Exploring Subnetwork Interactions in Heterogeneous Brain Network via Prior-Informed Graph Learning

Researchers have introduced KD-Brain, a novel graph learning framework that incorporates prior medical knowledge to model complex brain subnetwork interactions, achieving state-of-the-art performance in mental disorder diagnosis.

Computer Science > Machine Learning

Title: Exploring Subnetwork Interactions in Heterogeneous Brain Network via Prior-Informed Graph Learning

Modeling the complex interactions among functional subnetworks is crucial for the diagnosis of mental disorders and the identification of functional pathways. However, learning the interactions of the underlying subnetworks remains a significant challenge for existing Transformer-based methods due to the limited number of training samples. To address these challenges, we propose KD-Brain, a Prior-Informed Graph Learning framework for explicitly encoding prior knowledge to guide the learning process. Specifically, we design a Semantic-Conditioned Interaction mechanism that injects semantic priors into the attention query, explicitly navigating the subnetwork interactions based on their functional identities. Furthermore, we introduce a Pathology-Consistent Constraint, which regularizes the model optimization by aligning the learned interaction distributions with clinical priors. Additionally, KD-Brain leads to state-of-the-art performance on a wide range of disorder diagnosis tasks and identifies interpretable biomarkers consistent with psychiatric pathophysiology. Our code is available at this https URL.

Source: arXiv cs.AI Recent

More in this category

agentic-systems

Visual Graph Scaffolds for Structural Reasoning in Large Language Models

A new study reveals that visual graphs can serve as internal scaffolds to organize reasoning in Large Language Models (LLMs), rather than just acting as external knowledge sources. This visual guidance significantly outperforms flattened text representations in complex multi-hop reasoning tasks.

agentic-systems

Handoff Debt: The Rediscovery Cost When Coding Agents Take Over Interrupted Tasks

This paper introduces 'handoff debt', the cost of rediscovery when a successor coding agent resumes an interrupted task. It demonstrates that providing rich context during handoffs significantly improves efficiency, reducing prompt tokens by up to 63%.

NOW LET US Related – ChatHealthAI: Aligning Electronic Health Record Representations with Large Language Models for Grounded Clinical Reasoning

agentic-systems

ChatHealthAI: Aligning Electronic Health Record Representations with Large Language Models for Grounded Clinical Reasoning

Researchers have introduced ChatHealthAI, a multimodal reasoning framework that aligns structured Electronic Health Record (EHR) representations with Large Language Models (LLMs). This framework enhances clinical prediction accuracy while providing interpretable, language-based reasoning to support medical decision-making.

EXPLORE TOPICS