Human Activity Recognition of Individuals with Lower Limb Amputation in Free-Living Conditions: A Pilot Study

This pilot study aimed to investigate the implementation of supervised classifiers and a neural network for the recognition of activities carried out by Individuals with Lower Limb Amputation (ILLAs), as well as individuals without gait impairment, in free living conditions. Eight individuals with no gait impairments and four ILLAs wore a thigh-based accelerometer and walked on an improvised route in the vicinity of their homes across a variety of terrains. Various machine learning classifiers were trained and tested for recognition of walking activities. Additional investigations were made regarding the detail of the activity label versus classifier accuracy and whether the classifiers were capable of being trained exclusively on non-impaired individuals’ data and could recognize physical activities carried out by ILLAs. At a basic level of label detail, Support Vector Machines (SVM) and Long-Short Term Memory (LSTM) networks were able to acquire 77–78% mean classification accuracy, which fell with increased label detail. Classifiers trained on individuals without gait impairment could not recognize activities carried out by ILLAs. This investigation presents the groundwork for a HAR system capable of recognizing a variety of walking activities, both for individuals with no gait impairments and ILLAs.

Substantial Intra-Individual Variability in Post-Prandial Time to Peak in Controlled and Free-Living Conditions in Children with Type 1 Diabetes

The optimal time to bolus insulin for meals is challenging for children and adolescents with type 1 diabetes (T1D). Current guidelines to control glucose excursions do not account for individual differences in glycaemic responses to meals. This study aimed to examine the within- and between-person variability in time to peak (TTP) glycaemic responses after consuming meals under controlled and free-living conditions. Participants aged 8–15 years with T1D ≥ 1 year and using a continuous glucose monitor (CGM) were recruited. Participants consumed a standardised breakfast for six controlled days and maintained their usual daily routine for 14 free-living days. CGM traces were collected after eating. Linear mixed models were used to identify within- and between-person variability in the TTP after each of the controlled breakfasts, free-living breakfasts (FLB), and free-living dinners (FLD) conditions. Thirty participants completed the study (16 females; mean age and standard deviation (SD) 10.5 (1.9)). The TTP variability was greater within a person than the variability between people for all three meal types (between-person vs within-person SD; controlled breakfast 18.5 vs 38.9 minutes; FLB 14.1 vs 49.6 minutes; FLD 5.7 vs 64.5 minutes). For the first time, the study showed that within-person variability in TTP glycaemic responses is even greater than between-person variability.

High-Resolution Digital Phenotypes from Consumer Wearables Enhance Prediction of Cardiometabolic Risk Markers

Background: Consumer-grade wearable devices enable detailed recordings of heart rate and step counts in free-living conditions. Recent studies have shown that summary statistics from these wearable recordings have potential uses for longitudinal monitoring of health and disease states. However, the relationship between higher resolution physiological dynamics from wearables and known markers of health and disease remains largely uncharacterized. Objective: We aimed to (i) derive high resolution digital phenotypes from observational wearable recordings, (ii) characterize their ability to predict modifiable markers of cardiometabolic disease, and (iii) study their connections with genetic predispositions for cardiometabolic disease and with lifestyle factors. Methods: We introduce a principled framework to extract interpretable high resolution phenotypes from wearable data recorded in free-living conditions. The proposed framework standardizes handling of data irregularities, encodes contextual information about underlying physiological state at any given time, and generates a set of 66 minimally redundant features across active, sedentary and sleep states. We applied our approach on a multimodal dataset, from the SingHEART study (NCT02791152), that comprises of heart rate and step count time series from wearables, clinical screening profiles, whole genome sequences and lifestyle survey responses from 692 healthy volunteers. We employed machine learning to model non-linear relationships between the high resolution phenotypes and clinical risk markers for blood pressure, lipid and weight abnormalities. For each risk type, we performed model comparisons based on Brier Skill Scores (BSS) to assess predictive value of the high resolution features over and beyond typical baselines. We then examined associations between the wearable-derived features, polygenic risk for cardiometabolic disease, and lifestyle habits and health perceptions. Results: Compared to typical summary statistic measures like resting heart rate, we find that the high-resolution features collectively have greater predictive value for modifiable clinical markers associated with cardiometabolic disease risk (average improvement in Brier Skill Score=52.3%, P<.001). Further, we show that heart rate dynamics from different activity states contain distinct information about type of cardiometabolic risk, with dynamics in sedentary states being most predictive of lipid abnormalities and patterns in active states being most predictive of blood pressure abnormalities (P<.001). Finally, our results reveal that subtle heart rate dynamics in wearable recordings serve as physiological correlates of genetic predisposition for cardiometabolic disease, lifestyle habits and health perceptions. Conclusions: High resolution digital phenotypes recorded by consumer wearables in free-living states have the potential to enhance prediction of cardiometabolic disease risk, and could enable more proactive and personalized health management. Clinical Trial Registration ID #NCT02791152. Keywords: Wearable device, heart rate, cardiometabolic disease, risk prediction, digital phenotypes, polygenic risk scores, time series analysis, machine learning, free-living

