Table 5 Performance of TL-BLSTM-attention models based on EfficientNetB0 and VGG16 for different brain regions on unseen test data of TD-BRAIN dataset. The mean and standard deviation of each performance measure is presented.

From: A convolutional recurrent neural network with attention for response prediction to repetitive transcranial magnetic stimulation in major depressive disorder

  

Frontal

Central

Temporal

Parietal

Occipital

VGG16-BLSTM-Attention

ACC (%)

81.0 ± 4.2

79.2 ± 5.6

81.7 ± 4.1

79.9 ± 4.6

78.4 ± 3.8

SEN (%)

77.8 ± 3.5

77.4 ± 4.1

74.1 ± 3.0

79.5 ± 3.7

76.6 ± 4.4

SPE (%)

81.9 ± 3.4

80.5 ± 3.7

82.5 ± 3.6

80.4 ± 3.1

78.5 ± 3.5

AUC

0.82 ± 0.05

0.80 ± 0.04

0.83 ± 0.05

0.79 ± 0.02

0.79 ± 0.03

EfficientNetB0-BLSTM-Attention

ACC (%)

80.8 ± 3.7

80.7 ± 2.6

82.3 ± 3.3

79.7 ± 4.8

80.2 ± 3.2

SEN (%)

79.5 ± 1.9

78.1 ± 3.2

80.2 ± 3.4

78.3 ± 3.8

79.2 ± 2.9

SPE (%)

81.0 ± 3.0

81.2 ± 3.7

81.9 ± 2.8

79.0 ± 1.6

79.9 ± 3.1

AUC

0.82 ± 0.01

0.81 ± 0.03

0.83 ± 0.03

0.79 ± 0.05

0.79 ± 0.03

  1. Significant values are in bold.