Table 6.
Highest classification accuracies attained on the respective datasets that are relevant for SZ diagnosis.
| Work | Dataset | Modality | AI Technique | Highest Classification Accuracy (%) |
|---|---|---|---|---|
| Buettner et al. [72]—2019 | RepOD | EEG | RF, 10-fold cross-validation | 100 |
| Siuly et al. [70]—2023 | Kaggle Basic Sensory Task data | EEG | Deep ResNets, Softmax layer and deep features with SVM | 99.23 |
| Sobahi et al. [83]—2022 | MSU | EEG | ELM-based AE | 97.7 |
| Qureshi et al. [85]—2017 | COBRE | sMRI, fMRI | ELM | 99.29 |
| Nguyen et al. [96]—2018 | Clinical | sMRI | SVM, 10-fold cross-validation | 99.3 |
| Chatterjee et al. [116]—2018 | FBIRN | fMRI | Linear SVM | 99.5 |