Search

Accurate and reliable measurement of real-world walking activity is clinically relevant, particularly for people with mobility difficulties. Insights on walking can help understand mobility function, disease progression, and fall risks. People living in long-term residential care environments have heterogeneous and often pathological walking patterns, making it difficult for conventional algorithms paired with wearable sensors to detect their walking activity. We designed two walking bout detection algorithms for people living in long-term residential care. Both algorithms used thresholds on the magnitude of acceleration from a 3-axis accelerometer on the lower back to classify data as “walking” or “non-walking”. One algorithm had generic thresholds, whereas the other used personalized thresholds. To validate and evaluate the algorithms, we compared the classifications of walking/non-walking from our algorithms to the real-time research assistant annotated labels and the classification output from an algorithm validated on a healthy population. Both the generic and personalized algorithms had acceptable accuracy (0.83 and 0.82, respectively). The personalized algorithm showed the highest specificity (0.84) of all tested algorithms, meaning it was the best suited to determine input data for gait characteristic extraction. The developed algorithms were almost 60% quicker than the previously developed algorithms, suggesting they are adaptable for real-time processing.

Gait speed declines with age and slower walking speeds are associated with poor health outcomes. Understanding why we do not walk faster as we age, despite being able to, has implications for rehabilitation. Changes in regional oxygenated haemoglobin (HbO2) across the frontal lobe were monitored using functional near infrared spectroscopy in 17 young and 18 older adults while they walked on a treadmill for 5 min, alternating between 30 s of walking at a preferred and fast (120% preferred) speed. Gait was quantified using a triaxial accelerometer (lower back). Differences between task (preferred/fast) and group (young/old) and associations between regional HbO2 and gait were evaluated. Paired tests indicated increased HbO2 in the supplementary motor area (right) and primary motor cortex (left and right) in older adults when walking fast (p < 0.006). HbO2 did not significantly change in the young when walking fast, despite both groups modulating gait. When evaluating the effect of age (linear mixed effects model), greater increases in HbO2 were observed for older adults when walking fast (prefrontal cortex, premotor cortex, supplementary motor area and primary motor cortex) compared to young adults. In older adults, increased step length and reduced step length variability were associated with larger increases in HbO2 across multiple regions when walking fast. Walking fast required increased activation of motor regions in older adults, which may serve as a therapeutic target for rehabilitation. Widespread increases in HbO2 across the frontal cortex highlight that walking fast represents a resource-intensive task as we age.

Participating in habitual physical activity (HPA) may slow onset of dependency and disability for people with Parkinson’s disease (PwP). While cognitive and physical determinants of HPA are well understood, psychosocial influences are not. This pilot study aimed to identify psychosocial factors associated with HPA to guide future intervention development. Sixty-four PwP participated in this study; forty had carer informants. PwP participants wore a tri-axial accelerometer on the lower back continuously for seven days at two timepoints (18 months apart), measuring volume, pattern and variability of HPA. Linear mixed effects analysis identified relationships between demographic, clinical and psychosocial data and HPA from baseline to 18 months. Key results in PwP with carers indicated that carer anxiety and depression were associated with increased HPA volume (p < 0.01), while poorer carer self-care was associated with reduced volume of HPA over 18 months (p < 0.01). Greater carer strain was associated with taking longer walking bouts after 18 months (p < 0.01). Greater carer depression was associated with lower variability of HPA cross-sectionally (p = 0.009). This pilot study provides preliminary novel evidence that psychosocial outcomes from PwP’s carers may impact HPA in Parkinson’s disease. Interventions to improve HPA could target both PwP and carers and consider approaches that also support psychosocial wellbeing.

Laboratory-based gait assessments are indicative of clinical outcomes (e.g., disease identification). Real-world gait may be more sensitive to clinical outcomes, as impairments may be exaggerated in complex environments. This study aims to investigate how different environments (e.g., lab, real world) impact gait. Different walking bout lengths in the real world will be considered proxy measures of context. Data collected in different dementia disease subtypes will be analysed as disease-specific gait impairments are reported between these groups. Thirty-two people with cognitive impairment due to Alzheimer’s disease (AD), 28 due to dementia with Lewy bodies (DLB) and 25 controls were recruited. Participants wore a tri-axial accelerometer for six 10 m walks in lab settings, and continuously for seven days in the real world. Fourteen gait characteristics across five domains were measured (i.e., pace, variability, rhythm, asymmetry, postural control). In the lab, the DLB group showed greater step length variability (p = 0.008) compared to AD. Both subtypes demonstrated significant gait impairments (p < 0.01) compared to controls. In the real world, only very short walking bouts (

Falls are the leading cause of mortality, morbidity and poor quality of life in older adults with or without neurological conditions. Applying machine learning (ML) models to gait analysis outcomes offers the opportunity to identify individuals at risk of future falls. The aim of this study was to determine the effect of different data pre-processing methods on the performance of ML models to classify neurological patients who have fallen from those who have not for future fall risk assessment. Gait was assessed using wearables in clinic while walking 20 m at a self-selected comfortable pace in 349 (159 fallers, 190 non-fallers) neurological patients. Six different ML models were trained on data pre-processed with three techniques such as standardisation, principal component analysis (PCA) and path signature method. Fallers walked more slowly, with shorter strides and longer stride duration compared to non-fallers. Overall, model accuracy ranged between 48% and 98% with 43–99% sensitivity and 48–98% specificity. A random forest (RF) classifier trained on data pre-processed with the path signature method gave optimal classification accuracy of 98% with 99% sensitivity and 98% specificity. Data pre-processing directly influences the accuracy of ML models for the accurate classification of fallers. Using gait analysis with trained ML models can act as a tool for the proactive assessment of fall risk and support clinical decision-making.

