HARTH: A Human Activity Recognition Dataset for Machine Learning

Movement Behaviors in Youth on the Autism Spectrum: The HUNT Study, Norway

Research consistently show that autistic youth are less physically active compared to their neurotypical peers. However, there is limited understanding of how gender influences physical activity (PA) patterns among neurodiverse youth compared to the general population. This study aims to examine 24-hour movement behaviors - PA, sedentary behavior (SB), and sleep duration - among autistic youth (n = 71) in Norway, in comparison to peers with Attention-Deficit/Hyperactivity Disorder (ADHD) (n = 411) and the general youth population (n = 3805). The data is from the Young-HUNT4 study, linked with diagnostic information from the Norwegian Patient Registry. Variables explored are objective accelerometer-measured PA, SB, and sleep duration, self-reported participation in organized and unorganized physical activities, and screen activities. Results confirms that autistic youth engage in lower levels of moderate-to-vigorous PA, while demonstrating similar levels of light PA. They also spend more time sitting and comparable time sleeping. Autistic youth participate less in sport and were less likely to use commercial gyms. However, their participation in outdoor activities were similar to their peers. Regarding screen activities, autistic youth spent more time playing video games, while youth with ADHD were more engaged in social media. Among autistic youth, the only gender difference found was in video games. In conclusion, autistic youth are less physically active overall and spend significant time in SB. However, their comparable participation in light PA suggests opportunities for promoting further participation. Additionally, exergaming could offer a promising avenue to increase PA in this population.

Associations of sport and exercise participation in adolescence with body composition and device-measured physical activity in adulthood: longitudinal data from the Norwegian HUNT study

BACKGROUND

To examine whether adolescent sport and exercise participation is associated with adulthood moderate-to-vigorous physical activity (MVPA), body fat, skeletal muscle mass, and body mass index (BMI), and to explore whether the association between sport and exercise participation and adult body composition depends on adulthood MVPA level.

METHODS

Prospective study of 4603 adolescents aged 13-18 year (57.2% female) in the Norwegian Young-HUNT Study and follow-up ~ 11 or ~ 22 years later. Linear regression was used to estimate mean differences in accelerometer-measured MVPA and bioimpedance-measured body fat, muscle mass, and BMI in adulthood according to self-reported sport and exercise participation in adolescence.

RESULTS

Adolescents participating in sport/exercise every day accumulated more MVPA (48 min/week, 95% CI 23 to 73), had less body fat (-4.4%, 95% CI -5.4 to -3.2), more muscle mass (2.6%, 95% CI 2.0 to 3.2), and lower BMI (-1.1 kg/m2, 95% CI -1.7 to -0.5) as adults, compared to adolescents participating < 1 day/week. Joint analysis showed that adolescents who participated in sport/exercise ≥ 4 days/week, and who accumulated 150-299 min/week MVPA in adulthood, had less body fat (-5.8%, 95% CI -7.4 to -4.3) and more muscle mass (3.4%, 95% CI, 2.5 to 4.3) compared to those participating in sport/exercise ≤ 1 day/week and who accumulated < 150 MVPA min/week as adults. Compared to the same reference group, these associations were further strengthened among those who accumulated ≥ 300 min/week MVPA in adulthood and reported ≥ 4 days/week of sport/exercise for both body fat (-8.8%, 95% CI -10.3 to -7.4) and muscle mass (5.1%, 95% CI 4.3 to 5.9).

CONCLUSIONS

Adolescent sport and exercise participation is positively associated with MVPA, and skeletal muscle mass, and inversely associated with body fat and BMI in adulthood. These associations remained significant after adjusting for adult MVPA levels. A higher MVPA level in adulthood strengthens the association between adolescent sport/exercise participation and adult body composition.

