Table 1 Comparison of mental health perception technologies.
Study | Data source | Method | Target group | Application scenario | Core limitation/distinction |
|---|---|---|---|---|---|
Zhao et al.7 | ECG/EEG/motion sensors | Multimodal biosignal fusion | Clinical depression patients | Medical diagnosis | Reliance on costly medical equipment, limited applicability in education |
Alkurdi et al.8 | Wearable physiological data | Real-world noise filtering model | Individuals with anxiety | Daily monitoring | Lacks integration of behavioral–psychological cross-modal analysis |
Begum9 | Social media text | Chaotic Simulated Binary PSO (CS-BPSO) | Young internet users | Online screening | Ignores behavior patterns specific to sports training |
Atta et al.10 | Clinical interview records | Swarm-intelligence-optimized neural network | High-risk depression groups | Automated diagnosis | Requires medical data, limited adaptability to campus scenarios |
Song11 | Plantar pressure sensors | Biomechanics systematic review | Long-distance runners | Sports injury prevention | No quantitative model linking injury to depression |
Halkiopoulos et al.13 | AR/VR interaction logs | Machine learning cognitive therapy optimization | Patients with psychological disorders | Virtual therapy | High hardware costs, limited integration with daily teaching |
Troussas et al. 14 | Educational software interaction logs | Reinforcement learning with dynamic fuzzy weights | General student population | Adaptive learning systems | No focus on psychological state perception |
This study | Campus IoT behavioral data | GA–RF hybrid model | Sports education students | Teaching intervention support | (1) Focus on training load–mental state relationship (2) Low-cost integration with campus devices (3) Closed-loop intervention in educational settings |
