ObjectiveWe developed and investigated the feasibility of a machine learning–based automated rating for the 2 cardinal symptoms of Parkinson disease (PD): resting tremor and bradykinesia.MethodsUsing OpenPose, a deep learning–based human pose estimation program, we analyzed video clips for resting tremor and finger tapping of the bilateral upper limbs of 55 patients with PD (110 arms). Key motion parameters, including resting tremor amplitude and finger tapping speed, amplitude, and fatigue, were extracted to develop a machine learning–based automatic Unified Parkinson's Disease Rating Scale (UPDRS) rating using support vector machine (SVM) method. To evaluate the performance of this model, we calculated weighted κ and intraclass correlation coefficients (ICCs) between the model and the gold standard rating by a movement disorder specialist who is trained and certified by the Movement Disorder Society for UPDRS rating. These values were compared to weighted κ and ICC between a nontrained human rater and the gold standard rating.ResultsFor resting tremors, the SVM model showed a very good to excellent reliability range with the gold standard rating (κ 0.791; ICC 0.927), with both values higher than that of nontrained human rater (κ 0.662; ICC 0.861). For finger tapping, the SVM model showed a very good reliability range with the gold standard rating (κ 0.700 and ICC 0.793), which was comparable to that for nontrained human raters (κ 0.627; ICC 0.797).ConclusionMachine learning–based algorithms that automatically rate PD cardinal symptoms are feasible, with more accurate results than nontrained human ratings.Classification of EvidenceThis study provides Class II evidence that machine learning–based automated rating of resting tremor and bradykinesia in people with PD has very good reliability compared to a rating by a movement disorder specialist.

ObjectiveTo evaluate the differences in urodynamic findings between multiple system atrophy (MSA) and Parkinson disease (PD) and to identify the differential diagnostic ability of urodynamic study.MethodsWe reviewed patients with MSA or PD who underwent urodynamic studies between January 2011 and August 2018. Patients with probable MSA and PD determined by movement disorder specialists at our center were included. Patients with alleged MSA or PD from outside hospitals, atypical or secondary parkinsonism, and any history of pelvic operation or radiation therapy were excluded.ResultsA total of 219 patients, 107 with MSA (male:female 50:57) and 112 with PD (male:female 57:55), were included. Patients with MSA had shorter disease duration and were referred for urologic evaluation earlier (p < 0.001). Detrusor overactivity and associated urine leakage were prominent in PD (p < 0.001). Patients with MSA showed lower maximal flow rate (4.0 ± 5.8 vs 9.1 ± 8.3 mL/s, p < 0.001) and larger postvoid residual (290.8 ± 196.7 vs 134.0 ± 188.1 mL, p < 0.001) with decreased compliance (44.9% vs 10.7%, p < 0.001) and impaired contractility (24.9 ± 33.8 vs 65.7 ± 51.1, p < 0.001). Postvoid residual from a pressure-flow study had the highest sensitivity and specificity (74.8% and 75.9%), followed by detrusor pressure at maximal uroflow (72.6% and 70.5%), bladder contractility index, and postvoid residual from uroflowmetry (71.0% and 70.5%, respectively).ConclusionsPatients with MSA showed lower maximal flow rate, larger postvoid residual with decreased compliance, and impaired contractility, whereas patients with PD had higher incidence of detrusor overactivity and associated leakage. For differential diagnosis, postvoid residual from a pressure-flow study provided the best sensitivity and specificity.Classification of evidenceThis study provides Class III evidence that urodynamic measures can distinguish patients with MSA from those with PD.