Network-Aware Bilinear Tokenization for Brain Functional Connectivity Representation Learning

Leo Milecki, Qingyu Hu, Bahram Jafrasteh, Mert R. Sabuncu, Qingyu Zhao

arXiv:2605.14048 · 2026-05-16 공개 · arXiv · PDF

self-supervised-learning representation-learning ablation-study masked-autoencoder developmental-cohorts parcellation psychopathology-prediction brain-functional-connectivity

Abstract

Masked autoencoders (MAEs) have recently shown promise for self-supervised representation learning of resting-state brain functional connectivity (FC). However, a fundamental question remains unresolved: how should FC matrices be tokenized to align with the intrinsic modular organization of large-scale brain networks? Existing approaches typically adopt region-centric or graph-based schemes that treat FC as structurally homogeneous elements and overlook the large-scale network brain organization. We introduce NERVE (Network-Aware Representations of Brain Functional Connectivity via Bilinear Tokenization), a self-supervised learning framework that redefines FC tokenization by partitioning FC matrices into patches of intra- and inter-network connectivity blocks. Unlike image-based MAE, where fixed-size patches share a common tokenizer, FC patches defined by network pairs are heterogeneous in size and correspond to distinct functional roles. To resolve this problem, NERVE embeds FC patches through a novel structured bilinear factorization. This formulation preserves network identity and reduces parameter complexity from quadratic to linear scaling in the number of networks. We evaluate NERVE across three large-scale developmental cohorts (ABCD, PNC, and CCNP) for behavior and psychopathology prediction. Compared to structurally agnostic MAE variants and graph-based self-supervised baselines, the proposed network-aware formulation yields more stable and transferable representations, particularly in cross-cohort evaluation. Ablation studies confirm that the proposed bilinear network embedding and anatomically grounded parcellation are critical for performance. These findings highlight the importance of incorporating domain-specific structural priors into self-supervised learning for functional connectomics.

한국어 요약

📋 한 줄 요약

**[Neuroscience / Self-Supervised Learning]** 뇌 기능적 연결(FC) 행렬을 뇌 네트워크 구조에 맞춰 intra-/inter-network 블록 패치로 토큰화하고 bilinear factorization으로 임베딩하는 자기지도 학습 프레임워크 NERVE를 제안한다.

🎯 핵심 기여도

💡 핵심 아이디어

뇌의 FC는 픽셀처럼 균일한 데이터가 아니라, 모듈러 네트워크 구조에 따른 의미 단위의 블록들의 집합이다. 따라서 토큰화·임베딩 단계에서부터 네트워크 정체성과 기능적 역할을 보존해야 더 안정적이고 전이 가능한 표현을 얻을 수 있다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 자기지도 학습에 도메인 구조 prior를 어떻게 반영해야 하는지를 functional connectomics에서 모범적으로 보여줌. **한계**: 사용된 코호트가 발달기 데이터에 집중, 임상 적용을 위해서는 추가적 외부 검증과 안전성 평가 필요.

🚀 실용적 활용