Federated Learning for IoMT-Enhanced Human Activity Recognition with Hybrid LSTM-GRU Networks

The proliferation of wearable sensors and mobile devices has fueled advancements in human activity recognition (HAR), with growing importance placed on both accuracy and privacy preservation. In this paper, the author proposes a federated learning framework for HAR, leveraging a hybrid Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) model to enhance feature extraction and classification in decentralized environments. Utilizing three public datasets-UCI-HAR, HARTH, and HAR7+-which contain diverse sensor data collected from free-living activities, the proposed system is designed to address the inherent privacy risks associated with centralized data processing by deploying Federated Averaging for local model training. To optimize recognition accuracy, the author introduces a dual-feature extraction mechanism, combining convolutional blocks for capturing local patterns and a hybrid LSTM-GRU structure to detect complex temporal dependencies. Furthermore, the author integrates an attention mechanism to focus on significant global relationships within the data. The proposed system is evaluated on the three public datasets-UCI-HAR, HARTH, and HAR7+-achieving superior performance compared to recent works in terms of F1-score and recognition accuracy. The results demonstrate that the proposed approach not only provides high classification accuracy but also ensures privacy preservation, making it a scalable and reliable solution for real-world HAR applications in decentralized and privacy-conscious environments. This work showcases the potential of federated learning in transforming human activity recognition, combining advanced feature extraction methodologies and privacy-respecting frameworks to deliver robust, real-time activity classification.

The performance of a machine learning model in predicting accelerometer-derived walking speed

Background

Obtaining long-term measurements of walking speed in large-scale studies remains challenging. The aim of this study was to develop and evaluate the performance of a machine learning classifier in predicting slow (≤4 km/h), moderate (4.1-5.4 km/h), and brisk (≥5.5 km/h) walking speeds in adults based on dual and single accelerometer set-ups.

Methods

Twenty-four adults (mean age [SD, range] 36.1 [11.9, 23-62] years) participated in the study. Two tri-axial accelerometers positioned on the thigh and low back were used to record body movements. A measuring wheel with a speedometer along with video recording were used to define and record consecutive 5-min periods with the three walking speeds and jogging during conditions resembling free-living. In addition, we included a 5-min period with gradual increase and decrease in walking speed from slow to brisk and vice versa. The video recordings were labelled and used as ground truth for training an eXtreme Gradient Boosting (XGBoost) machine learning classifier. Windows of 1, 3, and 5 s duration were used to train the classifier. The performance of the classifier was evaluated by leave-one-out cross-validation.

Results

Total recording time was ∼600 min (∼25 min per participant). Performance metrics for predicting walking speeds (i.e., slow, moderate, brisk) and jogging were largely similar for the dual and single accelerometer set-ups as well as for the different window lengths. The highest overall accuracy was 91 % (SD 11 %, range 59-98 % for individual participants) using a dual accelerometer set-up and a 5-s window, whereas the lowest overall accuracy was 88 % (SD 11 %, range 51-96 % for individual participants) using a single thigh accelerometer set-up and a 1-s window.

Conclusions

A machine learning classifier can be used to accurately predict slow, moderate, and brisk walking speeds based on both a dual and single accelerometer set-up on the thigh and/or low back.

Device-measured physical activity in adults born preterm with very low birth weight and mediation by motor abilities

Physical activity (PA) is beneficial for several health outcomes. Adults born with very low birth weight (VLBW<1500g) undertake less PA than those born at term, have poorer motor abilities and may serve as a model on early life origins of PA. We therefore examined whether motor abilities mediate the association between being born with VLBW and device-measured PA. In a joint assessment of two longitudinal birth cohorts from Finland and Norway, PA was measured by two tri-axial accelerometers in 87 adults born preterm with VLBW and 109 controls born at term. We explored the mediating role of motor abilities assessed by standardized tests on the association between VLBW and device-measured PA. To do this, we examined group differences in metabolic equivalent of task (MET) min/day of moderate to vigorous PA (MVPA), light PA and sedentary. Analyses were adjusted for cohort, age and sex. MVPA was 40.4 (95% confidence interval [CI]: 13.3 to 69.4) MET min/day lower in the VLBW group than the control group. This was in part mediated through gross motor abilities, indicated by the indirect effect on the association between VLBW and MVPA being -15.6 (95% CI: -28.5 to -5.4) MET min/day. In conclusion, adults born preterm with VLBW undertake less MVPA than controls born at term, and gross motor abilities mediate this association. Interventions targeting motor abilities should be examined as potential ways to increase PA.

