Table 1 Characteristics of the participants

From: Deep learning-enabled accurate assessment of gait impairments in Parkinson’s disease using smartphone videos

Gait impairment score

All

0

1

2

P value

N

118

38

39

41

 

Sex, (F/M)

55/63

21/17

15/24

19/22

0.34

Age, years (mean ± SD)

62.9 ± 7.4

61.1 ± 5.0

62.8 ± 7.9

64.7 ± 8.4

0.09

Height, cm (mean ± SD)

166.2 ± 8.4

165.4 ± 7.2

168.0 ± 8.8

165.1 ± 9.0

0.26

Off/On Medication Test (N)

19

0

11

8

 
  1. We recruited 87 patients with PD and 31 healthy participants and classified them into three categories based on their gait impairment severity, assessed according to the MDS-UPDRS Part III-Gait scales. Here, we considered the UPDRS scores of healthy participants to be zero. In clinical practice, the severity of gait impairments of some patients with mild PD can be diagnosed as a UPDRS score of zero. To align with this, we incorporated six patients with a UPDRS score of 0 into the dataset. We used two-sided χ2 tests to analyze the statistical significance of inter-group differences for genders and one-way analysis of variance (ANOVA) for ages and heights. The p values show that the dataset has a balanced distribution of these categorical items across three groups (p > 0.05). Nineteen patients with PD in the independent test dataset further participated in an Off/On-medication test, in which their gait videos were recorded during off and on medication states, respectively.
Back to article page