ReviewGait and cognition: Mapping the global and discrete relationships in ageing and neurodegenerative disease
Introduction
Gait provides a marker of global health and is an important tool as a predictor for health status and survival in older adults (Hausdorff et al., 2001, Studenski et al., 2011). Gait is no longer regarded as purely a motor task. An extensive body of research has established that safe and effective gait requires input from higher cognitive areas (Hausdorff et al., 2005). Research over the past decade has refined our understanding of the relationship between gait and cognition to reveal compensatory cognitive strategies which vary as a function of age and pathology (Hausdorff et al., 2005). Cross sectional studies identify associations between gait and cognition in normal ageing and neurodegenerative disease (Yogev et al., 2005, Ijmker and Lamoth, 2012, Verlinden et al., 2013a, Lord et al., 2014) as supported by neuroimaging studies (Holtzer et al., 2014). Longitudinally, gait emerges as a strong and significant predictor of future cognitive impairment and dementia in older adults (Marquis et al., 2002, Verghese et al., 2007, Buracchio et al., 2010, Mielke et al., 2013).
Gait speed is universally used to reflect gait because of its utility and robust clinometric properties (Wade, 1992). However, due to its inherent complexity and because it is a multidimensional construct comprised of a number of discrete characteristics, gait cannot be represented by a single outcome. Although gait speed is sensitive to pathology, it is neither discriminative nor reflective of subtle and selective alterations of gait expressed in response to change in neuropathology in ageing and disease (Stolze et al., 2001, Verghese et al., 2007, Lord et al., 2014). Selective identification of gait characteristics is therefore critical for discrimination of pathology, identifying specific features of disease progression and discerning the effect of age for detection of shared neural correlates.
To allow for selectivity and specificity a comprehensive range of gait characteristics must be assessed, although there is high covariance among gait characteristics which needs to be accounted for. In response, several groups have proposed gait models that group gait characteristics into gait domains using data reduction techniques such as principle components analysis (Verghese et al., 2007, Hollman et al., 2011, Lord et al., 2013, Verlinden et al., 2013b). Whilst the models are comparable, there are subtle differences. For example Verghese and colleagues collated eight gait characteristics to form three domains; pace, variability and rhythm whereas Lord and colleagues obtained two further domains; asymmetry and postural control by collating 16 characteristics. Other models have produced more novel domains such as tandem and turning (Verlinden et al., 2013b) allowing for inclusion of more complex motor tasks. Independent domains of gait (and the characteristics thereof) can then be hypothesised to reflect independent neuroanatomical and functional substrates. Similarly for cognition, independent assessments are grouped to examine domains (Martin et al., 2013, Verlinden et al., 2013a, Lord et al., 2014) that are thought to represent independent neural substrates underlying cognitive functions.
Current understanding of disease pathology provides insight into potential associations of gait and cognition. For example, people with Alzheimer’s disease (AD) present foremost with deficits in amnestic ability predominantly due to amyloid deposition in the entorhinal cortex and hippocampus (Braak and Braak, 1995), with concordant findings of an association between atrophy of the hippocampus and decreased gait speed and step length (Callisaya et al., 2013). Similarly, people with Parkinson’s disease (PD) present with executive attention deficits due to compromised fronto-striatal circuitry (Stern et al., 1993, Burton et al., 2004) and attention also is significantly associated with reduced gait speed and step length (Lord et al., 2014). The question remains however if discrete gait domains share a different association dependent upon cognitive function? Furthermore, if the relationship between gait and cognitive variables is selective one may expect a different signature of impairments to emerge underpinned by the selective influence of pathology. A better understanding of this relationship would strengthen an understanding of the mechanisms of gait impairment, the shared neural and pathological correlates of gait and cognitive function and validate the role of gait as a surrogate biomarker of cognitive decline and pathology (Lord et al., 2014, Mollenhauer et al., 2014). The relationship between gait and cognition however is still an emerging area of work, largely due to recent advances in the understanding of gait and improvement in the ability to measure its discrete characteristics. To date a comprehensive investigation of the selective association between independent gait and cognition characteristics has not been undertaken.