Revolutionizing health monitoring: Integrating transformer models with multi-head attention for precise human activity recognition using wearable devices

BACKGROUND

A daily activity routine is vital for overall health and well-being, supporting physical and mental fitness. Consistent physical activity is linked to a multitude of benefits for the body, mind, and emotions, playing a key role in raising a healthy lifestyle. The use of wearable devices has become essential in the realm of health and fitness, facilitating the monitoring of daily activities. While convolutional neural networks (CNN) have proven effective, challenges remain in quickly adapting to a variety of activities.

OBJECTIVE

This study aimed to develop a model for precise recognition of human activities to revolutionize health monitoring by integrating transformer models with multi-head attention for precise human activity recognition using wearable devices.

METHODS

The Human Activity Recognition (HAR) algorithm uses deep learning to classify human activities using spectrogram data. It uses a pretrained convolution neural network (CNN) with a MobileNetV2 model to extract features, a dense residual transformer network (DRTN), and a multi-head multi-level attention architecture (MH-MLA) to capture time-related patterns. The model then blends information from both layers through an adaptive attention mechanism and uses a SoftMax function to provide classification probabilities for various human activities.

RESULTS

The integrated approach, combining pretrained CNN with transformer models to create a thorough and effective system for recognizing human activities from spectrogram data, outperformed these methods in various datasets - HARTH, KU-HAR, and HuGaDB produced accuracies of 92.81%, 97.98%, and 95.32%, respectively. This suggests that the integration of diverse methodologies yields good results in capturing nuanced human activities across different activities. The comparison analysis showed that the integrated system consistently performs better for dynamic human activity recognition datasets.

CONCLUSION

In conclusion, maintaining a routine of daily activities is crucial for overall health and well-being. Regular physical activity contributes substantially to a healthy lifestyle, benefiting both the body and the mind. The integration of wearable devices has simplified the monitoring of daily routines. This research introduces an innovative approach to human activity recognition, combining the CNN model with a dense residual transformer network (DRTN) with multi-head multi-level attention (MH-MLA) within the transformer architecture to enhance its capability.

Assessing the level of device-measured physical activity according to insomnia symptoms in 1,354 individuals with diabetes: the HUNT Study, Norway

Background

Insomnia symptoms that influence daytime functioning are common among adults with type 2 diabetes. However, no previous study has examined if levels of physical activity differ among adults with diabetes with and without insomnia symptoms. Thus, the aim of this study was to assess the difference in total physical activity (TPA) and moderate-to-vigorous physical activity (MVPA) levels in individuals with diabetes with and without insomnia symptoms.

Methods

This cross-sectional study included 1,354 participants with any type of diabetes who participated in the Norwegian HUNT4 Study, 2017-19. Participants were defined to have 'insomnia symptoms' if they reported difficulty initiating and/or maintaining sleep ≥ 3 nights/week during the last 3 months. MVPA (defined as moderate/brisk walking [> 4.0 km/h], running, and cycling), and TPA (MVPA including slow walking [≤ 4.0 km/h]) were determined from two accelerometers worn on the thigh and lower back. Analyses were stratified by age and sex.

Results

The median age was 67 years and 491 (36%) had insomnia symptoms and 37 (3%) had insomnia disorder. Among women, 28% with one or more insomnia symptoms fulfilled the recommended minimum level of physical activity, as compared to 34% in women without insomnia symptoms. The corresponding proportions in men were 48% and 45%. Women above 65 years with insomnia symptoms performed less TPA (-73 min/week, 95% CI -122 to -24) and MVPA (-33 min/week, 95% CI -50 to -15), compared to women without insomnia symptoms in the same age group. There was no clear difference in physical activity levels according to insomnia symptoms in men or women below 65 years. Women and men with insomnia disorder had substantially lower TPA (women: -192 min/week, 95% CI -278 to -106; men: -276 min/week, 95% CI -369 to -193) and MVPA (women: -37 min/week, 95% CI -63 to -11; men: -67 min/week, 95% CI -83 to -50) than those without insomnia symptoms.

Conclusions

This study showed that women above 65 years with insomnia symptoms and individuals with insomnia disorder performed less physical activity, suggesting that these subgroups may suffer from additional challenges that prevent them from engaging in regular physical activity.

Supplementary Information

The online version contains supplementary material available at 10.1186/s44167-024-00066-4.

