Freezing of gait assessment with inertial measurement units and deep learning: effect of tasks, medication states, and stops.

BACKGROUND: Freezing of gait (FOG) is an episodic and highly disabling symptom of Parkinson's Disease (PD). Traditionally, FOG assessment relies on time-consuming visual inspection of camera footage. Therefore, previous studies have proposed portable and automated solutions to annotate FOG. However, automated FOG assessment is challenging due to gait variability caused by medication effects and varying FOG-provoking tasks. Moreover, whether automated approaches can differentiate FOG from typical everyday movements, such as volitional stops, remains to be determined. To address these questions, we evaluated an automated FOG assessment model with deep learning (DL) based on inertial measurement units (IMUs). We assessed its performance trained on all standardized FOG-provoking tasks and medication states, as well as on specific tasks and medication states. Furthermore, we examined the effect of adding stopping periods on FOG detection performance. METHODS: Twelve PD patients with self-reported FOG (mean age 69.33 ± 6.02 years) completed a FOG-provoking protocol, including timed-up-and-go and 360-degree turning-in-place tasks in On/Off dopaminergic medication states with/without volitional stopping. IMUs were attached to the pelvis and both sides of the tibia and talus. A temporal convolutional network (TCN) was used to detect FOG episodes. FOG severity was quantified by the percentage of time frozen (%TF) and the number of freezing episodes (#FOG). The agreement between the model-generated outcomes and the gold standard experts' video annotation was assessed by the intra-class correlation coefficient (ICC). RESULTS: For FOG assessment in trials without stopping, the agreement of our model was strong (ICC (%TF) = 0.92 [0.68, 0.98]; ICC(#FOG) = 0.95 [0.72, 0.99]). Models trained on a specific FOG-provoking task could not generalize to unseen tasks, while models trained on a specific medication state could generalize to unseen states. For assessment in trials with stopping, the agreement of our model was moderately strong (ICC (%TF) = 0.95 [0.73, 0.99]; ICC (#FOG) = 0.79 [0.46, 0.94]), but only when stopping was included in the training data. CONCLUSION: A TCN trained on IMU signals allows valid FOG assessment in trials with/without stops containing different medication states and FOG-provoking tasks. These results are encouraging and enable future work investigating automated FOG assessment during everyday life.

Connecting real-world digital mobility assessment to clinical outcomes for regulatory and clinical endorsement-the Mobilise-D study protocol.

BACKGROUND: The development of optimal strategies to treat impaired mobility related to ageing and chronic disease requires better ways to detect and measure it. Digital health technology, including body worn sensors, has the potential to directly and accurately capture real-world mobility. Mobilise-D consists of 34 partners from 13 countries who are working together to jointly develop and implement a digital mobility assessment solution to demonstrate that real-world digital mobility outcomes have the potential to provide a better, safer, and quicker way to assess, monitor, and predict the efficacy of new interventions on impaired mobility. The overarching objective of the study is to establish the clinical validity of digital outcomes in patient populations impacted by mobility challenges, and to support engagement with regulatory and health technology agencies towards acceptance of digital mobility assessment in regulatory and health technology assessment decisions. METHODS/DESIGN: The Mobilise-D clinical validation study is a longitudinal observational cohort study that will recruit 2400 participants from four clinical cohorts. The populations of the Innovative Medicine Initiative-Joint Undertaking represent neurodegenerative conditions (Parkinson's Disease), respiratory disease (Chronic Obstructive Pulmonary Disease), neuro-inflammatory disorder (Multiple Sclerosis), fall-related injuries, osteoporosis, sarcopenia, and frailty (Proximal Femoral Fracture). In total, 17 clinical sites in ten countries will recruit participants who will be evaluated every six months over a period of two years. A wide range of core and cohort specific outcome measures will be collected, spanning patient-reported, observer-reported, and clinician-reported outcomes as well as performance-based outcomes (physical measures and cognitive/mental measures). Daily-living mobility and physical capacity will be assessed directly using a wearable device. These four clinical cohorts were chosen to obtain generalizable clinical findings, including diverse clinical, cultural, geographical, and age representation. The disease cohorts include a broad and heterogeneous range of subject characteristics with varying chronic care needs, and represent different trajectories of mobility disability. DISCUSSION: The results of Mobilise-D will provide longitudinal data on the use of digital mobility outcomes to identify, stratify, and monitor disability. This will support the development of widespread, cost-effective access to optimal clinical mobility management through personalised healthcare. Further, Mobilise-D will provide evidence-based, direct measures which can be endorsed by regulatory agencies and health technology assessment bodies to quantify the impact of disease-modifying interventions on mobility. TRIAL REGISTRATION: ISRCTN12051706.

