Open Access Repository

Associations between brain structure, cognition and gait in older people

Jayakody Arachchige Dona, SO ORCID: 0000-0001-5220-4927 2021 , 'Associations between brain structure, cognition and gait in older people', PhD thesis, University of Tasmania.

Full text not available from this repository.

Abstract

Background:
Dementia and falls are significant causes of disability among older people. Identifying people at risk of dementia and falls is an international priority to prevent the rising burden of these syndromes. Gait is a robust marker of functional independence. Poorer gait performance precedes dementia symptoms and is associated with greater risks of falls. Therefore, gait could be a useful functional marker to identify people at risk of dementia and falls from an early stage.
Gait speed is a strong indicator of overall health of a person and has received ample attention in the literature. However, gait variability (step to step fluctuations of gait measures) has emerged as a potentially more sensitive marker of adverse health outcomes in older people, including dementia and risk of falls. Currently, there is a gap in understanding of the underlying factors of gait variability. Firstly, gait variability is greater in older people compared to young people. However, evidence is only now emerging regarding if gait variability changes over time with advancing age. Secondly, a range of cognitive, sensorimotor and medical factors are cross-sectionally associated with greater gait variability, but it is less clear if these factors modify change in gait variability over time. Third, although poorer brain structure (greater burden of white matter lesion, amyloid deposition) is associated with greater variability, there is lack of understanding of brain networks or specific brain regions associated with gait variability.
Poorer gait performance during single- and dual-task walking is associated with dementia. In single-task walking (walking without performing an additional task), slow speed is associated with decline in specific cognitive domains. However, associations between other gait characteristics (i.e. gait variability, walking speed reserve [WSR, the difference in gait speed between fast and usual pace walking]) and cognitive decline are not extensively examined. In terms of dual-task walking, there is a gap in knowledge of the dual-task test and measure that is most strongly associated with cognition.
Aims:
This thesis had two broad aims. They were to examine 1) the underlying factors of gait variability and 2) the relationships between different gait characteristics with poorer cognition and future cognitive decline. The specific aims of each study were to examine,
Study 1.
1. the longitudinal associations between age and gait variability measures
2. the demographic, medical, sensorimotor and cognitive factors associated with change in gait variability and mean gait variability
Study 2.
1. if specific grey matter covariance patterns were associated with individual gait variability measures
2. if discrete covariance patterns of grey matter and gait variability measures were associated with cognitive functions
Study 3.
1. the associations between regional cortical thickness (regional mean cortical thickness and regional thickness ratio) and individual gait variability measures
Study 4.
1. whether baseline gait characteristics (variability, speed, WSR) were associated with decline in specific cognitive domains
2. whether the presence of ApoE4 modified any association between baseline gait and cognitive decline
Study 5.
1. the dual-task test and measure most strongly associated with global cognition and individual cognitive domains
2. dual-task interference patterns in relation to cognitive performance
Methods:
All studies (except study 5) used data from the Tasmanian Study of Cognition and Gait (TASCOG). TASCOG included community dwelling older people (age range 60-85), randomly selected from the Southern Tasmanian electoral roll. Exclusion criteria included diagnosis of dementia and Parkinson’s disease, inability to walk without an aide and contraindication to magnetic resonance imaging (MRI). Gait assessment was conducted with a 4.6-meter GAITRite walkway. Gait speed was obtained from the GAITRite software. Variability was calculated for double support time (DST), step length, step time and step width as the standard deviation of a measure averaged across all steps of six walks. Cognitive performance in executive function, processing speed, visuospatial function, and memory was assessed with a neuropsychological test battery. Sensorimotor functions were assessed with the short version of Physiological Profile Assessment. Demographics and medical history were recorded using self-reported questionnaires. MRI acquisition was performed with a 1.5-Tesla scanner.
Study 1.
This was a prospective longitudinal study. Variability in gait measures (the dependent variables) were assessed over mean 4.6 years at three time points (n=410 at baseline, n=285 at first phase, n=250 at second phase). The following baseline factors were the independent variables: 1) demographics (age, sex, education) 2) medical history (a cumulative index for cardiovascular disease, lower limb arthritis, mood, body mass index) 3) cognitive function in specific domains and 4) sensorimotor functions (postural sway, quadriceps strength, vision, proprioception, grip strength). Longitudinal mixed effect models were used to examine 1) change in gait variability over time 2) factors that modify change in gait variability or predict mean gait variability, adjusting for confounders.
Study 2.
This was a cross-sectional study (n= 351). Gait variability at baseline were the dependent variables. T1 images of participants were processed through a voxel-based morphometry pipeline to obtain grey matter probability maps. A whole-brain multivariate covariance-based analysis was performed to identify grey matter volume covariance patterns associated with each gait variability measure. The individual expressions of grey matter patterns were correlated with cognitive domains. Models were adjusted for age, sex, education, height and total intracranial volume.