Validation of two novel human activity recognition models for typically developing children and children with Cerebral Palsy

Human Activity Recognition models have potential to contribute to valuable and detailed knowledge of habitual physical activity for typically developing children and children with Cerebral Palsy. The main objective of the present study was to develop and validate two Human Activity Recognition models. One trained on data from typically developing children (n = 63), the second also including data from children with Cerebral Palsy (n = 16), engaging in standardised activities and free play. Our data was collected using accelerometers and ground truth was established with video annotations. Additionally, we aimed to investigate the influence of window settings on model performance. Utilizing the Extreme gradient boost (XGBoost) classifier, twelve sub-models were created, with 1-,3- and 5-seconds windows, with and without overlap. Both Human Activity Recognition models demonstrated excellent predictive capabilities (>92%) for standardised activities for both typically developing and Cerebral Palsy. From all window sizes, the 1-second window performed best for all test groups. Accuracy was slightly lower (>75%) for the Cerebral Palsy test group performing free play activities. The impact of window size and overlap varied depending on activity. In summary both Human Activity Recognition models effectively predict standardised activities, surpassing prior models for typically developing and children with Cerebral Palsy. Notably, the model trained on combined typically developing children and Cerebral Palsy data performed exemplary across all test groups. Researchers should select window settings aligned with their specific research objectives.

Early Intervention after Rape to prevent post-traumatic stress symptoms (the EIR-study): an internal pilot study of a randomized controlled trial

BACKGROUND

Rape is one of the trauma incidents with the highest risk of subsequent post-traumatic stress disorder. Early interventions, such as prolonged exposure therapy (PE), have shown promise in preventing PTSD following a sexual assault. The primary objective of this internal pilot trial was to examine the feasibility of the EIR study protocol, which used modified prolonged exposure therapy (mPE) as a preventive intervention after rape.

METHODS

This parallel two-arm clinical pilot study involved three sexual assault centers (SACs) in Trondheim, Oslo, and Vestfold, with data collected between June 2022 and March 2023. Women seeking assistance at one of these three SACs within 72 h after rape or attempted rape received acute medical treatment and forensic examinations. Women who wanted further psychosocial treatment were, if eligible and consenting, recruited to complete baseline assessments and a clinical interview before being randomized to one of two study arms. The intervention group prescribed up to five sessions of modified PE (mPE) in addition to treatment as usual (TAU), starting within the first 14 days after the rape incident, followed by weekly sessions. The other group received TAU. The present pilot evaluation is based on 22 participants, i.e., nine mPE + TAU and 13 TAU alone. Primary outcomes were predefined progression criteria regarding recruitment, retention, intervention implementation, a harm reporting system, and applying biological measurements and actigraphy.

RESULTS

During the 6-month recruitment period, 235 women visited the three SACs. After eligibility screening and consent, 22 (9.4%) women were randomized. Three months later, 14 (63.6%) participants completed the final assessments. Intervention implementation was successful using trained SAC personnel to deliver mPE. The harm reporting system was used according to the study's plan, and adverse and serious adverse events were detected during the trial. The biological measurements and actigraphy had substantial missing data but were still considered usable for statistical analyses.

CONCLUSION

It may be feasible to conduct a full-scale RCT of early intervention after rape by comparing mPE + TAU to TAU alone. Minor design refinements were made to the protocol to enhance the main study outcome.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT05489133. Registered on 15 July 2022, retrospectively.

A Machine Learning Model for Predicting Sleep and Wakefulness Based on Accelerometry, Skin Temperature and Contextual Information

Purpose

Body-worn accelerometers are commonly used to estimate sleep duration in population-based studies. However, since accelerometry-based sleep/wake-scoring relies on detecting body movements, the prediction of sleep duration remains a challenge. The aim was to develop and evaluate the performance of a machine learning (ML) model to predict accelerometry-based sleep duration and to explore if this prediction can be improved by adding skin temperature data, circadian rhythm based on the estimated midpoint of sleep, and cyclic time features to the model.

