Estimation of manual wheelchair-based activities in the free-living environment using a neural network model with inertial body-worn sensors

Humeral elevation workspace during daily life of adults with spinal cord injury who use a manual wheelchair compared to age and sex matched able-bodied controls

“…During MWC propulsion the humeral elevation required is between 25 and 55° at a self-selected speed (52, 53); however during walking the humeral elevation angles required are much lower (54). The difference in humeral elevation during mobility likely is not the only contributor to this increase, as MWC users move about 3 km less than able-bodied individuals and only spend a small amount of their day actually propelling themselves; estimates range from 16 to 54 minutes per day (50, 55). Another contributing factor to this discrepancy may be wheelchair setup; for example MWC users may not place their arms in a neutral resting position (0-30°) due to the location of their arm rest.…”

“…Continuing to map this workspace for individuals with SCI who use a MWC while they perform specific tasks (i.e., propulsion or transfers) may help us to further understand daily risk exposures and the contribution of specific tasks. Although out of the scope of this study, activity detection is an active area of research for our group (55).…”

Application and Reliability of Accelerometer-Based Arm Use Intensities in the Free-Living Environment for Manual Wheelchair Users and Able-Bodied Individuals

“…Two separate data collections were employed to (1) define the arm use intensity levels and (2) estimate the percentage of daily time MWC users and able-bodied individuals spend in each arm use intensity level and test the reliabilities ( Figure 1 ). First, in-lab IMU and video data, previously collected from a sample of MWC users with SCI [ 20 ], were utilized to determine the range of SMA magnitudes that defined each of the arm use intensity levels (stationary, low, mid, and high). After classification accuracy was assessed for the stationary and active threshold, the in-lab levels were applied to IMU data collected in the free-living environment from a different sample of MWC users with SCI and sex and age-matched able-bodied individuals.…”

“…IMU and video data were collected while participants performed six MWC-based activities as part of an ancillary study [ 20 , 21 ]. Two IMUs were secured with straps to each of the participant’s upper arms (Emerald, APDM, Inc., Portland, OR, USA) and data were collected at 128 Hz ( Figure 2 ).…”

“…VLC media player (VideoLAN Organization, Paris, France) was used to view all in-lab video data for analyses. For the six in-lab MWC-based activities and the time naturally occurring between them, one rater (EF) with more than eight years of movement and video analysis experience coded each second of data as either stationary or active [ 20 ]. A threshold between stationary and active arm use was calculated from a receiver operating characteristic (ROC) curve; the SMA value which maximized specificity and sensitivity for a subset of data (6 of 8 participants, ~78% of total data) was chosen as the threshold (0.67 g).…”