Table 2 Database characteristics

From: Self-supervised identification and elimination of harmful datasets in distributed machine learning for medical image analysis

Study

Number of centers

Parkinson’s disease

Healthy

participants

Mean age (std)

Males

(%)

Scanner vendor

ADNI

42

0

274

67.76 (4.27)

35%

Siemens, GE, and Philips

BIOCOG

1

45

49

71.29 (4.65)

56%

Siemens

C-BIG

1

66

10

64.65 (8.90)

49%

Siemens

HAMBURG

1

74

39

63.16 (9.95)

67%

Siemens

CCNA

6

57

0

68.35 (7.46)

65%

GE

JAPAN

1

30

15

66.15 (6.58)

44%

Siemens

NEUROCON

1

26

16

68.33 (11.0)

48%

Siemens

OASIS

1

0

27

66.33 (7.17)

63%

Siemens

CALGARY

1

79

42

70.78 (6.58)

60%

GE

MONTREAL

1

41

21

62.16 (5.93)

58%

Siemens

PPMI

24

384

162

61.10 (10.38)

64%

Siemens, GE, and Philips

SALD

1

0

78

61.65 (8.03)

100%

Siemens

TAOWU

1

17

20

64.64 (4.92)

54%

Siemens

UKBB

1

48

197

66.95 (7.58)

60%

Siemens