Skip to main content

Table 4 Overview of the number of trials, total duration (minutes), and FOG outcome of the training data for all models evaluated in this study

From: Freezing of gait assessment with inertial measurement units and deep learning: effect of tasks, medication states, and stops

Model name

#Trials

Total duration (min)

%TF

#FOG

Model_TUG

89

35.99

19.88

103

Model_360Turn

44

45.11

21.71

179

Model_Off

67

45.75

27.37

205

Model_On

66

35.36

12.53

77

Model_Clinical

133

81.11

20.90

282

Model_Stop

213

100.60

9.57

248

  1. A total of 133 trials were measured with standardized FOG-provoking tests excluding stops, which were used to train Model_Clinical, and 213 trials were measured with self-generated and researcher-imposed stops, which were used to train Model_Stop. Within the 133 trials, 89 trials of TUG tasks were used to train Model_TUG, and 44 were used to train Model_360Turn. Similarly, 67 trials Off medication were used to train Model_Off, and 66 trials were used to train Model_On. Trial durations are shown in minutes. FOG severity is quantified by means of the %TF and #FOG