Beyond Single-Source Cognitive Taskonomy:Multi-Source Task Relations through fMRI Transfer Learning

2026-06-24 · Source: Computer Vision and Pattern Recognition · Field: Science & Research — Artificial Intelligence & Machine Learning, Life Sciences & Biology, Mathematics & Computational Sciences · Depth: Expert, quick

Summary

An fMRI cognitive taskonomy has been extended from single-source to multi-source transfer across 23 Human Connectome Project task states, utilizing Boolean Integer Programming (BIP) for budget-constrained task allocation. Researchers trained 1,127 task-specific and transfer models, revealing that single-source transfer is directional and paradigm-structured; motor states transfer effectively within their paradigm but offer limited support to non-motor targets. Multi-source transfer's effectiveness hinges on the source set's composition, indicating that pairwise taskonomy alone cannot fully capture many-to-one task relations. Under supervision budgets, BIP consistently prioritizes direct supervision for several 0-back and 2-back working-memory states, suggesting their high importance due to integrated perceptual, attentional, and executive processes. These findings highlight a cross-paradigm-limited motor cluster and high-priority working-memory states.

Key takeaway

For AI Scientists developing fMRI analysis pipelines, understanding multi-source task relations is crucial. Your current single-source transfer models may miss complex dependencies, especially for working-memory tasks that integrate perceptual and executive processes. Consider implementing Boolean Integer Programming (BIP) for budget-constrained task allocation to identify high-priority states, optimizing your experimental design and data interpretation. This approach can reveal more nuanced cognitive taskonomies.

Key insights

Multi-source fMRI transfer learning reveals complex cognitive task relations beyond single-source analysis.

Principles

• Single-source fMRI transfer is paradigm-structured.
• Multi-source transfer depends on source set composition.
• Budget-constrained allocation prioritizes specific states.

Method

The study extends reconstruction-based fMRI taskonomy using Boolean Integer Programming (BIP) to analyze budget-constrained task allocation across 23 Human Connectome Project task states.

In practice

• Identify high-priority working-memory states.
• Optimize fMRI task allocation with BIP.
• Analyze multi-source task relations.

Topics

• fMRI Transfer Learning
• Cognitive Taskonomy
• Boolean Integer Programming
• Human Connectome Project
• Working Memory
• Neural Processes

Best for: AI Scientist, Research Scientist

Related on AIssential

• A homotopy-type-theoretic generalization of neurosymbolic inference — HoTT generalizes neurosymbolic inference, leveraging symmetries for improved shortcut-aware concept posteriors and calibration.
• Beyond representational alignment with brain-guided language models for robust reasoning — Large language models can be directly enhanced by task-evoked brain signals to improve reasoning performance.
• Isolating Nonlinear Independent Sources in fMRI with $\beta$-TCVAE Models — $\beta$-TCVAE effectively disentangles nonlinear fMRI data into biologically meaningful and spatially coherent brain networks.
• Memory Transfer Learning: How Memories are Transferred Across Domains in Coding Agents — Cross-domain memory transfer, especially with abstract insights, significantly boosts coding agent performance by sharing meta-knowledge.
• Pattern Selectivity is Not Task-Causal Structure: A Cross-Architecture Mechanistic Study of Composed-Task Circuits in 1B-Class Language Models — Specific mechanistic circuits for composed tasks do not transfer across 1B-class language models.
• Coarse-to-fine Hierarchical Architecture with Sequential Mamba for Brain Reconstruction — CHASMBrain's dual-stream Mamba and coarse-to-fine fMRI encoding architecture achieves high predictive accuracy and reveals visual cortex functional specialization.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computer Vision and Pattern Recognition.