The purpose of this review was therefore to undertake a detailed comparison of studies exploring discrete relationships of gait and cognitive domains. For this review, studies were limited to measuring gait under single task conditions. Gait under single task conditions reflects the ability of the cognitive system to control locomotion and to compensate for motor and cognitive deficit as a consequence of ageing and pathology. Therefore it is expected that changes in cognitive function would be reflected in changes in gait performance. Although dual task paradigm studies are extensively used to examine associations of gait and cognition, inconsistent findings are reported due to methodological issues such as diverse concurrent tasks, controlling for baseline task demand and different analysis of calculating dual task interference (Rochester et al., 2014). In addition to these inconsistencies, the underlying cognitive nature of dual task methodology remains unclear and does not reflect baseline cognitive influence on gait; therefore making it difficult to tease out direct underlying neural correlates.
In light of this, the aims of this review are to i) explore evidence for the associations between independent features of gait and cognitive function and ii) identify the longitudinal nature of relationships. We hypothesise that independent gait characteristics would be related to discrete cognitive functions in a specific rather than global manner and the pattern of association would be different with respect to pathology and ageing. In order to address this hypothesis we adopted a model of gait previously used in OA (Lord et al., 2013) (Fig. 1) and a comprehensive range of cognitive domains previously described (Emre et al., 2007) to improve consistency, reduce redundancy and retain independence between gait and cognitive features respectively to ease interpretation of results. We mapped individual gait domains (or respective characteristics) to individual cognitive functions to develop a matrix from which to identify discrete relationships. Three different cohorts were included: older adults (OA); those with cognitive impairment (CI); and people with PD, in order to explore gait-cognition associations in pathology and normal ageing. Cross-sectional and longitudinal study designs were examined to identify causality. It is hoped this review will not only to provide a clear understanding of current associations but in addition identify gaps in the literature to inform recommendations for future work.
Section snippets
Search strategy
Three databases were used for the search: Medline, Psychinfo and Scopus. For each of the databases used, three separate searches were performed for the three cohorts included in the review; OA, CI and PD. In total, nine separate searches were completed. Each search used the key terms ‘Gait’, ‘Cognition’ and either ‘Parkinson’s disease’, ‘dementia’ or ‘older adults’. For each of the key terms, a list of synonyms were correlated and entered into the search (Table 1). Where possible, MESH headings
Search yield
The search strategy generated a total of 43,828 papers; after exclusion criteria were applied the search strategy generated a total of 25,487 papers. After duplicates were removed, a total number of 22,128 papers were yielded from the search. The total number of papers were compiled into three databases; OA (n = 11609), CI (n = 7919) and PD (n = 2600). After the initial title screen, the total number of papers of interest for each group were OA (n = 168), CI (n = 119) and PD (n = 62).
After an abstract
Discussion
To our knowledge this is the first structured review to summarise the relationship between single task gait and cognition in older adults with and without cognitive impairment, and in PD. Key findings from this structured review are that for all groups the pace domain of gait (driven predominantly by gait speed) is associated with a broad range of cognitive functions but also selectively associated with executive attention. Gait speed is also a strong predictor of cognitive decline in OA,
Conflicts of Interest
No conflicts of interest are declared.
Acknowledgements
This research is supported by the National Institute for Health Research (NIHR) Newcastle Biomedical Research Unit (BRU) and centre (BRC) based at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The research was also supported by NIHR Newcastle CRF Infrastructure Funding. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.
References (108)
- et al.
Physical frailty predicts future cognitive decline—a four-year prospective study in 2737 cognitively normal older adults
J. Nutr. Health Aging
(2011) - et al.
Gait variability at fast-pace walking speed: a biomarker of mild cognitive impairment?
J. Nutr. Health Aging
(2013) - et al.
Theta band oscillation and synchrony in the hippocampal formation and associated structures: the case for its role in sensorimotor integration
Behav. Brain Res.
(2001) - et al.
Staging of alzheimer's disease-related neurofibrillary changes
Neurobiol. Aging
(1995) - et al.
Frequency of stages of Alzheimer-related lesions in different age categories
Neurobiol. Aging
(1997) - et al.
Stance time and step width variability have unique contributing impairments in older persons
Gait Posture
(2008) - et al.
Exploration of the neural substrates of executive functioning by functional neuroimaging
Neuroscience
(2006) - et al.
Insights into the neural control of locomotion from walking through doorways in Parkinson's disease
Neuropsychologia
(2010) - et al.
Cortical thickness is associated with gait disturbances in cerebral small vessel disease
NeuroImage
(2012) Dementia associated with Parkinson's disease
Lancet Neurol.
(2003)