Evaluating the ability of a predictive vision-based machine learning model to measure changes in gait in response to medication and DBS within individuals with Parkinson’s disease

Table 2 Macro-averaged precision, recall, f1-score, mean MDS-UPDRS-gait score prediction during on and off states and paired t-test significance value

Precision of rounded predictions (mean ± STD)	Recall of rounded predictions (mean ± STD)	F1-score of rounded predictions (mean ± STD)	Mean predicted MDS-UPDRS-gait score—OFF treatment	Mean predicted MDS-UPDRS-gait score—ON treatment	Paired t-test p-value for difference in model predicted scores in ON/OFF states	Mean clinician assessed MDS-UPDRS-gait score—OFF treatment	Mean clinician assessed MDS-UPDRS-gait score—ON treatment	Paired t-test p-value for difference in model clinician assessed scores in ON/OFF states
0.40 ± 0.12	0.30 ± 0.05	0.22 ± 0.07	1.22 ± 0.32	1.07 ± 0.24	0.017	1.71 ± 0.86	0.86 ± 0.67	< 0.001

The standard deviation represents the magnitude of the variation between the five repetitions of the experiments
MDS-UPDRS-gait Gait subscore (item 3.10) of the Movement Disorders Society revision of the Unified Parkinson’s Disease Rating Scale, STD Standard deviation

ISSN: 1475-925X