Study 3.
This was a cross-sectional study (n= 351). Individual gait variability measures were dependent variables. T1 images were processed through an automated FreeSufer (5.3) pipeline to obtain cortical thickness in 68 brain regions. Two thickness measures were adopted per region: 1) regional mean thickness and 2) regional thickness ratio (regional mean thickness/ mean thickness of the entire cortex). Bayesian regression models were used to determine associations between regional cortical thickness and gait variability measures. Models were adjusted for age, sex and height.
Study 4.
This was a prospective longitudinal study. Cognitive function in specific domains (measured at three time points over 4.6 years) were the dependent variables. Gait variability, gait speed and walking speed reserve (the difference between fast and usual speed) at baseline were the independent variables. The presence of ApoE4 allele was determined by whole blood DNA. Longitudinal mixed effect models were used to examine 1) associations between baseline gait characteristics and cognitive decline 2) if any associations were modified by ApoE4. Models were adjusted for age, sex and education.
Study 5.
This was a cross-sectional study that included 91 participants from the Tasmanian Healthy Brain Project. Under single- and dual-task, gait speed was obtained using a computerized Zeno walkway. Single-task cognitive performance was assessed when participants performed the following tasks in standing, for 30 seconds: 1) reciting alternate letters of the alphabet 2) counting backwards in 3s and 3) recalling words from a shopping list. For dual-task walking participants performed each cognitive task separately but while walking. Dual-task interference in gait and cognition were calculated as: (dual task–single task)/single task×100 and summed to obtain total interference. A neuropsychological test battery was used to assess cognitive performance in executive function, processing speed, working memory, verbal fluency, visuospatial function and verbal memory (recall and recognition). Raw test scores were subjected to principal component analysis to derive a global cognition score. Partial correlations were used to determine the strength of associations between single- and dual-task measures and cognitive scores, adjusting for age, sex and education.
Results:
The findings of the underlying factors of gait variability were as follows. Variability in step length, DST and step width increased over time. The rate of increase was modified by the presence of cardiovascular risk factors, lower education and weaker quadriceps, respectively. Specific medical, sensorimotor and cognitive factors predicted greater mean gait variability in all individual measures. Specific grey matter volume covariance patterns were associated with gait variability measures (except step time variability). The covariance patterns associated with DST and step length variability included multiple cortico-subcortical regions and also correlated with most cognitive domains. The pattern associated with step width variability included the cerebellum and a few visual-related regions and only correlated with memory. Smaller cortical thickness in multiple brain regions were associated with greater variability in step width and step time. Smaller thickness in a few specific frontal, temporal and occipital regions were associated with greater DST and step length variability. For each gait variability measure the associated regions were unique and important for different motor, sensory and cognitive functions. Smaller thickness ratio in a greater number of regions (compared to the regions identified by regional mean thickness) were associated with DST and step length variability.
The findings of the associations between different gait characteristics and cognition were as follows. Greater DST variability at baseline was associated with greater decline in memory. Slow gait speed at baseline was associated with greater decline in processing speed, visuospatial function and memory (only in ApoE4 carriers). Greater total interference during dual-task reciting alternate letters of the alphabet (DT-alpha) had the strongest associations with poorer global cognition, working memory and verbal memory recognition. People who adopted a mutual interference pattern (combined gait and cognitive interference) during DT-alpha had poorer cognition.
Conclusion:
Gait variability is not a unitary concept. Instead, different medical, sensorimotor and cognitive factors were associated with change in gait variability in individual measures. As greater variability is associated with adverse health outcomes (i.e. falls), preventing greater variability in older people is important. These findings suggest that in order to prevent adverse outcomes of greater variability, inventions should target different factors. Similarly, grey matter volume covariance patterns comprised of specific regions and smaller cortical thickness in specific brain regions were associated with each gait variability measure. This may potentially explain why individual gait variability measures had different underlying factors, and highlight that interventions to maintain brain health could be a potential pathway to prevent greater gait variability. Greater DST variability and slow gait speed may be useful motor biomarkers to identify people at risk of cognitive decline in specific domains. This will assist directing people to interventions to prevent accumulation of dementia pathology. Greater total interference and mutual interference pattern during DT-alpha could be useful dual-task measures to distinguish those with poorer cognition. However, clinically important cut scores should be identified to facilitate the use of these measures in the clinic.

Item Type: Thesis - PhD
Authors/Creators:Jayakody Arachchige Dona, SO
Keywords: Gait variability, cognitive decline, brain structure, falls, dementia, dual-task walking, older people
Copyright Information:

Copyright 2021 the author

Item Statistics: View statistics for this item

Actions (login required)

Item Control Page Item Control Page
TOP