Table 2. Measures and methods used to classify sedentary behavior, physical inactivity, and non-motor symptoms.

Publication	Study measures (sedentary behavior, physical (in)activity, and non-motor symptoms)	Measurement properties	Statistical methods used	Risk of bias score
Ellingson et al., 2019 [15]	Sedentary behavior Objectively measured sedentary time using activPAL and Actigraph accelerometers Self-reported sedentary time measured using Sedentary Behavior Questionnaire (SBQ). Non-motor symptoms PDQ-39 (cognitive domain)	Actigraph GT3X+ counts per second of each of the 3 axes were utilized. The specific cut-point used is unknown. ActivPAL’s proprietary software used in activity classification with >90% accuracy in classifying sedentary from non-sedentary behaviors [52]. The SBQ has acceptable test-retest reliability (ICC 0.51–0.93) [53]. Significant but small correlation between the PDQ-39 Cognitions score and the neurocognitive composite scores for delayed episodic memory (r = −0.13, p = 0.01) and processing speed (r = −0.12, p = .02) [29].	Spearman’s correlations and linear regression analysis	5
van Uem et al., 2018 [28]	Sedentary behavior Objectively measured sedentary behaviour using DynaPort MiniMod accelerometer Non-motor symptoms Geriatric Depression Scale PDQ-39 (cognitive domain) Mini-Mental State Examination (MMSE)	A minimum of 2 days of activity monitoring is required to obtain an ICC ≥ 0.7 for most activities [54]. Significant correlation with Hamilton Depression Rating Scale (r = 0.82 P<0.01) [55, 56]. The PDQ-39 correlates significantly with MMSE (r = -0.32, P<0.01) [57]. The PDQ-39 also correlates with the neurocognitive composite scores for delayed episodic memory (r = −0.13, p = 0.01) and processing speed (r = −0.12, p = .02).	Stepwise multivariate regression analyses	5
Jones et al., 2020 [29]	Physical (in)activity Physical Activity Scale for the Elderly (PASE) Non-motor symptoms Hopkins Verbal Learning Test-Revised, Judgment of Line Orientation, Symbols Digits Modalities Test, and Animal Fluency Test	Test-retest reliability (intraclass correlation coefficient) of 0.69 for household-related physical activity [58]. Hopkins Verbal Learning Test- Revised has a reliability of r = 0.74, P<0.01 for total recall, and predicts cognitive decline in PD (hazard ratio (HR) 0.98, P<0.01) [59]. Judgement of Line Orientation Test has a test-retest reliability of 0.90 with standard error of measurement of 1.8 points [60]. Symbol Digit Modalities Test predicts cognitive decline in PD (HR 0.98, P = 0.04) [61]. Animal Fluency is sensitive (0.88) and specific (0.96) in early detection of dementia [62].	Ordinal multilevel modeling (MLM)	5
Timblin et al., 2022 [30]	Physical (in)activity Physical Activity Scale for the Elderly (PASE) [58] Non-motor symptoms Hopkins Verbal Learning Test- Revised, Judgement of Line Orientation Test, Letter-Number Sequencing task, Symbols Digits Modalities Test, and Animal Fluency. Geriatric Depression Scale [55, 56]	Reported above. In addition to measurement properties reported above, the Letter-Number Sequencing task has good test-retest reliability (ICC = 0.64) in PD. Reported above.	Structural equation modeling (SEM)	5
Troutman et al., 2020 [33]	Sedentary behavior Objective measurement of sedentary behavior using Sensewear pro armband Non-motor symptoms Cognition: Parkinson’s Disease-Cognitive Rating Scale (PD-CRS)	At least 3 weekdays of monitoring required to achieve a reliability of 0.80 [63]. The PD-CRS has a test-retest reliability (ICC) >0.70, and sensitive (94%) and specific (94%) to detect PD dementia [54, 64].	Linear regression	6
Sulzer et al., 2021 [31] Longitudinal cohort study	Sedentary behavior Objectively measured sedentary behavior using DynaPort MiniMod accelerometer [54] Non-motor symptoms Cognition: Parkinson Neuropsychometric Dementia Assessment (PANDA). Geriatric Depression Scale [55, 56]	Reported above. The PANDA (cognition) had a specificity of 91% and a sensitivity of 90% for PD dementia and 77% for PD dementia plus PD-mild cognitive disorder [65]. Reported above.	Binary logistic regression	6
Prusynski et al., 2022 [32] Prospective observational study	Sedentary behavior/Non-motor symptoms Objective measurement of sedentary behavior, physical activity and sleep using Fitbit Charge HR activity monitor.	Fitbit devices can correctly identify sleep epochs with accuracy of 0.81 to 0.91, sensitivity of 0.87 and 0.99, and specificity of 0.10 and 0.52 [66].	Linear regression	3