Pull test estimation in Parkinson's disease patients using wearable sensor technology

摘要: Postural instability is one of the main motor impairment Parkinson's disease (PD). The Pull Test most common clinical examination to assess postural in PD. However, subjectivity and low discriminative power this test presents as a major drawback. In paper we propose novel methodology estimate scores from patients with We capture relationship between outcomes patients' foot motion patterns, using wearable sensors mounted on their shoes. 139 idiopathic performed four function tests, including walking repetitive motions, while acceleration orientation data was recorded. A total 684 features were extracted signals. Feature selection classification algorithms utilized score for each participant. Further, which would better predict score, depending patient's phenotype (i.e. bradykinetic, tremor-dominant or equivalent). When combining all phenotypes mean probability distribution achieved 0.75 (CI: [0.69 – 0.82]). Foot circling best predictive equivalent (mean = 0.79, CI: 0.87]) bradykinetic (mean: 0.75, [0.64 0.85]), 2×10 m. walk stop-and-go proved superior 0.85]). Overall, these results suggest that inertial can be used PD patients.