Patients and Methods

Twenty-nine adults (17 females), mean (SD) age 40.2 (15.0) years (range 17-70) participated in the study. Overnight polysomnography (PSG) was recorded in a sleep laboratory or at home along with body movement by two accelerometers with an embedded skin temperature sensor (AX3, Axivity, UK) positioned at the low back and thigh. The PSG scoring of sleep/wake was used as ground truth for training the ML model.

Results

Based on pure accelerometer data input to the ML model, the specificity and sensitivity for predicting sleep/wake was 0.52 (SD 0.24) and 0.95 (SD 0.03), respectively. Adding skin temperature data and contextual information to the ML model improved the specificity to 0.72 (SD 0.20), while sensitivity remained unchanged at 0.95 (SD 0.05). Correspondingly, sleep overestimation was reduced from 54 min (228 min, limits of agreement range [LoAR]) to 19 min (154 min LoAR).

Conclusion

An ML model can predict sleep/wake periods with excellent sensitivity and moderate specificity based on a dual-accelerometer set-up when adding skin temperature data and contextual information to the model.

Human Activity Recognition Algorithm with Physiological and Inertial Signals Fusion: Photoplethysmography, Electrodermal Activity, and Accelerometry

Inertial signals are the most widely used signals in human activity recognition (HAR) applications, and extensive research has been performed on developing HAR classifiers using accelerometer and gyroscope data. This study aimed to investigate the potential enhancement of HAR models through the fusion of biological signals with inertial signals. The classification of eight common low-, medium-, and high-intensity activities was assessed using machine learning (ML) algorithms, trained on accelerometer (ACC), blood volume pulse (BVP), and electrodermal activity (EDA) data obtained from a wrist-worn sensor. Two types of ML algorithms were employed: a random forest (RF) trained on features; and a pre-trained deep learning (DL) network (ResNet-18) trained on spectrogram images. Evaluation was conducted on both individual activities and more generalized activity groups, based on similar intensity. Results indicated that RF classifiers outperformed corresponding DL classifiers at both individual and grouped levels. However, the fusion of EDA and BVP signals with ACC data improved DL classifier performance compared to a baseline DL model with ACC-only data. The best performance was achieved by a classifier trained on a combination of ACC, EDA, and BVP images, yielding F1-scores of 69 and 87 for individual and grouped activity classifications, respectively. For DL models trained with additional biological signals, almost all individual activity classifications showed improvement (p-value < 0.05). In grouped activity classifications, DL model performance was enhanced for low- and medium-intensity activities. Exploring the classification of two specific activities, ascending/descending stairs and cycling, revealed significantly improved results using a DL model trained on combined ACC, BVP, and EDA spectrogram images (p-value < 0.05).

Quantitative Analysis of Mother Wavelet Function Selection for Wearable Sensors-Based Human Activity Recognition

Recent advancements in the Internet of Things (IoT) wearable devices such as wearable inertial sensors have increased the demand for precise human activity recognition (HAR) with minimal computational resources. The wavelet transform, which offers excellent time-frequency localization characteristics, is well suited for HAR recognition systems. Selecting a mother wavelet function in wavelet analysis is critical, as optimal selection improves the recognition performance. The activity time signals data have different periodic patterns that can discriminate activities from each other. Therefore, selecting a mother wavelet function that closely resembles the shape of the recognized activity's sensor (inertial) signals significantly impacts recognition performance. This study uses an optimal mother wavelet selection method that combines wavelet packet transform with the energy-to-Shannon-entropy ratio and two classification algorithms: decision tree (DT) and support vector machines (SVM). We examined six different mother wavelet families with different numbers of vanishing points. Our experiments were performed on eight publicly available ADL datasets: MHEALTH, WISDM Activity Prediction, HARTH, HARsense, DaLiAc, PAMAP2, REALDISP, and HAR70+. The analysis demonstrated in this paper can be used as a guideline for optimal mother wavelet selection for human activity recognition.

Can the bias of self-reported sitting time be corrected? A statistical model validation study based on data from 23 993 adults in the Norwegian HUNT study

BACKGROUND

Despite apparent shortcomings such as measurement error and low precision, self-reported sedentary time is still widely used in surveillance and research. The aim of this study was threefold; (i) to examine the agreement between self-reported and device-measured sitting time in a general adult population; (ii), to examine to what extent demographics, lifestyle factors, long-term health conditions, physical work demands, and educational level is associated with measurement bias; and (iii), to explore whether correcting for factors associated with bias improves the prediction of device-measured sitting time based on self-reported sitting time.

