Watching the walk
27 July 2011
Off-the-shelf software has been used by a team at Oxford Brookes University to develop a user-friendly way for clinicians to select variables to perform gait analysis that is lower-cost and more efficient than current methods, but still provides important clinical measures to assist rehabilitation.
Clinicians in the NHS identified the need for a widely accessible and valid objective motion analysis tool. A 2003 survey of 1826 physiotherapists in the UK revealed that only 23.1% of people with gait impairments were measured in gait laboratories. The main reasons cited were that it’s too time consuming, too expensive, and there is a lack of technical knowledge. These results were confirmed in a 2009 survey conducted by Oxford Brookes University.
Optical motion capture systems (OMCS) are the gold standard for kinematic gait analysis. However, these systems are confined to the specialist’s gait laboratory. Such facilities are relatively expensive, time consuming, and have a measurement volume that’s restricted to a single room. For extremely accurate gait assessments, such as pre-operation measures, a specialist’s gait laboratory is the best solution. However, the community physiotherapist or the hospital clinician would greatly benefit from a faster, easier-to-use solution that is also small, lightweight, and preferably wireless. Inertial measurement units (IMUs) may provide the basis for just such a tool.
Our goal is to create a user-friendly product that uses National Instruments’ LabVIEW to analyse IMU signals and produces important clinical aspects of gait accurately and easily. Thus, we began developing a motion analysis tool for clinicians to use by taking signals obtained from IMUs and evaluating the accuracy of the tool.
System set-up
We began the project using Xsens MTx sensors, which produced very accurate results in healthy and clinical populations. We used LabVIEW and the output of these sensors for the initial development of the Oxford Brookes Gait System. The MTx sensors are wireless and are the size of a matchbox, so the participant can carry a battery pack around their waist while walking. The data is stored on a computer directly via Bluetooth.
We use LabVIEW to analyse the data after it is recorded. Live analysis or visual feedback is not required at this stage; however, it is possible. Algorithms and models within LabVIEW require user input to select data and input variables such as leg length, weight and height.
We generate measurement reports containing basic temporal-spatial measurements such as step time, cadence, step frequency, stride length and walking speed. We can also obtain more complex analysis such as walking efficiency, energy, and balance analysis with no extra effort. Analysing these parameters using OMCS takes a relatively long time. However, using LabVIEW, these parameters can be collected, analysed and put in clinical context within five minutes when operated by an unskilled user. Currently we are creating a Web-based database so that users around the UK can centrally store data.
Implementation
Gait assessment involves placing the IMU over the projected centre of mass, which is located over the fourth lumbar vertebrae on the lower spine. The participants in clinics are generally instructed to perform standardised walking tests such as walking more than 10 minutes at a self-selected walking speed. Other standardised clinical measures such as a two to six minute walking test can provide insight into the development of gait-related fatigue and therefore mobility in daily life. These tests currently provide information only on walking speed or distance. By using the Oxford Brookes Gait System, a clinician can perform a more detailed analysis during these assessments and gain valuable information.
Clinical Trials
Our first trial started in 2008 when we began looking into the accuracy of the algorithms used in an early version of the Oxford Brookes Gait System. We found that vertical position after double integration was accurate to .5 mm while walking. This enabled us to derive other aspects of gait based on vertical centre of mass movement. We used the system in a clinical trial to measure the gait of people with neurological conditions, a study funded by the UK’s Department of Health.
We also investigated the validity of the models and algorithms used in our LabVIEW program for different clinical populations. We found that in people with Parkinson’s disease, muscular dystrophy, motor neuron disease, and multiple sclerosis, as well as stroke victims, these models could be used after applying a “personal correction factor” derived from user-input values within LabVIEW.
The project is currently funded by a Trust Technical Translation grant. The first stage of this work is to determine the inter-rater reliability of our gait analysis system between the expert user and clinicians.
Currently, we are looking into novel methods of gait analysis. By combining tools within LabVIEW, we could create new insights into planning and execution of human movement. We have been working on this new analyses for the last year with very promising results.
Patrick Esser, Helen Dawes, Johnny Collett and Ken Howells work at the Oxford Brookes University
Contact Details and Archive...
Related Articles...
Most Viewed Articles...