Objective: To study gait and balance through Inertial Measurement Units (IMUs) in patients with Parkinson´s disease (PD) at different stages and controls. Identify the most discriminative features, sensor locations and motor tasks.

Background: IMUs can provide quantitative objective information of the motor status in PD. Studies on gait and balance in PD patients with IMUs are limited in terms of sample size, number of sensors and tests performed. No agreement exists on the best sensor locations.

Method: 33 healthy control subjects and 40 PD patients, 12 H-Y I, 18 H-Y II and 10 H-Y III were recorded while they performed the Timed Up and Go Test (TUG), 15-meter walk, quiet standing on rigid surface and on foam. Quiet standing was measured in three conditions (eyes open, eyes closed and eyes open while performing a cognitive task). Subjects wore ten sensors located in the feet, shanks, thighs, wrists, lumbar region and chest. Spatiotemporal measures of gait (step length, step time, stride time, leg speed) and ranges of motion of ankles, knees, hips and trunk were calculated from signals coming from gait trials. Features corresponding to the TUG test included accelerometer-based features such as the Root Mean Square (RMS), jerk and range of the acceleration signal, TUG duration, duration of every phase of TUG and ranges of motion (RoM) of ankles, knees, hips and trunk.

Results: The TUG durations (specially the turn duration) the flexion/extension RoM of hips, knees and ankles during gait and during TUG transitions, the external/internal rotation RoM of the knees during TUG transitions and RMS values of feet and shanks in TUG exhibited high statistical differences between groups. Spatiotemporal measures of gait and accelerometer-based measures of quiet standing also showed significant differences across groups. Feet and chest were the most optimal places for a single sensor placement. For the use of two sensors, an ipsilateral foot – shank configuration would be the most appropriate. The TUG test obtained statistical differences in several features for every group. The Gait, Balance and Balance on foam only showed differences at advanced stages of the disease.

Conclusion: Inertial sensors can successfully capture motor differences between PD stages. A number of key features have been identified. Feet and chest were the most optimal places for single sensor placements. The TUG test showed the most discriminative capacity across PD stages.

« Back to MDS Virtual Congress 2020

MDS Abstracts - https://www.mdsabstracts.org/abstract/imu-based-study-on-gait-and-balance-in-parkinsons-disease-and-healthy-subjects/