Gait analysis with videogrammetry can differentiate healthy elderly, mild cognitive impairment, and Alzheimer's disease: A cross-sectional study

Abstract

Gait parameters have been investigated as an additional tool for differential diagnosis in neurocognitive disorders, especially among healthy elderly (HE), those with mild cognitive impairment (MCI), and Alzheimer's disease (AD) patients. A videogrammetry system could be used as a low-cost and clinically practical equipment to capture and analyze gait in older adults. The aim of this study was to select the better gait parameter to differentiate these groups among different motor test conditions with videogrammetry analyses. Different motor conditions were used in three specific assessments: 10-meter walk test (10mWT), timed up and go test (TUGT), and treadmill walk test (TWT). These tasks were compared among HE (n = 17), MCI (n = 23), and AD (n = 23) groups. One-way ANOVA, Kruskal-Wallis, and Bonferroni post-hoc tests were used to compare variables among groups. Then, an effect size (ES) and a linear regression analysis were calculated. The gait parameters showed significant differences among groups in all conditions, but not in TWT. Controlled by confounding variables, the gait velocity in 10mWT at usual speed, and TUGT in dual-task condition, predicts 39% and 53% of the difference among diagnoses, respectively. Finally, these results suggest that a low-cost and practical video analysis could be able to differentiate HE, those with MCI, and AD patients in clinical assessments.

Introduction

The pathophysiological process of Alzheimer's disease (AD) can begin 10–20 years before the clinical diagnosis (Sperling et al., 2011). Mild cognitive impairment (MCI) is a transitional state of cognitive loss, higher than expected for age group and educational level, but does not fulfill the clinical criteria for probable AD yet (Petersen et al., 2001). Although gait parameters also change with age, recently impairments in motor tasks have been associated with early stages of dementia (Beauchet et al., 2018). According to Liu-Ambrose et al. (2008), this motor decline may occur up to 12 years before the diagnosis of dementia.

Motor and cognitive function areas of brain support the gait action through an interactive connection, exerting substantial influence on how an individual moves (Makeig et al., 2009). Mainly brain areas associated with executive and sensorimotor integration are active when voluntary movements are performed (J.A. Cohen et al., 2016). Although velocity decline coexists with cognitive impairments (Buracchio et al., 2010) parameters such as gait cycle time, step and stride lengths, are also able to differentiate among healthy elderly (HE), those with MCI, and AD patients (Gillain and Petermans, 2013).

According to Deschamps (2018), along with the clinical symptoms, human movement could be a possible biomarker of mental disorders. Bahureksa et al. (2017) show that motor tests could assess this pathological behavior in different tasks present in activities of daily living (ADLs) environment. Technological instruments that are used for gait analysis as well as three-dimensional motion capture cameras, inertial sensors, and electronic walkway have shown significant differences between HE and those with the onset of dementia (Nadkarni et al., 2009; Remelius et al., 2012). However, to use these instruments, a significant investment is required to import the technological equipment, to construct a specific test room (a standard environment with darkness and temperature control), and to train a team for calibrating the system, and editing complex data. Hence, a low-cost and clinically practical equipment is necessary to differentiate the biomechanics of healthy and ill elderly people.

In the present study, we propose that a single camera and free software can be used to capture and analyze the gait, respectively, in real-life locations. Hypothetically, the videogrammetry can satisfy these assumptions because it has been validated (Dunn et al., 2014). However, the mobility-task type can have a significant influence on the results of gait analysis. Current research in this area focuses on a dual-task (motor and cognitive) test to differentiate the mobility between HE, MCI, and AD (Auvinet et al., 2017). Though, in a recently published meta-analysis (de Oliveira Silva et al., 2019), we show that the single-task mobility could distinguish between the prodromal and early stages of AD. Kirtley (2006) has reported that the treadmill is a safe and comfortable type of equipment to evaluate the gait of elderly individuals with low-functional capacity, especially when there is no space to perform the 4-m test. In addition, investigations assume that the treadmill and overground gait kinematics have similar patterns for a self-selected speed (Riley et al., 2007; Watt et al., 2010).

Following the Canadian Gait Consortium Guidelines (Beauchet et al., 2017), we selected the usual overground walk tests for clinical and spatiotemporal analysis (timed up and go and 4 m at steady-state walking). Furthermore, on the basis of the increasing number of reports of surveys using the treadmill gait test (Simoni et al., 2013; Sloot et al., 2014), we evaluated the self-selected and imposed speed among the HE, MCI, and AD groups. The present study is the first one that compares gait parameters among different groups (HE, MCI, and AD) in individual tasks (usual and faster overground gait speed, single and dual-task, self-selected and imposed treadmill gait speed) with videogrammetry analyses. This cross-sectional study aims to identify the better gait parameter using videogrammetric analyses to distinguish between HE, MCI, and AD in different motor tasks.