Sleep Brain and Cardiac Activity Predict Cognitive Flexibility and Conceptual Reasoning Using Deep Learning

2025-05-30

AI Summary

This study investigates the specificity of associations between sleep microstructure and executive functions—specifically cognitive flexibility and conceptual reasoning. We propose CogPSGFormer, the first end-to-end multi-scale convolutional-Transformer hybrid architecture designed to model whole-night multimodal physiological signals—including EEG, ECG/HRV, and interpretable spectral power features—for task-free, individualized assessment of executive function. By leveraging cross-modal complementary representation learning and feature disentanglement, we systematically validate the predictive specificity of sleep microstructural markers for higher-order cognition. Evaluated on the STAGES dataset (N=817), CogPSGFormer achieves 80.3% accuracy in stratifying participants into high- versus low-performance groups on the Penn Conditional Exclusion Test (PCET), significantly outperforming baseline models. The implementation is publicly released to facilitate reproducibility and further research.

Technology Category

Cognitive Modeling & Cognitive Systems: (Computational) Cognitive ArchitecturesIntelligent Robots: Multimodal Perception & Sensor FusionMachine Learning: Neuro-Symbolic Learning

Application Category

Security and Privacy: Large-scale security measurementsSearch and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAGUser Modeling, Personalization and Recommendation: Psychology-informed user models and recommender systems

Abstract

Despite extensive research on the relationship between sleep and cognition, the connection between sleep microstructure and human performance across specific cognitive domains remains underexplored. This study investigates whether deep learning models can predict executive functions, particularly cognitive adaptability and conceptual reasoning from physiological processes during a night's sleep. To address this, we introduce CogPSGFormer, a multi-scale convolutional-transformer model designed to process multi-modal polysomnographic data. This model integrates one-channel ECG and EEG signals along with extracted features, including EEG power bands and heart rate variability parameters, to capture complementary information across modalities. A thorough evaluation of the CogPSGFormer architecture was conducted to optimize the processing of extended sleep signals and identify the most effective configuration. The proposed framework was evaluated on 817 individuals from the STAGES dataset using cross-validation. The model achieved 80.3% accuracy in classifying individuals into low vs. high cognitive performance groups on unseen data based on Penn Conditional Exclusion Test (PCET) scores. These findings highlight the effectiveness of our multi-scale feature extraction and multi-modal learning approach in leveraging sleep-derived signals for cognitive performance prediction. To facilitate reproducibility, our code is publicly accessible (https://github.com/boshrakh95/CogPSGFormer.git).

Problem

• Predict cognitive flexibility from sleep physiology using deep learning
• Explore sleep microstructure links to executive function performance
• Develop multi-modal model for sleep-based cognition classification

Innovation

• Multi-scale convolutional-transformer model for sleep data
• Integrates ECG and EEG with power bands
• Achieves 80.3% accuracy in cognitive prediction