Task-guided Spatiotemporal Network with Diffusion Augmentation for EEG-based Dementia Diagnosis and MMSE Prediction

2026-04-26

AI Summary

This study addresses the challenge of task interference in traditional multi-task approaches for EEG-based dementia diagnosis and Mini-Mental State Examination (MMSE) score prediction, which arises from entangled features across tasks. To mitigate this issue, the authors propose a Task-Guided Spatio-Temporal Network (TGSN) that integrates multi-band EEG features, leverages a diffusion model for data augmentation, and incorporates a gated spatio-temporal attention mechanism along with a task-guided query module to disentangle task-specific representations. Evaluated on the XY02 dataset, TGSN achieves classification accuracies of 97.78% and 83.93% for binary (AD/FTD) and ternary (AD/FTD/VCI) diagnostic tasks, respectively—outperforming the best baseline by 16.39% and 8.28%. Furthermore, it reduces MMSE prediction root mean square error (RMSE) to 1.93 and 2.38. The model also demonstrates strong cross-dataset generalization on the DS004504 dataset.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Patients with dementia typically exhibit cognitive impairment, which is routinely assessed using the Mini-Mental State Examination (MMSE). Concurrently, their underlying neurophysiological abnormalities are reflected in Electroencephalography (EEG), providing a basis for joint modeling. However, traditional multi-task approaches suffer from feature entanglement, which leads to inter-task interference when handling heterogeneous objectives.To address this challenge, we propose a task-guided spatiotemporal network (TGSN) with diffusion augmentation for EEG-based dementia diagnosis and MMSE prediction. Specifically, TGSN integrates a multi-band feature fusion module to capture complementary spectral information from EEG. Meanwhile, a pre-trained data augmentation module utilizing a diffusion process is introduced toincrease sample diversity. To model the complex spatiotemporal patterns of EEG, we propose a gated spatiotemporal attention module that captures long-range spatial dependencies and temporal dynamics. Moreover, we design a task-guided query module to achieve task-specific feature extraction, thereby mitigating task interference. The effectiveness of TGSN is evaluated on the XY02 dataset. Experimental results demonstrate that the proposed network outperforms several state-of-the-art methods, achieving classification accuracies of 97.78\% for Alzheimer's Disease (AD)/Frontotemporal Dementia (FTD) and 83.93\% for AD/FTD/Vascular Cognitive Impairment (VCI), which exceed the best baselines by 16.39\% and 8.28\%, respectively. In parallel, it reduces the RMSE for MMSE prediction to 1.93 and 2.38, achieving significant error reductions of 1.44 and 1.43 compared to the best baselines. Additionally, validation on the DS004504 dataset demonstrates strong cross-dataset generalization...

Problem

• EEG-based dementia diagnosis
• MMSE prediction
• multi-task learning
• task interference
• feature entanglement

Innovation

• task-guided spatiotemporal network
• diffusion augmentation
• multi-band feature fusion
• gated spatiotemporal attention
• EEG-based dementia diagnosis