Machine learning for activity recognition : hip versus wrist data

Problem addressed Wrist-worn accelerometers are associated with greater compliance. However, validated algorithms for predicting activity type from wrist-worn accelerometer data are lacking. This study compared the activity recognition rates of an activity classifier trained on acceleration signal collected on the wrist and hip. Methodology 52 children and adolescents (mean age 13.7 +/- 3.1 year) completed 12 activity trials that were categorized into 7 activity classes: lying down, sitting, standing, walking, running, basketball, and dancing. During each trial, participants wore an ActiGraph GT3X+ tri-axial accelerometer on the right hip and the non-dominant wrist. Features were extracted from 10-s windows and inputted into a regularized logistic regression model using R (Glmnet + L1). Results Classification accuracy for the hip and wrist was 91.0% +/- 3.1% and 88.4% +/- 3.0%, respectively. The hip model exhibited excellent classification accuracy for sitting (91.3%), standing (95.8%), walking (95.8%), and running (96.8%); acceptable classification accuracy for lying down (88.3%) and basketball (81.9%); and modest accuracy for dance (64.1%). The wrist model exhibited excellent classification accuracy for sitting (93.0%), standing (91.7%), and walking (95.8%); acceptable classification accuracy for basketball (86.0%); and modest accuracy for running (78.8%), lying down (74.6%) and dance (69.4%). Potential Impact Both the hip and wrist algorithms achieved acceptable classification accuracy, allowing researchers to use either placement for activity recognition.

Prediction of activity type in preschool children using machine learning techniques

Objectives Recent research has shown that machine learning techniques can accurately predict activity classes from accelerometer data in adolescents and adults. The purpose of this study is to develop and test machine learning models for predicting activity type in preschool-aged children. Design Participants completed 12 standardised activity trials (TV, reading, tablet game, quiet play, art, treasure hunt, cleaning up, active game, obstacle course, bicycle riding) over two laboratory visits. Methods Eleven children aged 3–6 years (mean age = 4.8 ± 0.87; 55% girls) completed the activity trials while wearing an ActiGraph GT3X+ accelerometer on the right hip. Activities were categorised into five activity classes: sedentary activities, light activities, moderate to vigorous activities, walking, and running. A standard feed-forward Artificial Neural Network and a Deep Learning Ensemble Network were trained on features in the accelerometer data used in previous investigations (10th, 25th, 50th, 75th and 90th percentiles and the lag-one autocorrelation). Results Overall recognition accuracy for the standard feed forward Artificial Neural Network was 69.7%. Recognition accuracy for sedentary activities, light activities and games, moderate-to-vigorous activities, walking, and running was 82%, 79%, 64%, 36% and 46%, respectively. In comparison, overall recognition accuracy for the Deep Learning Ensemble Network was 82.6%. For sedentary activities, light activities and games, moderate-to-vigorous activities, walking, and running recognition accuracy was 84%, 91%, 79%, 73% and 73%, respectively. Conclusions Ensemble machine learning approaches such as Deep Learning Ensemble Network can accurately predict activity type from accelerometer data in preschool children.

The effect of the addition of hip strengthening exercises to a lumbopelvic exercise programme for the treatment of non-specific low back pain : a randomized controlled trial

