Martin, Suzanne (2021) Gait Adaptability and Biofeedback in Older Adults with Diabetes. PhD thesis, Victoria University.

Abstract

In response to changes in the environment, the central nervous system frequently updates motor commands using sensory feedback related to the requirements of movements and adapts the internal model of gait. Therefore, gait adaptability is ability to modulate gait parameters in order to adjust trajectories of feet in the anatomical planes to avoid obstacles or step on targets. Although older adults with diabetes mellitus (diabetes) report falls more frequently than healthy older adult, a few studies have been conducted to investigate the effects of diabetes on gait adaptability in which participants with diabetes could plan their performance a few steps ahead. However, falls occur when participants have not enough time to plan and respond. This thesis had three aims: (i) to investigate the effects of diabetes on gait adaptability, (ii) to investigate the effects of biofeedback (visualised real-time performance) on foot displacement adjustments in the sagittal plane, and (iii) to investigate the agreement between foot displacement adjustments quantified by the treadmill and overground adaptability tests for future application of biofeedback tools for assessing foot displacement adjustment. To address all aims, 16 young adults (Group I), 16 healthy older adults (Group II) and 16 older adults with diabetes (Group III) were recruited. Exclusion criteria were musculoskeletal injury, uncorrected vision, a fall within a year before participation, cognition issues, and diabetic- or ageing-related neuropathy. To address the first aim, participants walked in baseline and then completed overground gait adaptability tests (40 trials) with four random conditions: step shortening, step lengthening, obstacle avoiding, and walking through. Step length targets were 40% of the baseline step length longer or shorter than the mean baseline step length. The obstacle presented a 5 cm height across the walkway. A Vicon three-dimensional motion capture system was used to quantify spatiotemporal parameters of a gait cycle. To address the second aim, participants walked on a motorised treadmill for a few minutes. Preferred speed for each participant was determined. Customised MATLAB programs used marker trajectory data collected by a three-dimensional motion capture system streamed by a Visual3D Server. They then computed mean step length and minimum toe clearance of each participant from three 60-second trials collected during a 10-minute walking at preferred speed. Four subject- specific targets included two step length targets (baseline step length ± 10% baseline step length) and two minimum toe clearance targets (2.5 cm and 3.5 cm higher than the mean minimum toe clearance). Targets values for each participant were entered in the biofeedback system before the participant completed adaptability tests. Participant could see continuously a graphical display of step length or vertical trajectory of the toe mark in real time on a monitor installed in front of the treadmill. Targets were randomly presented as horizontal lines discretely appeared every 10 steps and disappeared after 10 steps three times. Participants adapted their step lengths and minimum toe clearance heights with presented targets on the monitor using biofeedback (real-time distances between their step length/ minimum toe clearance heights and presented targets) without being aware of the order of targets. Overground gait adaptability parameters (step velocity, stance time, swing time, double support time, and step length) and errors of foot displacement adjustments (differences between real and desired step lengths/ minimum toe clearance heights) in the overground and treadmill gait adaptability tests were quantified. For statistical analyses of data related to the first two aims, analysis of variance (ANOVA) was used to test the main effects of group and condition at a significance level of 0.05. To address the third aim, Bland and Altman plots, scatter plots of difference and mean errors quantified by two treadmill and overground gait adaptability tests and the limits of agreement, were used in response to each condition to investigate the agreement between the tests. Sixteen young adults (Group I), 14 healthy older adults (Group II) and 13 older adults with diabetes (Group III) completed both overground and treadmill tests. Groups were not significantly different in gait spatiotemporal parameters (step length, stance time, swing time, double support time, step velocity) when they walked normally at their preferred speed. However, they were different in gait spatiotemporal parameters when they tried to meet goal-tasks in adaptability tests. In Group III, stance and double support times significantly increased when they adapted the trajectory of their feet to step length targets and the obstacle height. Increased stance and double support times did not increase the accuracy of foot displacement adjustments, as older adults with diabetes in Group III showed the greatest errors of step length and minimum toe clearance adjustments. The participants in Group III could use visual feedback to significantly reduce errors of their step length and minimum toe clearance adjustments during an online correction. However, they had the greatest errors in their responses with and without biofeedback compared with other groups. Errors of step length and minimum toe clearance adjustments without biofeedback in the treadmill adaptability tests were in agreement with those in the overground gait adaptability tests. This may suggest efficacy of the application of the treadmill adaptability tools to assess precise foot displacement adaptation in a feedforward model for fall prevention in older adults with diabetes.

Search Google Scholar

Repository staff login