Auto detection and segmentation of daily living activities during a Timed Up and Go task in people with Parkinson’s disease using multiple inertial sensors

Abstract

BackgroundWearable sensors have the potential to provide clinicians with access to motor performance of people with movement disorder as they undergo intervention. However, sensor data often have to be manually classified and segmented before they can be processed into clinical metrics. This process can be time consuming. We recently proposed detection and segmentation algorithms based on peak detection using Inertial Measurement Units (IMUs) to automatically identify and isolate common activities during daily living such as standing up, walking, turning, and sitting down. These algorithms were developed using a homogenous population of healthy older adults. The aim of this study was to investigate the transferability of these algorithms in people with Parkinson’s disease (PD).MethodsA modified Timed Up And Go task was used since it is comprised of these activities, all performed in a continuous fashion. Twelve older adults diagnosed with early PD (Hoehn & Yahr ≤ 2) were recruited for the study and performed three trials of a 10 and 5-m TUG during OFF state. They were outfitted with 17 IMUs covering each body segment. Raw data from IMUs were detrended, normalized and filtered to reveal kinematics peaks that corresponded to different activities. Segmentation was accomplished by identifying the first minimum or maximum to the right and the left of these peaks. Segmentation times were compared to results from two examiners who visually segmented the activities. Specificity and sensitivity were used to evaluate the accuracy of the detection algorithms.ResultsUsing the same IMUs and algorithms developed in the previous study, we were able to detect these activities with 97.6% sensitivity and 92.7% specificity (n = 432) in PD population. However, with modifications to the IMUs selection, we were able to detect these activities with 100% accuracy. Similarly, applying the same segmentation to PD population, we were able to isolate these activities within ~500 ms of the visual segmentation. Re-optimizing the filtering frequencies, we were able to reduce this difference to ~400 ms.ConclusionsThis study demonstrates the agility and transferability of using a system of IMUs to accurately detect and segment activities in daily living in people with movement disorders.

The Quebec Parkinson Network: A Researcher-Patient Matching Platform and Multimodal Biorepository

“…Data collected from inertial measurement units (IMUs) worn by patients with PD were used to assess motor symptoms including tremor, bradykinesia, and freezing of gait, as well as drug-induced dyskinesia [29]. Using this inertial sensor technology, algorithms were developed that can detect and segment movements in patients performing a Timed Up and Go task [30] or carrying out activities of daily living in a simulated living space [31]. IMU data were also used to identify a turning signature that discriminates between older-age adults and older-age patients with PD, as well as between patients in ON and OFF medication states [32].…”

The Use of Inertial Measurement Units for the Study of Free Living Environment Activity Assessment: A Literature Review

“…Authors of these studies set up algorithms that allow, from previous annotated observations of participants performing activities (labeled training data), to determine which type of activity is performed during an analysis of a signal section (see Section 5 for more details on machine learning procedures for activity classification). Some of these studies point towards the study of machine learning systems [ 62 , 63 , 64 ], while others focus on other factors such as sensor implementation [ 65 ]. This is the group of aims that contains the greatest number of studies (28 papers).…”

“…Some clinical trials focus on healthy participants, some compare the patients included with healthy participants [ 39 , 42 , 71 , 72 ], while others dwell specifically upon cohorts with a medical condition. Perriot et al [ 7 ] intended to improve posture detection in COPD participants and Nguyen et al [ 63 ] focused on activity classification in patients with PD. In such work, the aim is either to compare the results obtained on certain patients with specific pathologies to control patients or to evaluate the impact that changes in instrumentation (position of sensors, etc.)…”

“…[ 27 , 28 , 54 , 61 , 65 , 74 , 86 ]). Moreover, only three studies use more than ten IMUs—these studies use the IGS-180 suit consisting of seventeen IMUs [ 18 , 63 , 70 ]. Besides, 41% of the studies use only one sensor for all of their acquisitions [ 39 , 64 , 72 , 79 ].…”

Application of the ActiGraph GT9X IMU for the assessment of turning during walking and running

“…Based on previous literature, it is unclear what low-pass filter should be used for the gyroscope and magnetometer data. Previous studies that examined turning have used low-pass filters between 0.7-1.5 Hz (18,51,52). Thus, for the current study, the data were filtered with a second-order low-pass Butterworth filter at various frequencies (0.25, 0.50, 0.75, 1.00, 1.25, 1.50, 1.75, 2.00 Hz).…”