A glutamate synthase mutant of Bradyrhizobium sp. strain ORS285 is unable to induce nodules on Nod factor-independent Aeschynomene species

The Bradyrhizobium sp. strain ORS285 is able to establish a nitrogen-fixing symbiosis with both Nod factor (NF) dependent and NF-independent Aeschynomene species. Here, we have studied the growth characteristics and symbiotic interaction of a glutamate synthase (GOGAT; gltD::Tn5) mutant of Bradyrhizobium ORS285. We show that the ORS285 gltD::Tn5 mutant is unable to use ammonium, nitrate and many amino acids as nitrogen source for growth and is unable to fix nitrogen under free-living conditions. Moreover, on several nitrogen sources, the growth rate of the gltB::Tn5 mutant was faster and/or the production of the carotenoid spirilloxanthin was much higher as compared to the wild-type strain. The absence of GOGAT activity has a drastic impact on the symbiotic interaction with NF-independent Aeschynomene species. With these species, inoculation with the ORS285 gltD::Tn5 mutant does not result in the formation of nodules. In contrast, the ORS285 gltD::Tn5 mutant is capable to induce nodules on NF-dependent Aeschynomene species, but these nodules were ineffective for nitrogen fixation. Interestingly, in NF-dependent and NF-independent Aeschynomene species inoculation with the ORS285 gltD::Tn5 mutant results in browning of the plant tissue at the site of the infection suggesting that the mutant bacteria induce plant defence responses.

Relationship Between Fasting and Postprandial Glucose Levels and the Gut Microbiota

Postprandial hyperglycemia increases the risk of mortality among patients with type 2 diabetes and cardiovascular diseases. Additionally, the gut microbiota and type 2 diabetes and cardiovascular disease are known to be correlated. Currently, fasting blood glucose is the primary index for the clinical diagnosis of diabetes; however, postprandial blood glucose is associated with the risk of developing type 2 diabetes and cardiovascular disease and mortality. Therefore, the dynamic change in blood glucose levels under free-living conditions is considered an important and better marker than fasting glucose levels, to study the relationship between glucose levels and microbiota. Here, we investigated the relationship between fasting and postprandial glucose levels and microbiota under free-living conditions, for one week in the older adults. The results revealed a significant correlation between peak glucose levels after dinner and the gut bacteria, particularly, Bacteroides, Clostridiales Clostridiaceae group, Anaerostipes, Clostridiales [Mogibacteriaceae] group, Holdemania, and Bilophila. Together, these findings suggest that the glucose levels after dinner are a better predictor of microbiota conditions than fasting glucose levels.

Validity, Reliability and Sensitivity to Change of Three Consumer-Grade Activity Trackers in Controlled and Free-Living Conditions among Older Adults

Wrist-worn consumer-grade activity trackers are popular devices, developed mainly for personal use. This study aimed to explore the validity, reliability and sensitivity to change of movement behaviors metrics from three activity trackers (Polar Vantage M, Garmin Vivoactive 4s and Garmin Vivosport) in controlled and free-living conditions when worn by older adults. Participants (n = 28; 74 ± 5 years) underwent a videotaped laboratory protocol while wearing all three trackers. On a separate occasion, participants (n = 17 for each of the trackers) wore one (randomly assigned) tracker and a research-grade activity monitor ActiGraph wGT3X-BT simultaneously for six consecutive days. Both Garmin trackers showed excellent performance for step counts, with a mean absolute percentage error (MAPE) below 20% and intraclass correlation coefficient (ICC2,1) above 0.90 (p < 0.05). The MAPE for sleep time was within 10% for all the trackers tested, while it was far beyond 20% for all other movement behaviors metrics. The results suggested that all three trackers could be used for measuring sleep time with a high level of accuracy, and both Garmin trackers could also be used for step counts. All other output metrics should be used with caution. The results provided in this study could be used to guide choice on activity trackers aiming for different purposes—individual use, longitudinal monitoring or in clinical trial setting.

Your Physical Activity Is in Your Hand—Objective Activity Tracking Among University Students in Hungary, One of the Most Obese Countries in Europe

