent Brain Cognitive Development Dataset, and demonstrated its effectiveness in predicting individual differences in intelligence. Our model achieved lower mean squared errors and higher correlation scores than existing relevant graph architectures and other traditional machine learning models for all of the intelligence prediction tasks. The middle frontal gyrus exhibited a significant contribution to both fluid and crystallized intelligence, suggesting their pivotal role in these cognitive processes. Total composite scores identified a diverse set of brain regions to be relevant which underscores the complex nature of total intelligence.

Figures (3)

• Figure 1: Overall Architecture of BrainRGIN . The static FNC matrix is extracted from a resting state fMRI time series data. Three blocks of BrainRGIN are used with attention-based readout functions followed by a fully connected layer for prediction.
• Figure 2: Regions Significant in Fluid and Crystallized Intelligence Prediction
• Figure 3: Significant regions expressed as Connectivity Networks for Total Composite Scores