Improving Brain Disorder Diagnosis with Advanced Brain Function Representation and Kolmogorov-Arnold Networks

2025-04-04

AI Summary

Predefined brain atlases introduce selection bias and insufficient individual specificity in functional connectivity (FC) quantification. To address this, we propose AFBR-KAN—a novel end-to-end Transformer–Kolmogorov–Arnold Network (KAN) hybrid classification framework for fMRI-based辅助 diagnosis of autism spectrum disorder (ASD). Our key innovation is the first integration of KAN into the Transformer encoder—replacing conventional multi-layer perceptrons—to enable data-driven, dynamic FC modeling and interpretable brain functional representation learning. By eliminating atlas dependency, AFBR-KAN enhances individual specificity and cross-site generalizability. Evaluated on multi-center resting-state fMRI datasets, it achieves significantly improved ASD classification accuracy over state-of-the-art methods. The implementation is publicly available. This work establishes a new paradigm for neuroimaging-assisted diagnosis that jointly ensures interpretability, robustness, and clinical adaptability.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Humans and AI: Brain-Sensing and Analysis

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Quantifying functional connectivity (FC), a vital metric for the diagnosis of various brain disorders, traditionally relies on the use of a pre-defined brain atlas. However, using such atlases can lead to issues regarding selection bias and lack of regard for specificity. Addressing this, we propose a novel transformer-based classification network (AFBR-KAN) with effective brain function representation to aid in diagnosing autism spectrum disorder (ASD). AFBR-KAN leverages Kolmogorov-Arnold Network (KAN) blocks replacing traditional multi-layer perceptron (MLP) components. Thorough experimentation reveals the effectiveness of AFBR-KAN in improving the diagnosis of ASD under various configurations of the model architecture. Our code is available at https://github.com/tbwa233/ABFR-KAN

Problem

• Improving brain disorder diagnosis with advanced representation
• Addressing selection bias in functional connectivity quantification
• Enhancing autism diagnosis using transformer-based KAN networks

Innovation

• Transformer-based classification network for brain disorder diagnosis
• Kolmogorov-Arnold Network blocks replace traditional MLP components
• Effective brain function representation to reduce selection bias