What can gait analysis tell us about dementia and its subtypes? : an integrated study of brain and behaviour

Abstract

Discrete gait characteristics are associated with select cognitive functions, potentially reflecting underlying neural processes. Therefore, different dementia subtypes may have unique signatures of gait impairment, reflecting their different underlying disease pathologies. As such, gait may be a useful tool to aid differential diagnosis of dementia disease subtypes, such as Alzheimer’s disease (AD) and Lewy body disease (LBD), which includes dementia with Lewy bodies (DLB) and Parkinson’s disease dementia (PDD). A large structured review undertaken as part of this thesis highlighted a lack of studies investigating a comprehensive range of gait characteristics in well-classified dementia subtypes. Thus, the primary aim of this thesis was to investigate the potential of discrete gait characteristics to differentiate dementia disease subtypes in both laboratory and free-living environments. There was a particular emphasis on discriminating AD and DLB, as clinical similarities between these subtypes can lead to misdiagnosis and incorrect management and treatment of disease. 110 people participated in this observational cross sectional study. Participants with mild cognitive impairment and dementia related to AD (n = 36), DLB (n = 30) and PDD (n = 15), and controls (n = 29) underwent gait assessment in controlled laboratory environments. Additionally, body-worn monitors continuously collected gait data over seven days in free living environments, providing information about spatiotemporal gait characteristics, and the quantity, variability and pattern of habitual walking activity. Participants completed a battery of cognitive tests and, associations between gait and cognitive variables were examined across all testing environments. Selective patterns of gait impairment differentiated AD and LBD subtypes in the laboratory, while PDD could be discriminated from all disease subtypes in free-living environments. When considering patterns of gait impairment across different walking bout lengths, there was promising evidence that gait could also differentiate AD and LBD in free-living environments. Gait-cognition associations appeared dependent on disease subtype, potentially reflecting underlying pathology. Additionally, differences in habitual walking behaviour between controls and dementia subtypes was found and associated with motor disease severity, balance confidence and executive dysfunction. This thesis is the first to describe gait in well-characterised dementia disease subtypes in both laboratory and free-living conditions. It provides novel evidence to support a role for quantitative gait analysis and discrete characteristics as clinical biomarkers to aid differential diagnosis and further enhance understanding of the complex relationship between gait and cognition