Longitudinal changes over thirty-six months in postural control dynamics and cognitive function in people with Parkinson's disease

Highlights

• Trunk acceleration regularity was less in people with Parkinson’s disease.

• No decrease in trunk acceleration regularity was found with disease progression.

• Weak correlations were found between trunk dynamics and cognitive function.

• Weak correlations were found between trunk dynamics and clinical motor score.

Abstract

Background

Postural instability is a common motor feature in people with Parkinson's disease (PD) together with non-motor features such as cognitive dysfunction. Management of postural instability is challenging as it is often resistant to dopaminergic therapy. Greater knowledge of postural control is essential to understand postural instability in PD.

Research question

This study aimed to answer how postural control differs in people with PD compared to healthy older adults (HOA). Additionally, postural control changes over a 36 month period and its relationship to cognitive impairment and motor scores were investigated.

Methods

The study group consisted of 50 people diagnosed with PD and 59 HOAs, recruited as part of an incident cohort study (ICICLE-GAIT). Participants stood still for 2 min, eyes open and arms by their side. A single tri-axial accelerometer (Axivity AX3, York, UK) on the lower back recorded acceleration. Measurements were taken at 18, 36 and 54 months after recruitment. Sample entropy (SampEn), which measures signal predictability, was determined for the accelerometry data. Cognitive tests included the Montreal Cognitive Assessment and the Unified Parkinson’s Disease Rating Scale (UPDRS III) quantified motor function. Linear mixed models, regression analysis and correlation analysis were applied to the data.

Results

indicated that SampEn was greater for the PD group at all three time-points and along all three axes. However, there was no increase of SampEn with disease progression. Higher SampEn values were associated with greater cognitive impairment and lower UPDRS III, although correlations were weak. There was a difference between axial directions and cognitive and motor scores.

Significance

People with PD exhibit decreased regularity of trunk dynamics when standing compared to HOAs. Nonlinear accelerometer metrics along all three axes are therefore a potential biomarker of PD. The relationship between trunk dynamics and cognitive function indicates common neural pathways.

Introduction

Parkinson’s disease (PD) is the second most common neurodegenerative disease with a worldwide prevalence greater than 1% in adults over 70 [1]. Early motor symptoms include rigidity, bradykinesia, tremor and gait impairment with postural instability emerging as the disease progresses [2]. Postural instability is a known predictor of falls with consequential impact on wellbeing [3]. Over 20% of people with PD have also been reported to have mild cognitive impairment (MCI) at initial diagnosis [4]. Changes in cognitive function are associated with postural instability [5,6] and with an increased risk of falls [7,8]. A neuroimaging study supports the link between cognitive function and falling in PD with reduced grey matter reported in the posterior caudate (associated with cognitive function) in fallers compared to fallers [8]. Management of postural instability is challenging as it is often resistant to dopaminergic therapy [9]. Nonpharmaceutical interventions include different treadmill training protocols which show limited improvements in gait and balance [10,11].

Clinically, assessment of postural instability is the Pull test, included in the Unified Parkinson’s Disease Rating Scale (UPDRS III), which is insensitive to small changes in postural stability [12]. Objectively, postural stability has been assessed through centre of pressure (COP) parameters recorded from a force platform [13] and more recently, from body-worn monitors such as accelerometers [[14], [15], [16]]. The derived postural parameters are generally linear, providing information about signal characteristics averaged across time. However, as postural stability is a function of postural control, parameters that reflect underlying neural control mechanisms may yield additional information. For example, people with PD have reduced automatic control and greater conscious control of posture [17,18]. This altered regulatory mechanism may present as a change in regularity or predictability of postural parameters.

The purpose of this study was to investigate predictability of postural trunk dynamics through nonlinear analysis of tri-axial accelerometer signals. We included vertical accelerometry data as postural control operates in three dimensions and changes in perception of the postural vertical in older adults have been postulated [19]. We compared postural trunk predictability in people with recently diagnosed PD to postural trunk predictability in healthy older adults (HOA). Additionally, we investigated changes in this parameter over a 36 month period and its relationship to cognitive dysfunction. We hypothesised that (1) Postural trunk predictability is different in people with PD due to the loss of automatic control; (2) Postural trunk predictability changes over time in people with PD, as a result of increasing loss of automatic motor behaviour; (3) Postural trunk predictability is correlated with clinical measures of cognitive dysfunction, reflecting common dysfunctional neural pathways. This analysis may offer information about the underlying postural control mechanisms, its application as an early biomarker and as an indicator of disease progression.