Automated freezing of gait assessment with marker-based motion capture and multi-stage spatial-temporal graph convolutional neural networks.

BACKGROUND: Freezing of gait (FOG) is a common and debilitating gait impairment in Parkinson's disease. Further insight into this phenomenon is hampered by the difficulty to objectively assess FOG. To meet this clinical need, this paper proposes an automated motion-capture-based FOG assessment method driven by a novel deep neural network. METHODS: Automated FOG assessment can be formulated as an action segmentation problem, where temporal models are tasked to recognize and temporally localize the FOG segments in untrimmed motion capture trials. This paper takes a closer look at the performance of state-of-the-art action segmentation models when tasked to automatically assess FOG. Furthermore, a novel deep neural network architecture is proposed that aims to better capture the spatial and temporal dependencies than the state-of-the-art baselines. The proposed network, termed multi-stage spatial-temporal graph convolutional network (MS-GCN), combines the spatial-temporal graph convolutional network (ST-GCN) and the multi-stage temporal convolutional network (MS-TCN). The ST-GCN captures the hierarchical spatial-temporal motion among the joints inherent to motion capture, while the multi-stage component reduces over-segmentation errors by refining the predictions over multiple stages. The proposed model was validated on a dataset of fourteen freezers, fourteen non-freezers, and fourteen healthy control subjects. RESULTS: The experiments indicate that the proposed model outperforms four state-of-the-art baselines. Moreover, FOG outcomes derived from MS-GCN predictions had an excellent (r = 0.93 [0.87, 0.97]) and moderately strong (r = 0.75 [0.55, 0.87]) linear relationship with FOG outcomes derived from manual annotations. CONCLUSIONS: The proposed MS-GCN may provide an automated and objective alternative to labor-intensive clinician-based FOG assessment. Future work is now possible that aims to assess the generalization of MS-GCN to a larger and more varied verification cohort.

Exercise- and strategy-based physiotherapy-delivered intervention for preventing repeat falls in people with Parkinson's: the PDSAFE RCT.

BACKGROUND: People with Parkinson's disease are twice as likely to experience a fall as a healthy older person, often leading to debilitating effects on confidence, activity levels and quality of life. OBJECTIVE: To estimate the effect of a physiotherapy programme for fall prevention among people with Parkinson's disease. DESIGN: A multicentre, pragmatic, investigator-masked, individually randomised controlled trial (RCT) with prespecified subgroup analyses. SETTING: Recruitment from NHS hospitals and clinics and community and social services in eight English regions with home-based interventions. PARTICIPANTS: A total of 474 people with Parkinson's disease (i.e. Hoehn and Yahr scale stages 1-4) were recruited: 238 were assigned to a physiotherapy programme and 236 were assigned to usual care. Random allocation was 50 : 50. INTERVENTIONS: All participants received routine care; the usual-care group received an information digital versatile disc (DVD) and a single advice session at trial completion. The intervention group had an individually tailored, progressive, home-based fall avoidance strategy training programme with balance and strengthening exercises: PDSAFE. MAIN OUTCOME MEASURES: The primary outcome was the risk of repeat falling, collected by self-report monthly diaries between 0 and 6 months after randomisation. Secondary outcomes included near-falls, falls efficacy, freezing of gait (FoG), health-related quality of life, and measurements taken using the Mini-Balance Evaluation Systems Test (Mini-BESTest), the Chair Stand Test (CST), the Geriatric Depression Scale, the Physical Activity Scale for the Elderly and the Parkinson's Disease Questionnaire. RESULTS: PDSAFE is the largest RCT of falls management among people with Parkinson's disease: 541 patients were screened for eligibility. The average age was 72 years, and 266 out of 474 (56%) participants were men. Of the 474 randomised participants, 238 were randomised to the intervention group and 236 were randomised to the control group. No difference in repeat falling within 6 months of randomisation was found [PDSAFE group to control group odds ratio (OR) 1.21, 95% confidence interval (CI) 0.74 to 1.98; p = 0.447]. An analysis of secondary outcomes demonstrated better balance (Mini-BESTest: mean difference 0.95, 95% CI 0.24 to 1.67; p = 0.009), functional strength (CST: p = 0.041) and falls efficacy (Falls Efficacy Scale - International: mean difference 1.6, 95% CI -3.0 to -0.19; p = 0.026) with near-falling significantly reduced with PDSAFE (OR 0.67, 95% CI 0.53 to 0.86; p = 0.001) at 6 months. Prespecified subgroup analysis (i.e. disease severity and FoG) revealed a PDSAFE differing effect; the intervention may be of benefit for people with moderate disease but may increase falling for those in the more severe category, especially those with FoG. LIMITATIONS: All participants were assessed at primary outcome; only 73% were assessed at 12 months owing to restricted funding. CONCLUSIONS: PDSAFE was not effective in reducing repeat falling across the range of people with Parkinson's disease in the trial. Secondary analysis demonstrated that other functional tasks and self-efficacy improved and demonstrated differential patterns of intervention impact in accordance with disease severity and FoG, which supports previous secondary research findings and merits further primary evaluation. FUTURE WORK: Further trials of falls prevention on targeted groups of people with Parkinson's disease are recommended. TRIAL REGISTRATION: Current Controlled Trials ISRCTN48152791. FUNDING: This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 23, No. 36. See the NIHR Journals Library website for further project information. Sarah E Lamb is funded by the NIHR Collaboration for Leadership in Applied Health Research and Care (CLAHRC) at Oxford Health NHS Foundation Trust, the NIHR Oxford Biomedical Research Centre at the Oxford University Hospitals NHS Foundation Trust and CLAHRC Oxford. Victoria A Goodwin is supported by the NIHR Collaborations for Leadership in Applied Health Research and Care in the South West Peninsula (PenCLAHRC). Lynn Rochester is supported by the NIHR Newcastle Biomedical Research Centre based at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The research was also supported by the NIHR Newcastle Clinical Research Facility Infrastructure funding. Helen C Roberts is supported by CLAHRC Wessex and the NIHR Southampton Biomedical Research Centre.