Objectives To compare the efficacy of two exercise programs in reducing pain and disability for individuals with non-specific low back pain and to examine the underlying mechanical factors related to pain and disability for individuals with NSLBP. Design A single-blind, randomized controlled trial. Methods: Eighty participants were recruited from eleven community-based general medical practices and randomized into two groups completing either a lumbopelvic motor control or a combined lumbopelvic motor control and progressive hip strengthening exercise therapy program. All participants received an education session, 6 rehabilitation sessions including real time ultrasound training, and a home based exercise program manual and log book. The primary outcomes were pain (0-100mm visual analogue scale), and disability (Oswestry Disability Index V2). The secondary outcomes were hip strength (N/kg) and two-dimensional frontal plane biomechanics (°) measure during the static Trendelenburg test and while walking. All outcomes were measured at baseline and at 6-week follow up. Results There was no statistical difference in the change in pain (xˉ = -4.0mm, t= -1.07, p =0.29, 95%CI -11.5, 3.5) or disability (xˉ = -0.3%, t= -0.19, p =0.85, 95%CI -3.5, 2.8) between groups. Within group comparisons revealed clinically meaningful reductions in pain for both Group One (xˉ =-20.9mm, 95%CI -25.7, -16.1) and Group Two (xˉ =-24.9, 95%CI -30.8, -19.0). Conclusion Both exercise programs had similar efficacy in reducing pain. The addition of hip strengthening exercises to a motor control exercise program does not appear to result in improved clinical outcome for pain for individuals with non-specific low back pain.

The potential for new understandings of normal and abnormal cognition by integration of neuroimaging and behavioral data : Not an exercise in carrying coals to Newcastle

Discovering the means to prevent and cure schizophrenia is a vision that motivates many scientists. But in order to achieve this goal, we need to understand its neurobiological basis. The emergent metadiscipline of cognitive neuroscience fields an impressive array of tools that can be marshaled towards achieving this goal, including powerful new methods of imaging the brain (both structural and functional) as well as assessments of perceptual and cognitive capacities based on psychophysical procedures, experimental tasks and models developed by cognitive science. We believe that the integration of data from this array of tools offers the greatest possibilities and potential for advancing understanding of the neural basis of not only normal cognition but also the cognitive impairments that are fundamental to schizophrenia. Since sufficient expertise in the application of these tools and methods rarely reside in a single individual, or even a single laboratory, collaboration is a key element in this endeavor. Here, we review some of the products of our integrative efforts in collaboration with our colleagues on the East Coast of Australia and Pacific Rim. This research focuses on the neural basis of executive function deficits and impairments in early auditory processing in patients using various combinations of performance indices (from perceptual and cognitive paradigms), ERPs, fMRI and sMRI. In each case, integration of two or more sources of information provides more information than any one source alone by revealing new insights into structure-function relationships. Furthermore, the addition of other imaging methodologies (such as DTI) and approaches (such as computational models of cognition) offers new horizons in human brain imaging research and in understanding human behavior.

Cold water immersion enhances recovery of submaximal muscle function after resistance exercise

We investigated the effect of cold water immersion (CWI) on the recovery of muscle function and physiological responses following high-intensity resistance exercise. Using a randomized, cross-over design, 10 physically active men performed high-intensity resistance exercise, followed by one of two recovery interventions: 10 min of cold water immersion at 10°C, or 10 min active recovery (low-intensity cycling). After the recovery interventions, maximal muscle function was assessed after 2 h and 4 h by measuring jump height and isometric squat strength. Submaximal muscle function was assessed after 6 h by measuring the average load lifted during six sets of 10 squats at 80% 1RM. Intramuscular temperature (1 cm) was also recorded, and venous blood samples were analyzed for markers of metabolism, vasoconstriction and muscle damage. CWI did not enhance recovery of maximal muscle function. However, during the final three sets of the submaximal muscle function test, the participants lifted a greater load (p<0.05; 38%; Cohen’s d 1.3) following CWI compared with active recovery. During CWI, muscle temperature decreased 6°C below post-exercise values, and remained below pre-exercise values for another 35 min. Venous blood O2 saturation decreased below pre-exercise values for 1.5 h after CWI. Serum endothelin-1 concentration did not change after CWI, whereas it decreased after active recovery. Plasma myoglobin concentration was lower, whereas plasma interleukin-6 concentration was higher after CWI compared with active recovery. These results suggest that cold water immersion after resistance exercise allow athletes to complete more work during subsequent training sessions, which could enhance long-term training adaptations.

