Multimodal physiological signals provide a more reliable data source for depression detection. For instance, combining electroencephalography (EEG) and pupil area (PA) signals can enhance depression recognition. However, EEG acquisition is challenging, limiting the practical use of EEG-based multimodal approaches, while PA signals are more accessible. Additionally, while existing explainability methods for time series models can quantify the contribution of each feature, they often fail to provide a comprehensive understanding of how these contributions drive performance improvements, limiting insights into the underlying mechanisms. To address these limitations and enhance the generalizability of PA-based depression detection models, this paper proposes a cross-modal knowledge distillation method, using an EEG and PA-based multimodal teacher model and a PA-based unimodal student model. Through knowledge distillation, complex multimodal features are transferred to the PA-based model, enhancing its performance. We also introduce Entropy-GradCAM (E-GCAM), an explainability method combining information entropy and gradient-weighted class activation mapping (Grad-CAM), to clarify mechanisms behind the student model’s performance gains. Quantitative results show that knowledge-distilled time series models encode more useful information, consistent with observed student model improvements. Experimental results demonstrate that the proposed method achieves optimal performance on two datasets, effectively reducing reliance on multimodal data and increasing the practicality of depression recognition models.

Accurate identification of critical states and their primary driving factors in complex biological processes is crucial for providing early warning signals of catastrophic shifts. Although existing methods based on dynamic network biomarkers (DNBs) have made some progress, they often fall short in fully utilizing the information from single-sample networks and struggle with the computational challenges posed by ultrahigh-dimensional data. Here, we introduce a comprehensive and rapid single-sample DNB method (crsDNB). It introduces a global community detection process and integrates it with a local network perspective to achieve self-adaptive identification of single-sample DNBs. In addition, it proposes targeted parallel optimization strategies to enhance computational efficiency. Validated using six transcriptomic datasets related to male aging and cancer, the crsDNB successfully identified critical states prior to decisive transitions, achieving significant improvements in both computational and biological significance metrics. Consequently, the crsDNB can identify personalized biomarkers for each sample more accurately and efficiently, providing a powerful new tool for determining critical states in complex biological processes.