Postural instability is considered one of the most incapacitating motor symptoms and a primary cause of falls in Parkinson's disease (PD), compromising patients' autonomy and well-being. The traditional clinical examination used to evaluate this symptom designed by pull test is difficult to standardize and is not sensitive to subtle but significant postural changes. Inertial measurement units have emerged as a portable and cost-effective solution to measure on-body patients' postural sway allowing them to obtain more sensitive metrics able to capture postural instability. However, further studies are required to monitor patients' postural conditions under dynamic conditions.

The proposed research focused on investigating the hypothesis of whether it is possible to differentiate between all the scores of the pull test through postural and gait metrics extracted from raw acceleration and angular velocity signals from the centre of mass of patients with PD acquired while performing basic daily tasks. A new cross-sectional study was conducted with 23 patients to determine which gait and postural-associated metrics are considered significant to distinguish between the different levels of pull test, and which metrics are more correlated with the pull test score.

Achieved results showed that most of the estimated metrics can differentiate the pull test scores ( , R ). The duration of the activity, root-mean-square and range of motion of vertical and mediolateral angular velocity, as also most of the gait-associated metrics, presented the most significant differences in all trials which involved motion tasks, such as sitting, lying, walking and turning.

Overall, promising results were achieved as the statistical analysis revealed that gait and postural metrics estimated under dynamic conditions were considered relevant to distinguish between the scores of the pull test.