Metabolic and hormonal responses to isoenergetic high-intensity interval exercise and continuous moderate-intensity exercise

This study investigated the effects of high-intensity interval training (HIIT) vs. work-matched moderate-intensity continuous exercise (MOD) on metabolism and counterregulatory stress hormones. In a randomized and counterbalanced order, 10 well-trained male cyclists and triathletes completed a HIIT session [81.6 ± 3.7% maximum oxygen consumption (V̇o2 max); 72.0 ± 3.2% peak power output; 792 ± 95 kJ] and a MOD session (66.7 ± 3.5% V̇o2 max; 48.5 ± 3.1% peak power output; 797 ± 95 kJ). Blood samples were collected before, immediately after, and 1 and 2 h postexercise. Carbohydrate oxidation was higher (P = 0.037; 20%), whereas fat oxidation was lower (P = 0.037; −47%) during HIIT vs. MOD. Immediately after exercise, plasma glucose (P = 0.024; 20%) and lactate (P < 0.01; 5.4×) were higher in HIIT vs. MOD, whereas total serum free fatty acid concentration was not significantly different (P = 0.33). Targeted gas chromatography-mass spectromtery metabolomics analysis identified and quantified 49 metabolites in plasma, among which 11 changed after both HIIT and MOD, 13 changed only after HIIT, and 5 changed only after MOD. Notable changes included substantial increases in tricarboxylic acid intermediates and monounsaturated fatty acids after HIIT and marked decreases in amino acids during recovery from both trials. Plasma adrenocorticotrophic hormone (P = 0.019), cortisol (P < 0.01), and growth hormone (P < 0.01) were all higher immediately after HIIT. Plasma norepinephrine (P = 0.11) and interleukin-6 (P = 0.20) immediately after exercise were not significantly different between trials. Plasma insulin decreased during recovery from both HIIT and MOD (P < 0.01). These data indicate distinct differences in specific metabolites and counterregulatory hormones following HIIT vs. MOD and highlight the value of targeted metabolomic analysis to provide more detailed insights into the metabolic demands of exercise.

Ibuprofen supplementation and its effects on NF-KB activation in skeletal muscle following resistance exercise

Resistance exercise triggers a subclinical inflammatory response that plays a pivotal role in skeletal muscle regeneration. Nuclear factor‐κB (NF‐κB) is a stress signalling transcription factor that regulates acute and chronic states of inflammation. The classical NF‐κB pathway regulates the early activation of post‐exercise inflammation; however there remains scope for this complex transcription factor to play a more detailed role in post‐exercise muscle recovery. Sixteen volunteers completed a bout of lower body resistance exercise with the ingestion of three 400 mg doses of ibuprofen or a placebo control. Muscle biopsy samples were obtained prior to exercise and at 0, 3 and 24 h post‐exercise and analysed for key markers of NF‐κB activity. Phosphorylated p65 protein expression and p65 inflammatory target genes were elevated immediately post‐exercise independent of the two treatments. These changes did not translate to an increase in p65 DNA binding activity. NF‐κB p50 protein expression and NF‐κB p50 binding activity were lower than pre‐exercise at 0 and 3 h post‐exercise, but were elevated at 24 h post‐exercise. These findings provide novel evidence that two distinct NF‐κB pathways are active in skeletal muscle after resistance exercise. The initial wave of activity involving p65 resembles the classical pathway and is associated with the onset of an acute inflammatory response. The second wave of NF‐κB activity comprises the p50 subunit, which has been previously shown to resolve an acute inflammatory program. The current study showed no effect of the ibuprofen treatment on markers of the NF‐κB pathway, however examination of the within group effects of the exercise protocol suggests that this pathway warrants further research.

