Related Experiment Video
Updated: May 1, 2026
05:55
Modeling the Functional Network for Spatial Navigation in the Human Brain
Published on: October 13, 2023
1.7K
Association-sensory spatiotemporal hierarchy and functional gradient-regularised recurrent neural network with
Subati Abulikemu1,2, Puria Radmard3, Michail Mamalakis4,5,6
1Department of Psychiatry, University of Cambridge, Cambridge, UK. ss2905@cam.ac.uk.
NPJ Systems Biology and Applications
|April 29, 2026
Summary
Schizophrenia compresses the brain's sensory-to-association (AS) hierarchy, reducing functional differentiation. This neural compression may destabilize cognitive computations, impacting working memory.
Area of Science:
- Neuroscience
- Cognitive Science
- Psychiatry
Background:
- The human neocortex exhibits a functional sensory-to-association (AS) hierarchy.
- Understanding structural alterations in schizophrenia and their impact on neural dynamics is crucial.
Purpose of the Study:
- Investigate structural changes in the AS hierarchy in schizophrenia.
- Determine the implications of these changes for neural dynamics and cognitive computation.
Main Methods:
- Utilized a large fMRI dataset (N=355) to extract individual AS gradients.
- Quantified hierarchical organization using gradient range and estimated neural timescales.
- Employed gradient-regularized recurrent neural networks (RNNs) for working memory tasks.
Main Results:
- Schizophrenia demonstrated a compressed AS hierarchy, indicating reduced functional differentiation.
- Neural timescales were attenuated in schizophrenia, particularly in specialized regions.
- RNNs with greater gradient range exhibited more efficient learning and stable neural states.
Conclusions:
- AS gradient de-differentiation in schizophrenia may destabilize neural computations.
- Empirical timescale flattening and model-based evidence support this computational hypothesis.
