English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Structural magnetic resonance imaging (sMRI) provides accurate estimates ofthe brain's structural organization and learning invariant brainrepresentations from sMRI is an enduring issue in neuroscience. Previous deeprepresentation learning models ignore the fact that the brain, as the core ofhuman cognitive activity, is distinct from other organs whose primary attributeis anatomy. Therefore, capturing the semantic structure that dominatesinterindividual cognitive variability is key to accurately representing thebrain. Given that this high-level semantic information is subtle, distributed,and interdependently latent in the brain structure, sMRI-based models need tocapture fine-grained details and understand how they relate to the overallglobal structure. However, existing models are optimized by simple objectives,making features collapse into homogeneity and worsening simultaneousrepresentation of fine-grained information and holistic semantics, causing alack of biological plausibility and interpretation of cognition. Here, wepropose MCIAT, a unified framework that combines Multi-task Collaborativepre-training and Individual-Adaptive-Tokens fine-tuning. Specifically, we firstsynthesize restorative learning, age prediction auxiliary learning andadversarial learning as a joint proxy task for deep semantic representationlearning. Then, a mutual-attention-based token selection method is proposed tohighlight discriminative features. The proposed MCIAT achieves state-of-the-artdiagnosis performance on the ADHD-200 dataset compared with several sMRI-basedapproaches and shows superior generalization on the MCIC and OASIS datasets.Moreover, we studied 12 behavioral tasks and found significant associationsbetween cognitive functions and MCIAT-established representations, whichverifies the interpretability of our proposed framework.

Multi-task Collaborative Pre-training and Individual-adaptive-tokens Fine-tuning: A Unified Framework for Brain Representation Learning