METHODS

A statistical validation model study based on data from 23 993 adults in the Trøndelag Health Study (HUNT4), Norway. Participants reported usual sitting time on weekdays using a single-item questionnaire and wore two AX3 tri-axial accelerometers on the thigh and low back for an average of 3.8 (standard deviation [SD] 0.7, range 1-5) weekdays to determine their sitting time. Statistical validation was performed by iteratively adding all possible combinations of factors associated with bias between self-reported and device-measured sitting time in a multivariate linear regression. We randomly selected 2/3 of the data (n = 15 995) for model development and used the remaining 1/3 (n = 7 998) to evaluate the model.

RESULTS

Mean (SD) self-reported and device-measured sitting time were 6.8 (2.9) h/day and 8.6 (2.2) h/day, respectively, corresponding to a mean difference of 1.8 (3.1) h/day. Limits of agreement ranged from - 8.0 h/day to 4.4 h/day. The discrepancy between the measurements was characterized by a proportional bias with participants device-measured to sit less overestimating their sitting time and participants device-measured to sit more underestimating their sitting time. The crude explained variance of device-measured sitting time based on self-reported sitting time was 10%. This improved to 24% when adding age, body mass index and physical work demands to the model. Adding sex, lifestyle factors, educational level, and long-term health conditions to the model did not improve the explained variance.

CONCLUSIONS

Self-reported sitting time had low validity and including a range of factors associated with bias in self-reported sitting time only marginally improved the prediction of device-measured sitting time.

Human behavior recognition based on sparse transformer with channel attention mechanism

Human activity recognition (HAR) has recently become a popular research field in the wearable sensor technology scene. By analyzing the human behavior data, some disease risks or potential health issues can be detected, and patients' rehabilitation progress can be evaluated. With the excellent performance of Transformer in natural language processing and visual tasks, researchers have begun to focus on its application in time series. The Transformer model models long-term dependencies between sequences through self-attention mechanisms, capturing contextual information over extended periods. In this paper, we propose a hybrid model based on the channel attention mechanism and Transformer model to improve the feature representation ability of sensor-based HAR tasks. Extensive experiments were conducted on three public HAR datasets, and the results show that our network achieved accuracies of 98.10%, 97.21%, and 98.82% on the HARTH, PAMAP2, and UCI-HAR datasets, respectively, The overall performance is at the level of the most advanced methods.

SONAR, a nursing activity dataset with inertial sensors

Accurate and comprehensive nursing documentation is essential to ensure quality patient care. To streamline this process, we present SONAR, a publicly available dataset of nursing activities recorded using inertial sensors in a nursing home. The dataset includes 14 sensor streams, such as acceleration and angular velocity, and 23 activities recorded by 14 caregivers using five sensors for 61.7 hours. The caregivers wore the sensors as they performed their daily tasks, allowing for continuous monitoring of their activities. We additionally provide machine learning models that recognize the nursing activities given the sensor data. In particular, we present benchmarks for three deep learning model architectures and evaluate their performance using different metrics and sensor locations. Our dataset, which can be used for research on sensor-based human activity recognition in real-world settings, has the potential to improve nursing care by providing valuable insights that can identify areas for improvement, facilitate accurate documentation, and tailor care to specific patient conditions.

Physical activity and associations with health-related quality of life in adults born small for gestational age at term: a prospective cohort study

BACKGROUND

Adults born small for gestational age (SGA) have increased risk of adverse health outcomes. Physical activity (PA) is a key determinant of health and health-related quality of life (HRQoL). We aimed to investigate if being born SGA at term is associated with lower objectively measured and self-reported PA during adulthood. We also examined if objectively measured and self-reported PA were associated with HRQoL.