Strong associations exist between quality of life and physical activity for those living in aged residential care (ARC). Suitable and reliable tools are required to quantify physical activity for descriptive and evaluative purposes. We calculated the number of days required for reliable walking outcomes indicative of physical activity in an ARC population using a trunk-worn device. ARC participants (n = 257) wore the device for up to 7 days. Reasons for data loss were also recorded. The volume, pattern, and variability of walking was calculated. For 197 participants who wore the device for at least 3 days, linear mixed models determined the impact of week structure and number of days required to achieve reliable outcomes, collectively and then stratified by care level. The average days recorded by the wearable device was 5.2 days. Day of the week did not impact walking activity. Depending on the outcome and level of care, 2–5 days was sufficient for reliable estimates. This study provides informative evidence for future studies aiming to use a wearable device located on the trunk to quantify physical activity walking out in the ARC population.

Wearable inertial sensors can be used to monitor mobility in real-world settings over extended periods. Although these technologies are widely used in human movement research, they have not yet been qualified by drug regulatory agencies for their use in regulatory drug trials. This is because the first generation of these sensors was unreliable when used on slow-walking subjects. However, intense research in this area is now offering a new generation of algorithms to quantify Digital Mobility Outcomes so accurate they may be considered as biomarkers in regulatory drug trials. This perspective paper summarises the work in the Mobilise-D consortium around the regulatory qualification of the use of wearable sensors to quantify real-world mobility performance in patients affected by Parkinson’s Disease. The paper describes the qualification strategy and both the technical and clinical validation plans, which have recently received highly supportive qualification advice from the European Medicines Agency. The scope is to provide detailed guidance for the preparation of similar qualification submissions to broaden the use of real-world mobility assessment in regulatory drug trials.

Parkinson’s disease (PD) is a common neurodegenerative disorder resulting in a range of mobility deficits affecting gait, balance and turning. In this paper, we present: (i) the development and validation of an algorithm to detect turns during gait; (ii) a method to extract turn characteristics; and (iii) the classification of PD using turn characteristics. Thirty-seven people with PD and 56 controls performed 180-degree turns during an intermittent walking task. Inertial measurement units were attached to the head, neck, lower back and ankles. A turning detection algorithm was developed and validated by two raters using video data. Spatiotemporal and signal-based characteristics were extracted and used for PD classification. There was excellent absolute agreement between the rater and the algorithm for identifying turn start and end (ICC ≥ 0.99). Classification modeling (partial least square discriminant analysis (PLS-DA)) gave the best accuracy of 97.85% when trained on upper body and ankle data. Balanced sensitivity (97%) and specificity (96.43%) were achieved using turning characteristics from the neck, lower back and ankles. Turning characteristics, in particular angular velocity, duration, number of steps, jerk and root mean square distinguished mild-moderate PD from controls accurately and warrant future examination as a marker of mobility impairment and fall risk in PD.

Neurological patients can have severe gait impairments that contribute to fall risks. Predicting falls from gait abnormalities could aid clinicians and patients mitigate fall risk. The aim of this study was to predict fall status from spatial-temporal gait characteristics measured by a wearable device in a heterogeneous population of neurological patients. Participants (n = 384, age 49–80 s) were recruited from a neurology ward of a University hospital. They walked 20 m at a comfortable speed (single task: ST) and while performing a dual task with a motor component (DT1) and a dual task with a cognitive component (DT2). Twenty-seven spatial-temporal gait variables were measured with wearable sensors placed at the lower back and both ankles. Partial least square discriminant analysis (PLS-DA) was then applied to classify fallers and non-fallers. The PLS-DA classification model performed well for all three gait tasks (ST, DT1, and DT2) with an evaluation of classification performance Area under the receiver operating characteristic Curve (AUC) of 0.7, 0.6 and 0.7, respectively. Fallers differed from non-fallers in their specific gait patterns. Results from this study improve our understanding of how falls risk-related gait impairments in neurological patients could aid the design of tailored fall-prevention interventions.

Parkinson’s disease (PD) is a common age-related neurodegenerative disease. Gait impairment is frequent in the later stages of PD contributing to reduced mobility and quality of life. Digital biomarkers such as gait velocity and step length are predictors of motor and cognitive decline in PD. Additional gait parameters may describe different aspects of gait and motor control in PD. Sample entropy (SampEnt), a measure of signal predictability, is a nonlinear approach that quantifies regularity of a signal. This study investigated SampEnt as a potential biomarker for PD and disease duration. Real-world gait data over a seven-day period were collected using an accelerometer (Axivity AX3, York, UK) placed on the low back and gait metrics extracted. SampEnt was determined for the stride time, with vector length and threshold parameters optimized. People with PD had higher stride time SampEnt compared to older adults, indicating reduced gait regularity. The range of SampEnt increased over 36 months for the PD group, although the mean value did not change. SampEnt was associated with dopaminergic medication dose but not with clinical motor scores. In conclusion, this pilot study indicates that SampEnt from real-world data may be a useful parameter reflecting clinical status although further research is needed involving larger populations.