Altering the redox state of skeletal muscle by glutathione depletion increases the exercise-activation of PGC-1α

We investigated the relationship between mitochondrial biogenesis, cell signalling and antioxidant enzymes by depleting skeletal muscle glutathione with diethyl maleate (DEM) which resulted in a demonstrable increase in oxidative stress during exercise. Animals were divided into six groups: (1) sedentary control rats; (2) sedentary rats treated with DEM; (3) exercise control rats euthanized immediately after exercise; (4) exercise rats + DEM; (5) exercise control rats euthanized 4 h after exercise, and; (6) exercise rats + DEM euthanized 4 h after exercise. Exercising animals ran on the treadmill at a 10% gradient at 20 m/min for the first 30 min. The speed was then increased every 10 min by 1.6 m/min until exhaustion. There was a reduction in total glutathione in the skeletal muscle of DEM treated animals compared to the control animals (P<0.05). Within the control group, total glutathione was higher in the sedentary group compared to after exercise (P<0.05). DEM treatment also significantly increased oxidative stress, as measured by increased plasma F2-isoprostanes (P<0.05). Exercising animals given DEM showed a significantly greater increase in peroxisome proliferator activated receptor γ coactivator-1α(PGC-1α) mRNA compared to the control animals that were exercised (P<0.05). This study provides novel evidence that by reducing the endogenous antioxidant glutathione in skeletal muscle and inducing oxidative stress through exercise, PGC-1α gene expression was augmented. These findings further highlight the important role of exercise induced oxidative stress in the regulation of mitochondrial biogenesis.

Ramipril - improving exercise tolerance in intermittent claudication?

Peripheral artery disease (PAD) is one of the most common manifestations of systemic atherosclerosis. It is estimated that 10-15% of the general population is affected by PAD, whereby the narrowed arteries lead to reduced blood flow to the extremeties - particularly the legs. While many people have mild or no systems with PAD, approximately one-third of people experience intermittent claudication (IC).

Physical and functional implications of aquatic exercise for nursing home residents with dementia

Exercise has reported benefits for those with dementia. In the current study we investigated the feasibility of delivery and the physical and functional benefits of an innovative aquatic exercise program for adults with moderate to severe dementia living in a nursing home aged care facility. Ten adults (88.4 years, inter quartile range 12.3) participated twice weekly for 12 weeks. Anthropometric and grip strength data, and measures of physical function and balance were collected at baseline and post-intervention. Feasibility was assessed by attendance, participation, enjoyment and recruitment. Following exercise, participant's left hand grip strength had improved significantly (p = .017). Small to moderate effect sizes were observed for other measures. A number of delivery challenges emerged, but participant enjoyment, benefits and attendance suggest feasibility. Aquatic exercise shows promise as an intervention among those with dementia who live in a nursing home aged care facility. Greater program investigation is warranted.

Effect of moderate- and high-intensity acute exercise on appetite in obese individuals

Purpose The effect of acute exercise, and exercise intensity, on appetite control in obese individuals requires further study. The aim of this study was to compare the effects of acute isocaloric bouts (250 kcal) of high-intensity intermittent cycling (HIIC) and moderate-intensity continuous cycling (MICC) or short-duration HIIC (S-HIIC) (125 kcal) and a resting control condition on the appetite hormone responses, subjective feelings of appetite, energy intake (EI), and food reward in overweight/obese individuals. Methods This study is a randomized crossover study on 12 overweight/obese volunteers. Participants were assigned to the control, MICC, HIIC, and S-HIIC conditions, 1 wk apart, in a counterbalanced order. Exercise was performed 1 h after a standard breakfast. An ad libitum test lunch was served 3 h after breakfast. Fasting/postprandial plasma samples of insulin, acylated ghrelin, polypeptide YY3–36, and glucagon-like peptide 1 and subjective feelings of appetite were measured every 30 min for 3 h. Nutrient and taste preferences were measured at the beginning and end of each condition using the Leeds Food Preference Questionnaire. Results Insulin levels were significantly reduced, and glucagon-like peptide 1 levels significantly increased during all exercise bouts compared with those during rest. Acylated ghrelin plasma levels were lower in the MICC and HIIC, but not in S-HIIC, compared with those in control. There were no significant differences for polypeptide YY3–36 plasma levels, hunger or fullness ratings, EI, or food reward. Conclusions Our findings suggest that, in overweight/obese individuals, isocaloric bouts of moderate- or high-intensity exercise lead to a similar appetite response. This strengthens previous findings in normal-weight individuals that acute exercise, even at high intensity, does not induce any known physiological adaptation that would lead to increased EI.