METHODS

As part of the 32-year follow-up in the NTNU Low Birth Weight in a Lifetime Perspective study, SGA and non-SGA control participants wore two tri-axial accelerometers for seven days (37 SGA, 43 control), and completed the International Physical Activity Questionnaire (IPAQ) (42 SGA, 49 control) and the Short Form 36 Health Survey (SF-36) (55 SGA, 67 control). Group differences in objectively measured daily metabolic equivalent of task (MET) minutes spent sedentary (lying, sitting), on feet (standing, walking, running, cycling), on the move (walking, running, cycling) and running/cycling, and group differences in self-reported daily MET minutes spent walking and in moderate and vigorous PA were examined using linear regression. Associations with SF-36 were explored in a general linear model.

RESULTS

Mean (SD) daily MET minutes on the move were 218 (127) in the SGA group and 227 (113) in the control group. There were no group differences in objectively measured and self-reported PA or associations with HRQoL. In the SGA group, one MET minute higher objectively measured time on the move was associated with 4.0 (95% CI: 0.6-6.5, p = 0.009) points higher SF-36 physical component summary.

CONCLUSION

We found no differences in objectively measured and self-reported PA or associations with HRQoL between term-born SGA and non-SGA control participants in adulthood.

Exploring Regularization Methods for Domain Generalization in Accelerometer-Based Human Activity Recognition

The study of Domain Generalization (DG) has gained considerable momentum in the Machine Learning (ML) field. Human Activity Recognition (HAR) inherently encompasses diverse domains (e.g., users, devices, or datasets), rendering it an ideal testbed for exploring Domain Generalization. Building upon recent work, this paper investigates the application of regularization methods to bridge the generalization gap between traditional models based on handcrafted features and deep neural networks. We apply various regularizers, including sparse training, Mixup, Distributionally Robust Optimization (DRO), and Sharpness-Aware Minimization (SAM), to deep learning models and assess their performance in Out-of-Distribution (OOD) settings across multiple domains using homogenized public datasets. Our results show that Mixup and SAM are the best-performing regularizers. However, they are unable to match the performance of models based on handcrafted features. This suggests that while regularization techniques can improve OOD robustness to some extent, handcrafted features remain superior for domain generalization in HAR tasks.

Reducing the Impact of Sensor Orientation Variability in Human Activity Recognition Using a Consistent Reference System

Sensor- orientation is a critical aspect in a Human Activity Recognition (HAR) system based on tri-axial signals (such as accelerations); different sensors orientations introduce important errors in the activity recognition process. This paper proposes a new preprocessing module to reduce the negative impact of sensor-orientation variability in HAR. Firstly, this module estimates a consistent reference system; then, the tri-axial signals recorded from sensors with different orientations are transformed into this consistent reference system. This new preprocessing has been evaluated to mitigate the effect of different sensor orientations on the classification accuracy in several state-of-the-art HAR systems. The experiments were carried out using a subject-wise cross-validation methodology over six different datasets, including movements and postures. This new preprocessing module provided robust HAR performance even when sudden sensor orientation changes were included during data collection in the six different datasets. As an example, for the WISDM dataset, sensors with different orientations provoked a significant reduction in the classification accuracy of the state-of-the-art system (from 91.57 ± 0.23% to 89.19 ± 0.26%). This important reduction was recovered with the proposed algorithm, increasing the accuracy to 91.46 ± 0.30%, i.e., the same result obtained when all sensors had the same orientation.

AccNet24: A deep learning framework for classifying 24-hour activity behaviours from wrist-worn accelerometer data under free-living environments

OBJECTIVE

Although machine learning techniques have been repeatedly used for activity prediction from wearable devices, accurate classification of 24-hour activity behaviour categories from accelerometry data remains a challenge. We developed and validated a deep learning-based framework for classifying 24-hour activity behaviours from wrist-worn accelerometers.

METHODS

Using an openly available dataset with free-living wrist-based raw accelerometry data from 151 participants (aged 18-91 years), we developed a deep learning framework named AccNet24 to classify 24-hour activity behaviours. First, the acceleration signal (x, y, and z-axes) was segmented into 30-second nonoverlapping windows, and signal-to-image conversion was performed for each segment. Deep features were automatically extracted from the signal images using transfer learning and transformed into a lower-dimensional feature space. These transformed features were then employed to classify the activity behaviours as sleep, sedentary behaviour, and light-intensity (LPA) and moderate-to-vigorous physical activity (MVPA) using a bidirectional long short-term memory (BiLSTM) recurrent neural network. AccNet24 was trained and validated with data from 101 and 25 randomly selected participants and tested with the remaining unseen 25 participants. We also extracted 112 hand-crafted time and frequency domain features from 30-second windows and used them as inputs to five commonly used machine learning classifiers, including random forest, support vector machines, artificial neural networks, decision tree, and naïve Bayes to classify the 24-hour activity behaviour categories.

