Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Data-Efficient Model for Psychological Resilience Prediction based on Neurological Data

Zhi Zhang, Yan Liu, Mengxia Gao, Yu Yang, Jiannong Cao, Wai Kai Hou, Shirley Li, Sonata Yau, Yun Kwok Wing, Tatia M. C. Lee

TL;DR

This paper tackles the challenge of predicting psychological resilience from limited EEG and resting-state fMRI data. It introduces a data-efficient framework built on modality-specific Neuro Kolmogorov-Arnold Networks (KAN), a trait-informed chunking strategy to augment data, and a noise-informed inference mechanism to handle low-SNR signals. Through experiments on RESIL and LEMON, the approach outperforms state-of-the-art baselines, and ablation studies confirm the value of each component. The work also provides interpretability analyses linking neural patterns to behavioral factors, highlighting its potential for practical resilience assessment and future neuropsychological research.

Abstract

Psychological resilience, defined as the ability to rebound from adversity, is crucial for mental health. Compared with traditional resilience assessments through self-reported questionnaires, resilience assessments based on neurological data offer more objective results with biological markers, hence significantly enhancing credibility. This paper proposes a novel data-efficient model to address the scarcity of neurological data. We employ Neuro Kolmogorov-Arnold Networks as the structure of the prediction model. In the training stage, a new trait-informed multimodal representation algorithm with a smart chunk technique is proposed to learn the shared latent space with limited data. In the test stage, a new noise-informed inference algorithm is proposed to address the low signal-to-noise ratio of the neurological data. The proposed model not only shows impressive performance on both public datasets and self-constructed datasets but also provides some valuable psychological hypotheses for future research.

Data-Efficient Model for Psychological Resilience Prediction based on Neurological Data

TL;DR

This paper tackles the challenge of predicting psychological resilience from limited EEG and resting-state fMRI data. It introduces a data-efficient framework built on modality-specific Neuro Kolmogorov-Arnold Networks (KAN), a trait-informed chunking strategy to augment data, and a noise-informed inference mechanism to handle low-SNR signals. Through experiments on RESIL and LEMON, the approach outperforms state-of-the-art baselines, and ablation studies confirm the value of each component. The work also provides interpretability analyses linking neural patterns to behavioral factors, highlighting its potential for practical resilience assessment and future neuropsychological research.

Abstract

Psychological resilience, defined as the ability to rebound from adversity, is crucial for mental health. Compared with traditional resilience assessments through self-reported questionnaires, resilience assessments based on neurological data offer more objective results with biological markers, hence significantly enhancing credibility. This paper proposes a novel data-efficient model to address the scarcity of neurological data. We employ Neuro Kolmogorov-Arnold Networks as the structure of the prediction model. In the training stage, a new trait-informed multimodal representation algorithm with a smart chunk technique is proposed to learn the shared latent space with limited data. In the test stage, a new noise-informed inference algorithm is proposed to address the low signal-to-noise ratio of the neurological data. The proposed model not only shows impressive performance on both public datasets and self-constructed datasets but also provides some valuable psychological hypotheses for future research.

Paper Structure

This paper contains 15 sections, 13 equations, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Kolmogorov-Arnold Networks
  4. Multimodal Representation Learning
  5. Methodology
  6. Neuro Kolmogorov-Arnold Networks
  7. Triat-informed Multimodal Representation Learning
  8. Noise-informed Psychological Resilience Prediction
  9. Experiments
  10. Comparative Experiments
  11. Ablation Study
  12. Interpretability Experiments
  13. Exploration Experiments
  14. Case Study
  15. Conclusions and Future Work

Figures (5)

  • Figure 1: The overall framework of the data-efficient model for psychological resilience prediction.
  • Figure 2: Interpretability analysis of EEG topography and functional connectivity on the RESIL dataset.
  • Figure 3: 2D visualization of EEG and fMRI features for low resilience and high resilience groups.
  • Figure 4: Co-occurrence network of important regions in fMRI, electrodes in EEG, and scales in behavioral data based on the learned model of RESIL dataset.
  • Figure 5: Case study grouped by similarities on the RESIL dataset.