A comparison between conductive and infrared devices for measuring mean skin temperature at rest, during exercise in the heat, and recovery

Purpose: Skin temperature assessment has historically been undertaken with conductive devices affixed to the skin. With the development of technology, infrared devices are increasingly utilised in the measurement of skin temperature. Therefore, our purpose was to evaluate the agreement between four skin temperature devices at rest, during exercise in the heat, and recovery. Methods: Mean skin temperature (T̅sk) was assessed in thirty healthy males during 30 min rest (24.0± 1.2°C, 56 ± 8%), 30 min cycle in the heat (38.0 ± 0.5°C, 41 ± 2%), and 45 min recovery(24.0 ± 1.3°C, 56 ± 9%). T̅sk was assessed at four sites using two conductive devices(thermistors, iButtons) and two infrared devices (infrared thermometer, infrared camera). Results: Bland–Altman plots demonstrated mean bias ± limits of agreement between the thermistors and iButtons as follows (rest, exercise, recovery): -0.01 ± 0.04, 0.26 ± 0.85, -0.37 ± 0.98°C; thermistors and infrared thermometer: 0.34 ± 0.44, -0.44 ± 1.23, -1.04 ± 1.75°C; thermistors and infrared camera (rest, recovery): 0.83 ± 0.77, 1.88 ± 1.87°C. Pairwise comparisons of T̅sk found significant differences (p < 0.05) between thermistors and both infrared devices during resting conditions, and significant differences between the thermistors and all other devices tested during exercise in the heat and recovery. Conclusions: These results indicate poor agreement between conductive and infrared devices at rest, during exercise in the heat, and subsequent recovery. Infrared devices may not be suitable for monitoring T̅sk in the presence of, or following, metabolic and environmental induced heat stress.

Exercise during chemotherapy for ovarian cancer (Echo) : study design features and outcomes of a Cancer Australia and Cancer Council Australia funded randomised, controlled trial

Ovarian cancer is the most common cause of gynaecological cancer death, with an overall 5-year relative survival of 43%. Impaired physical wellbeing and overall quality of life (QoL) represent major concerns for women during and following ovarian cancer treatment, predict survival and are amenable to change through interventions. Exercise, now considered an important part of overall management of a number of cancers, improves short-term outcomes (e.g., function, fatigue, QoL) during chemotherapy...

Cytokine expression and secretion by skeletal muscle cells : regulatory mechanisms and exercise effects

Cytokines are important mediators of various aspects of health and disease, including appetite, glucose and lipid metabolism, insulin sensitivity, skeletal muscle hypertrophy and atrophy. Over the past decade or so, considerable attention has focused on the potential for regular exercise to counteract a range of disease states by modulating cytokine production. Exercise stimulates moderate to large increases in the circulating concentrations of interleukin (IL)-6, IL-8, IL-10, IL-1 receptor antagonist, granulocyte-colony stimulating factor, and smaller increases in tumor necrosis factor-α, monocyte chemotactic protein-1, IL-1β, brain-derived neurotrophic factor, IL-12p35/p40 and IL-15. Although many of these cytokines are also expressed in skeletal muscle, not all are released from skeletal muscle into the circulation during exercise. Conversely, some cytokines that are present in the circulation are not expressed in skeletal muscle after exercise. The reasons for these discrepant cytokine responses to exercise are unclear. In this review, we address these uncertainties by summarizing the capacity of skeletal muscle cells to produce cytokines, analyzing other potential cellular sources of circulating cytokines during exercise, and discussing the soluble factors and intracellular signaling pathways that regulate cytokine synthesis (e.g., RNA-binding proteins, microRNAs, suppressor of cytokine signaling proteins, soluble receptors).