BackgroundPhysical activity has not been objectively measured in prospective cohorts with sufficiently large numbers to reliably detect associations with multiple health outcomes. Technological advances now make this possible. We describe the methods used to collect and analyse accelerometer measured physical activity in over 100,000 participants of the UK Biobank study, and report variation by age, sex, day, time of day, and season.MethodsParticipants were approached by email to wear a wrist-worn accelerometer for seven days that was posted to them. Physical activity information was extracted from 100Hz raw triaxial acceleration data after calibration, removal of gravity and sensor noise, and identification of wear / non-wear episodes. We report age- and sex-specific wear-time compliance and accelerometer measured physical activity, overall and by hour-of-day, week-weekend day and season.Results103,712 datasets were received (44.8% response), with a median wear-time of 6.9 days (IQR:6.5–7.0). 96,600 participants (93.3%) provided valid data for physical activity analyses. Vector magnitude, a proxy for overall physical activity, was 7.5% (2.35mg) lower per decade of age (Cohen’s d = 0.9). Women had a higher vector magnitude than men, apart from those aged 45-54yrs. There were major differences in vector magnitude by time of day (d = 0.66). Vector magnitude differences between week and weekend days (d = 0.12 for men, d = 0.09 for women) and between seasons (d = 0.27 for men, d = 0.15 for women) were small.ConclusionsIt is feasible to collect and analyse objective physical activity data in large studies. The summary measure of overall physical activity is lower in older participants and age-related differences in activity are most prominent in the afternoon and evening. This work lays the foundation for studies of physical activity and its health consequences. Our summary variables are part of the UK Biobank dataset and can be used by researchers as exposures, confounding factors or outcome variables in future analyses.

Recently, deep learning (DL) methods have been introduced very successfully into human activity recognition (HAR) scenarios in ubiquitous and wearable computing. Especially the prospect of overcoming the need for manual feature design combined with superior classification capabilities render deep neural networks very attractive for real-life HAR applications. Even though DL-based approaches now outperform the state-of-the-art in a number of recognition tasks, still substantial challenges remain. Most prominently, issues with real-life datasets, typically including imbalanced datasets and problematic data quality, still limit the effectiveness of activity recognition using wearables. In this paper we tackle such challenges through Ensembles of deep Long Short Term Memory (LSTM) networks. LSTM networks currently represent the state-of-the-art with superior classification performance on relevant HAR benchmark datasets. We have developed modified training procedures for LSTM networks and combine sets of diverse LSTM learners into classifier collectives. We demonstrate that Ensembles of deep LSTM learners outperform individual LSTM networks and thus push the state-of-the-art in human activity recognition using wearables. Through an extensive experimental evaluation on three standard benchmarks (Opportunity, PAMAP2, Skoda) we demonstrate the excellent recognition capabilities of our approach and its potential for real-life applications of human activity recognition.

HighlightsRecent developments in automatically detecting compromised pig health and welfare.Five categories of behaviour are reviewed.Behaviours mapped to sensors that are feasible for automated detection.Progress towards levels of automation through detection, monitoring and fully automatic detection of behavioural change.Challenges for automated detection of behavioural changes are multifaceted and require trade-offs to develop such systems.

BackgroundPhysical activity and sedentary behaviour are key moderators of cardiovascular disease risk and metabolic control. Despite the importance of a physically active lifestyle, little is known about the effects of stroke on physical activity. We assessed physical activity and sedentary behaviour at three time points following stroke compared to a healthy control group.MethodsPhysical activity and sedentary behaviour were objectively measured using a portable multi-sensor array in 31 stroke participants (73±9 years, National Institute of Health Stroke Scale 2±2, mobile 10 metres with/without aid) within seven days and at three and six months. Stroke data were compared with an age, sex and body mass index matched healthy control group (n = 31).ResultsWithin seven days of stroke, total energy expenditure and physical activity were significantly lower and sedentary time higher in the stroke group compared to controls (total energy expenditure 1840±354 vs. 2220±489 kcal, physical activity 28±32 vs. 79±46 min/day, steps 3111±2290 vs. 7996±2649, sedentary time 1383±42 vs. 1339±44 min/day, p<0.01). At three months physical activity levels had increased (64±58 min/day) but plateaued by six months (66±68 min/day).ConclusionsPhysical activity levels are reduced immediately post-stroke and remain below recommended levels for health and wellbeing at the three and six month time points. Clinicians should explore methods to increase physical activity and reduce sedentary behaviour in both the acute and later stages following stroke.

Preterm birth is associated with increased risks of neurological and motor impairments such as cerebral palsy. The risks are highest in those born at the lowest gestations. Early identification of those most at risk is challenging meaning that a critical window of opportunity to improve outcomes through therapy-based interventions may be missed. Clinically, the assessment of spontaneous general movements is an important tool, which can be used for the prediction of movement impairments in high risk infants. Movement recognition aims to capture and analyze relevant limb movements through computerized approaches focusing on continuous, objective, and quantitative assessment. Different methods of recording and analyzing infant movements have recently been explored in high risk infants. These range from camera-based solutions to body-worn miniaturized movement sensors used to record continuous time-series data that represent the dynamics of limb movements. Various machine learning methods have been developed and applied to the analysis of the recorded movement data. This analysis has focused on the detection and classification of atypical spontaneous general movements. This article aims to identify recent translational studies using movement recognition technology as a method of assessing movement in high risk infants. The application of this technology within pediatric practice represents a growing area of inter-disciplinary collaboration, which may lead to a greater understanding of the development of the nervous system in infants at high risk of motor impairment.