People with Parkinson's disease fall often. Falls are scary and make moving about harder. The PDSAFE trial tested a new 'home physiotherapy' programme for reducing falls. People with Parkinson's disease were allocated to one of two groups by chance: they either received the PDSAFE exercises or just normal care. The costs were looked at and people were asked for their views of the PDSAFE exercises. To take part, people had to have Parkinson's disease, live in their own home, be able to walk, have had at least one fall in the previous year and pass a memory test. PDSAFE was taught by physiotherapists and included exercises and fall avoidance strategies. Everyone had to record falls on a monthly calendar, and balance, strength and walking were tested. To our knowledge, this was the largest falls trial looking at people with Parkinson's disease in the world: 541 people took part. The number of falls an individual reported differed a lot between people. When all people with Parkinson's disease in the trial were considered, the physiotherapy programme did not reduce falls in the first 6 months. However, it was found that some people had fewer falls after taking part in the exercises, whereas others did not. Those with more severe Parkinson's disease (i.e. problems with movement, memory and freezing of gait) fell more often after the PDSAFE intervention, even though their balance and confidence improved. Those with good memory, moderate disease and two or three falls in the previous year reacted well to PDSAFE and had fewer falls. It was found that PDSAFE reduced near-falls (about to fall but managed to save themselves) and improved balance and confidence. The physiotherapists and those who took part liked the programme and felt that it helped, but it was expensive to run. In conclusion, a falls prevention programme should be based on each person's needs and a different treatment should be used for those with more severe Parkinson's disease.

A roadmap for implementation of patient-centered digital outcome measures in Parkinson's disease obtained using mobile health technologies.

Obtaining reliable longitudinal information about everyday functioning from individuals with Parkinson's disease (PD) in natural environments is critical for clinical care and research. Despite advances in mobile health technologies, the implementation of digital outcome measures is hindered by a lack of consensus on the type and scope of measures, the most appropriate approach for data capture (eg, in clinic or at home), and the extraction of timely information that meets the needs of patients, clinicians, caregivers, and health care regulators. The Movement Disorder Society Task Force on Technology proposes the following objectives to facilitate the adoption of mobile health technologies: (1) identification of patient-centered and clinically relevant digital outcomes; (2) selection criteria for device combinations that offer an acceptable benefit-to-burden ratio to patients and that deliver reliable, clinically relevant insights; (3) development of an accessible, scalable, and secure platform for data integration and data analytics; and (4) agreement on a pathway for approval by regulators, adoption into e-health systems and implementation by health care organizations. We have developed a tentative roadmap that addresses these needs by providing the following deliverables: (1) results and interpretation of an online survey to define patient-relevant endpoints, (2) agreement on the selection criteria for use of device combinations, (3) an example of an open-source platform for integrating mobile health technology output, and (4) recommendations for assessing readiness for deployment of promising devices and algorithms suitable for regulatory approval. This concrete implementation guidance, harmonizing the collaborative endeavor among stakeholders, can improve assessments of individuals with PD, tailor symptomatic therapy, and enhance health care outcomes. © 2019 International Parkinson and Movement Disorder Society.

