Table 3 Performance comparison of our proposed multi-task, multi-stage deep transfer learning model using only connectome data versus only clinical data versus combined brain connectome and clinical data, for the joint prediction of abnormal cognitive, language, and motor outcomes at 2 years corrected age in very preterm infants.

70.9 ± 4.5

70.0 ± 6.9

71.7 ± 5.5

2.4 ± 0.4

28.3 ± 5.5

0.78 ± 0.06

69.1 ± 4.3

76.0 ± 4.6

62.1 ± 3.7

2.0 ± 0.2

37.9 ± 3.7

0.79 ± 0.04

81.5 ± 3.2

74.0 ± 4.9

88.9 ± 3.1

6.6 ± 1.9

11.1 ± 3.1

0.86 ± 0.05

62.8 ± 3.7

63.2 ± 6.8

62.4 ± 5.7

1.6 ± 0.2

37.6 ± 5.7

0.60 ± 0.05

47.3 ± 3.2

41.0 ± 3.5

53.6 ± 4.1

0.9 ± 0.1

46.4 ± 4.1

0.51 ± 0.03

68.9 ± 2.3

60.0 ± 4.0

77.8 ± 3.2

2.6 ± 0.3

22.2 ± 3.2

0.66 ± 0.03

70.1 ± 4.2

72.0 ± 7.3

68.1 ± 5.3

2.2 ± 0.3

31.9 ± 5.3

0.66 ± 0.06

71.7 ± 3.3

74.0 ± 3.6

69.3 ± 3.6

2.4 ± 0.3

30.7 ± 3.6

0.72 ± 0.02

73.9 ± 2.4

76.0 ± 4.9

71.7 ± 3.4

2.6 ± 0.2

28.3 ± 3.4

0.84 ± 0.02