Table 1 Summary of the previous studies performing classification of neurological disorders using MRI and with clear data leakage (see also Supplementary Table S1 for a detailed description).
From: Effect of data leakage in brain MRI classification using 2D convolutional neural networks
Disorder | References | Groups (number of subjects) | Machine learning model | Data split method | Type of data leakage | Accuracy (%) |
|---|---|---|---|---|---|---|
AD/MCI | Gunawardena et al.36 | AD-MCI-HC (36) | 2D CNN | 4:1 train/test slice-level split | Wrong split | 96.00 |
Hon and Khan21 | AD-HC (200) | 2D CNN (VGG16) | 4:1 train/test slice-level split | Wrong split | 96.25 | |
Jain et al.37 | AD-MCI-HC (150) | 2D CNN (VGG16) | 4:1 train/test slice-level split | Late and wrong split | 95.00 | |
Khagi et al.38 | AD-HC (56) | 2D CNN (AlexNet, GoogLeNet,ResNet50, new CNN) | 6:2:2 train/validation/test slice-level split | Wrong split | 98.00 | |
Sarraf et al.22 | AD-HC (43) | 2D CNN (LeNet-5) | 3:1:1 train/validation/test slice-level split | Wrong split | 96.85 | |
Wang et al.39 | MCI-HC (629) | 2D CNN | Data augmentation + 10:3:3 train/validation/test split by MRI slices | Wrong split and augmentation before split | 90.60 | |
Puranik et al.40 | AD/EMCI-HC (75) | 2D CNN | 17:3 train/test split by MRI slices | Wrong split | 98.40 | |
Basheera et al.41 | AD-MCI-HC (1820) | 2D CNN | 4:1 train/test split by MRI slices | Wrong split | 90.47 | |
Nawaz et al.42 | AD-MCI-HC (1726) | 2D CNN | 6:2:2 slice level split | Wrong split | 99.89 |
