Estimation of gait cycle characteristics by trunk accelerometry
Introduction
A large number of reports have been presented on the significance of gait parameters to diagnose impairments in balance control, assess functional ability, and predict risk of falling. Such parameters include cadence, step length and walking speed, which can be measured without the need of fixed laboratory equipment. By use of instrumented walkways more detailed records of temporal and spatial parameters may be obtained.
Variables reported typically change with walking speed, however (Winter, 1991), and differences in walking speed may therefore confound the results. This poses an important restriction on the interpretation of such data. Still, a common procedure is to have subjects walk at a self-selected preferred speed without controlling for differences in speed between sessions or subjects (Winter et al., 1990; Lord et al., 1996; Wolfson et al., 1985; Hallett et al., 1993). Some investigators therefore administer paced walking, where either walking speed is controlled as on a treadmill (Crane and Demer, 2000) or cadence is standardized by a metronome (Krebs et al., 2002). However, such constraints may affect walking behavior and thus restrict the validity of the results.
Most commonly, the average outcome over a given walking distance is reported, but variability between steps or strides may give additional information. Measures like within-subject step-length variability and within-subject step-width variability, however, require costly equipment like an instrumented walkway (Grabiner et al., 2001) or cumbersome analysis of ink marks (Sekiya et al., 1997; Helbostad and Moe-Nilssen, 2003), which have restricted their use in field research.
In recent years, low inertia piezoresistant accelerometers have become widespread at low cost, and accelerometry has been described for various biomechanical purposes (Aminian et al., 1999b; Nigg, 1994; Yack and Berger, 1993). Still, however, accelerometry is not in common use for gait analysis, possibly because unwanted variability caused by the gravity factor was not adequately dealt with in the past. Procedures have now been described to eliminate the gravity component (Moe-Nilssen, 1998a), and to assess trunk accelerations in walking (Moe-Nilssen, 1998b) and standing (Browne and O’Hare, 2001) as measures of balance control. The operational simplicity and high capacity of present day accelerometric technology suggest a method suitable for use in settings not restricted to a laboratory.
Several authors have attempted to derive common gait cycle parameters from accelerometry data. Thus Evans et al. (1991) used a uniaxial accelerometer to identify each heel strike. Auvinet et al. (1999) applied a biaxial accelerometry unit to derive measures of cycle frequency, stride symmetry, and stride regularity at preferred walking speed, while Aminian et al. (1999a) measured temporal parameters using two accelerometers, also at preferred speed. None of these investigators, however, reported how their chosen outcome measures might be affected by changes in preferred walking speed between sessions and subjects. This source of variability may camouflage relations among other variables of interest.
It is a research issue to develop procedures utilizing the advantages of ambulatory technology, at the same time eliminating unwanted variability associated with self-administered walking speeds. In this paper, we suggest a protocol for estimation of well-known parameters like cadence, step length, and measures of gait regularity and symmetry from data obtained by a single triaxial accelerometer, also describing procedures to control for the confounding effect of differing gait speed, when subjects are walking at self-administered speeds.
Section snippets
Instrumentation and data acquisition
Linear acceleration was measured along three orthogonal axes using a low-inertia (15 g) triaxial piezoresistant accelerometer snugly secured to the test subjects by a fixation belt over the L3 region of the spine. The accelerometer was connected to a battery-operated portable PCMCIA data logger also worn by the subject. Analog signals were low-pass filtered at 55 Hz before being sampled at 128 Hz. The digitized signals were stored on interchangeable 20 Mb PCMCIA cards, and subsequently transferred
Discussion
We have demonstrated how gait cycle periodicity of trunk acceleration data can be analyzed by unbiased autocorrelation procedures to give cadence, step length and measures of gait regularity and symmetry. Further we have suggested procedures to control for variability introduced by differences in walking speed, still allowing subjects to walk at self-administered speeds. Reliability and validity of the derived parameters are issues for ongoing research.
The idea of analyzing gait data by
References (24)
- et al.
Review of the different methods for assessing standing balance
Physiotherapy
(2001) - et al.
Effects of vestibular and cerebellar deficits on gaze and torso stability during ambulation
Otolaryngology Head and Neck Surgery
(2000) - et al.
Age-related changes in spatial and temporal gait variables
Archives of Physical Medicine and Rehabilitation
(2001) A new method for evaluating motor control in gait under real-life environmental conditions. Part 1the instrument
Clinical Biomechanics
(1998)A new method for evaluating motor control in gait under real-life environmental conditions. Part 2gait analysis
Clinical Biomechanics
(1998)Test–retest reliability of trunk accelerometry during standing and walking
Archives of Physical Medicine and Rehabilitation
(1998)- et al.
Dynamic adjustments of walking behavior dependent on noxious input in experimental low back pain
Pain
(1999) - et al.
Gait and balance in the elderly. Two functional capacities that link sensory and motor ability to falls
Clinical Geriatric Medicine
(1985) - et al.
Temporal feature estimation during walking using miniature accelerometersan analysis of gait improvement after hip arthroplasty
Medical & Biological Engineering & Computing
(1999) - et al.
Physical activity monitoring based on accelerometryvalidation and comparison with video observation
Medical & Biological Engineering & Computing
(1999)
Accelerometric gait analysis for use in hospital outpatients
Revue du Rhumatisme
Utilisation of an accelerometric device in equine gait analysis
Equine Veterinary Journal Supplement
Cited by (500)
Laboratory-assessed gait cycle entropy for classifying walking limitations among community-dwelling older adults
2024, Experimental GerontologyIndividual differences in children's movement variability in a virtual reality playground task
2024, Human Movement ScienceAge and walking conditions differently affect domains of gait
2023, Human Movement Science