Inadequate physical activity is currently one of the leading risk factors for mortality worldwide. University students are a high-risk group in terms of rates of obesity and lack of physical activity. In recent years, activity trackers have become increasingly popular for measuring physical activity. The aim of the present study is to examine whether university students in Hungary meet the health recommendations (10,000 steps/day) for physical activity and investigate the impact of different variables (semester-exam period, days-weekdays, days, months, sex) on the level of physical activity in free-living conditions for 3 months period. In free-living conditions, 57 healthy university students (male: 25 female: 32 mean age: 19.50 SD = 1.58) wore MiBand 1S activity tracker for 3 months. Independent sample t-tests were used to explore differences between sexes. A One-way analysis of variance (ANOVA) was used to explore differences in measures among different grouping variables and step count. A Two-way ANOVA was conducted to test for differences in the number of steps by days of the week, months, seasons and for sex differences. Tukey HSD post-hoc tests were used to examine significant differences. Students in the study achieved 10,000 steps per day on 17% of days (minimum: 0%; maximum: 76.5%; median: 11.1%). Unfortunately, 70% of the participants did not comply the 10,000 steps at least 80% of the days studied. No statistical difference were found between sexes. However, significant differences were found between BMI categories (underweight &lt;18.50 kg/m2; normal range 18.50–24.99 kg/m2; overweight: 25.00–29.99 kg/m2 obese &gt; 30 kg/m2, the number of steps in the overweight category was significantly lower (F = 72.073, p &lt; 0.001). The average daily steps were significantly higher in autumn (t = 11.457, p &lt; 0.001) than in winter. During exam period average steps/day were significantly lower than during fall semester (t = 13.696, p &lt; 0.001). On weekdays, steps were significantly higher than on weekends (F = 14.017, p &lt; 0.001), and even within this, the greatest physical activity can be done by the middle of the week. Our data suggest that university students may be priority groups for future physical activity interventions. Commercial activity trackers provide huge amount of data for relatively low cost therefore it has the potential to objectively analyze physical activity and plan interventions.

Comparison of Energy Expenditure Assessed Using Wrist- and Hip-Worn ActiGraph GT3X in Free-Living Conditions in Young and Older Adults

The World Health Organization has presented their recommendations for energy expenditure to improve public health. Activity trackers do represent a modern solution for measuring physical activity, particularly in terms of steps/day and energy expended in physical activity (active energy expenditure). According to the manufacturer's instructions, these activity trackers can be placed on different body locations, mostly at the wrist and the hip, in an undifferentiated manner. The objective of this study was to compare the absolute error rate of active energy expenditure measured by a wrist-worn and hip-worn ActiGraph GT3X+ over a 24-h period in free-living conditions in young and older adults. Over the period of a 24-h period, 22 young adults and 22 older adults were asked to wear two ActiGraph GT3X+ at two different body locations recommended by the manufacturer, namely one around the wrist and one above the hip. Freedson algorithm was applied for data analysis. For both groups, the absolute error rate tended to decrease from 1,252 to 43% for older adults and from 408 to 46% for young participants with higher energy expenditure. Interestingly, for both young and older adults, the wrist-worn ActiGraph provided a significantly higher values of active energy expenditure (943 ± 264 cal/min) than the hip-worn (288 ± 181 cal/min). Taken together, these results suggest that caution is needed when using active energy expenditure as an activity tracker-based metric to quantify physical activity.

Hill Runner's Physiology, Performance and Nutrition: A Descriptive Study

Objectives: The aim of this descriptive study was to characterise anthropometric variables, aerobic capacity, running performance and energy intake and expenditure of hill runners in free-living conditions, and to investigate the relationship between age, anthropometric variables, aerobic capacity and running performance.Methods: Twenty-eight hill runners participated in this study (17 males and 11 females; aged 18–65 years). Body fat percentage estimate, sum of eight skinfolds (triceps, subscapular, biceps, iliac crest, supraspinale, abdominal, front thigh and medial calf) and maximal oxygen capacity (VO2max) were assessed in a laboratory setting. Participants also completed a timed hill run (Dumyat Hill, Scotland, ascent: 420 m, distance: 8 km) while wearing a portable gas analyzer to assess oxygen consumption (VO2). Energy intake and energy expenditure were assessed in free-living conditions over three consecutive days different from the testing days through self-reported food diaries and accelerometers.Results: VO2max assessed in the lab (51.2 ± 7.6 ml·min−1·kg−1) showed a weak negative relationship with age [rs(23) = −0.38, p = 0.08]. Neither body fat percentage (median 12.4; IQR 10.1–17.1) nor the sum of skinfolds (median 81.8; IQR 62.4–97.8 mm) correlated with age [rs(28) = 0.001, p = 0.10 and 26 rs(28) = −0.02, p = 0.94, respectively]. The observed intensity of the hill run was 89 ± 6% of the age predicted maximum heart rate and 87 ± 9% of the VO2max observed in the lab. Hill running performance correlated with VO2max [r(21) = 0.76, p &lt; 0.001], age [rs(26) = −0.44, p = 0.02] and with estimated body fat percentage and sum of skinfolds [rs(26) = −0.66, p &lt; 0.001 and rs(26) = −0.49, p = 0.01, respectively]. Energy intake negatively correlated with age [rs(26) = −0.43, p = 0.03], with the overall energy intake being significantly lower than the total energy expenditure (2273 ± 550 vs. 2879 ± 510 kcal·day−1; p &lt; 0.001; d = 1.05).Conclusion: This study demonstrated that hill running performance is positively associated with greater aerobic capacity and negatively associated with increases in adiposity and age. Further, the study highlights that hill runners are at risk of negative energy balance.