Table 2 Performance metrics for three of the 3D-DenseNet models are provided after tenfold cross-validation.

From: Twinned neuroimaging analysis contributes to improving the classification of young people with autism spectrum disorder

Data type

Mean ± SE (SD)

Accuracy %

Sensitivity %

Specificity %

Precision %

F1 score %

sMRI

65.1 ± 3.36 (5.4)

71.7 ± 4 (6.4)

58.3 ± 6.64 (10.7)

64.3 ± 3.75 (6)

67.55 ± 2.8 (4.55)

ALFF

72 ± 3.1 (4.9)

66.5 ± 5.3 (8.6)

77.5 ± 6.5 (10.5)

75.2 ± 4.9 (7.9)

69.9 ± 3.15 (5.1)

ALFF + sMRI

76.9 ± 2.34 (3.8)

70.8 ± 5.1 (8.2)

83.14 ± 6.3 (11.6)

82.4 ± 5 (8.15)

75.6 ± 2.4 (3.9)

  1. Accuracies, sensitivities, specificities, precisions, and F1 scores are shown as mean, standard error (SE), and standard deviation (SD).
Back to article page