Detection of age and gender differences in walking using mobile wearable sensors.

Abstract

Previous studies have demonstrated differences in gait speed and ground reaction forces (GRF) based on age as well as sex during lab-based testing. With advancements in wearable technology, it may be possible to assess differences in loading parameters in non-lab settings using portable data collection methods. The purpose of this study is to determine if wearable sensors (loadsol®) are valid for assessing peak force, impulse and loading rate (LR) in older adults and determine if the insole can detect sex and age differences in these parameters during walking. Twenty young (22.2 ± 2.9 years) and 23 older adults (68.1 ± 5.8 years) walked at a self-selected speed on a flat, inclined and declined instrumented treadmill (randomized order). Force data was simultaneously collected from the treadmill (1440 Hz) and insoles (100 Hz) during each condition. To assess validity, an ICC(3,k) and a Bland-Altman plot was generated for each variable and condition in the older adults. To determine age and gender differences, an ANCOVA (covary: walking speed) was completed for each variable. All ICCs were greater than 0.88 for vGRF, impulse and loading rate. The Bland-Altman plots report a bias of less than 2% for vGRF, -8 to -15 % for impulse and -5 to 2% for loading rate. The ANCOVA results indicate that the loadsol® has the ability to detect differences between age groups in peak vGRF in the flat, declined and inclined conditions which are in agreement with the differences the force plates detected. Similarly, the loadsol® and force plates agreed with age-based differences in the flat and inclined condition, but the loadsol® missed the declined LR difference. The ability to collect data in nontraditional settings has the potential to broaden the research questions investigated, explore clinical applications and increase the generalizability.