A multi-centre, randomised controlled trial of the effectiveness of PDSAFE to prevent falls among people with Parkinson's: study protocol.

BACKGROUND: Falls amongst people with Parkinson's (PwP) result in significant disability and reduced quality of life. There is emerging evidence that exercise-based and physiotherapeutic interventions are of benefit for improving fall risk factors, such as balance. However, the benefit, in terms of preventing falls, is mixed. The development of effective interventions has been identified as the highest research priority for this population. The aim of this trial is to establish the effectiveness and cost-effectiveness of a novel, home-based physiotherapy programme, compared with usual care, on falls amongst PwP. METHODS/DESIGN: A UK multi-centre, community-based, single blind, randomised controlled trial with twelve month follow-up, and nested economic evaluation and qualitative studies will be undertaken. Six hundred PwP who live in their own home, have had one or more falls in the previous year and an MMSE score of ≥24 will be recruited. Those living in care homes and those needing assistance from another person to walk indoors will not be eligible. The intervention is a physiotherapist delivered, individually tailored and progressive, home-based programme (PDSAFE) comprising task orientated movement strategy training, functional lower limb strengthening and balance training, of six months duration. Unsupervised daily home exercises and strategies will be practised and supported using technology. Control participants will receive usual care. Data collection will include falls, cognitive state, balance and mobility, fear of falling, freezing of gait, mood, quality of life, carer quality of life and resource use. Data will be collected at baseline, three, six and twelve months. Longitudinal semi-structured interviews will be undertaken with forty participants to explore the expectations and experiences of participants. The primary outcome is risk of repeat falling at six months post-randomisation. DISCUSSION: The aims of this trial are to establish the effectiveness and cost-effectiveness of a novel, home-delivered physiotherapy intervention (PDSAFE) compared with usual care on risk of falling for PwP who have a history of falling. PDSAFE is a novel intervention that builds upon the existing literature and targeting known risk factors, being the first study that uses a novel delivery modus (technology) in conjunction with traditional physiotherapeutic approaches. TRIAL REGISTRATION: Current Controlled Trials ISRCTN48152791.

Age-related differences in attentional cost associated with postural dual tasks: increased recruitment of generic cognitive resources in older adults.

Dual-task designs have been used widely to study the degree of automatic and controlled processing involved in postural stability of young and older adults. However, several unexplained discrepancies in the results weaken this literature. To resolve this problem, a careful selection of dual-task studies that met certain methodological criteria are considered with respect to reported interactions of age (young vs. older adults)×task (single vs. dual task) in stable and unstable postural conditions. Our review shows that older adults are able to perform a postural dual task as well as younger adults in stable conditions. However, when the complexity of the postural task is increased by dynamic conditions (surface and surround), performance in postural, concurrent, or both tasks is more affected in older relative to young adults. In light of neuroimaging studies and new conceptual frameworks, these results demonstrate an age-related increase of controlled processing of standing associated with greater intermittent adjustments.

The effect of cues on gait variability--reducing the attentional cost of walking in people with Parkinson's disease.

Parkinson's disease (PD) subjects have increased gait variability, reflecting greater attentional demand during walking. This study aimed to investigate the attentional cost of three cueing strategies by examining their effect on gait variability. Fourteen PD and 12 age matched control subjects were studied under single and dual walking tasks. Gait variability of PD subjects tended to reduce with all cues, the most consistent reductions in variability seen with a combination of an attentional strategy (focussing on big steps) and a rhythmical auditory cue. The reduction in gait variability of PD subjects with cues, suggests they may reduce the attentional cost of walking.

Reliability and validity of trunk assessment for people with multiple sclerosis.

BACKGROUND AND PURPOSE: Standardized scales are a prerequisite for rehabilitation and research. This study was designed to determine the reliability and validity of scores on items of the trunk assessment of the Melsbroek Disability Scoring Test (MDST) and Trunk Impairment Scale (TIS) in people with multiple sclerosis (MS). SUBJECTS: Thirty people with MS participated in the study. METHODS: Interrater and test-retest reliability and construct validity were assessed. RESULTS: Kappa and weighted kappa values for the items of the trunk assessment of the MDST ranged from .74 to .95, and the kappa and weighted kappa values for the TIS items ranged from .46 to 1.00. Intraclass correlation coefficients for interrater and test-retest agreement were .93 and .92, respectively, for the trunk assessment of the MDST and .97 and .95, respectively, for the TIS. Bland-Altman analysis showed consistency of scores without observer bias. Construct validity was established. DISCUSSION AND CONCLUSION: The MDST and TIS provide reliable assessments of the trunk and are valid scales for measuring trunk performance in people with MS.