RESULTS

Using the same training, validation, and test data and window size, the classification accuracy of AccNet24 outperformed the accuracy of the other five machine learning classification algorithms by 16%-30% on unseen data.

CONCLUSION

AccNet24, relying on signal-to-image conversion, deep feature extraction, and BiLSTM achieved consistently high accuracy (>95 %) in classifying the 24-hour activity behaviour categories as sleep, sedentary, LPA, and MVPA. The next generation accelerometry analytics may rely on deep learning techniques for activity prediction.

Validation of an Activity Type Recognition Model Classifying Daily Physical Behavior in Older Adults: The HAR70+ Model

Activity monitoring combined with machine learning (ML) methods can contribute to detailed knowledge about daily physical behavior in older adults. The current study (1) evaluated the performance of an existing activity type recognition ML model (HARTH), based on data from healthy young adults, for classifying daily physical behavior in fit-to-frail older adults, (2) compared the performance with a ML model (HAR70+) that included training data from older adults, and (3) evaluated the ML models on older adults with and without walking aids. Eighteen older adults aged 70-95 years who ranged widely in physical function, including usage of walking aids, were equipped with a chest-mounted camera and two accelerometers during a semi-structured free-living protocol. Labeled accelerometer data from video analysis was used as ground truth for the classification of walking, standing, sitting, and lying identified by the ML models. Overall accuracy was high for both the HARTH model (91%) and the HAR70+ model (94%). The performance was lower for those using walking aids in both models, however, the overall accuracy improved from 87% to 93% in the HAR70+ model. The validated HAR70+ model contributes to more accurate classification of daily physical behavior in older adults that is essential for future research.

Modified prolonged exposure therapy as Early Intervention after Rape (The EIR-study): study protocol for a multicenter randomized add-on superiority trial

BACKGROUND

Sexual assault and rape are the traumatic life events with the highest probability for posttraumatic stress disorder (PTSD), which can have devastating consequences for those afflicted by the condition. Studies indicate that modified prolonged exposure (mPE) therapy may be effective in preventing the development of PTSD in recently traumatized individuals, and especially for people who have experienced sexual assault. If a brief, manualized early intervention can prevent or reduce post-traumatic symptoms in women who have recently experienced rape, healthcare services targeted for these populations (i.e., sexual assault centers, SACs) should consider implementing such interventions as part of routine care.

METHODS/DESIGN

This is a multicenter randomized controlled add-on superiority trial that enrolls patients attending sexual assault centers within 72 h after rape or attempted rape. The objective is to assess whether mPE shortly after rape can prevent the development of post-traumatic stress symptoms. Patients will be randomized to either mPE plus treatment as usual (TAU) or TAU alone. The primary outcome is the development of post-traumatic stress symptoms 3 months after trauma. Secondary outcomes will be symptoms of depression, sleep difficulties, pelvic floor hyperactivity, and sexual dysfunction. The first 22 subjects will constitute an internal pilot trial to test acceptance of the intervention and feasibility of the assessment battery.

DISCUSSION

This study will guide further research and clinical initiatives for implementing strategies for preventing post-traumatic stress symptoms after rape and provide new knowledge about which women may benefit the most from such initiatives and for revising existing treatment guidelines within this area.

TRIAL REGISTRATION

ClinicalTrials.gov NCT05489133. Registered on 3 August 2022.

Real-Time Risk Assessment Detection for Weak People by Parallel Training Logical Execution of a Supervised Learning System Based on an IoT Wearable MEMS Accelerometer

Activity monitoring has become a necessary demand for weak people to guarantee their safety. The paper proposed a Parallel Training Logical Execution (PTLE) system using machine learning (ML) models on a microelectromechanical system (MEMS) accelerometer to detect coughs, falls, and other normal activities. When there are many categories, the ML prediction can be confused between these activities with each other. The PTLE system trains several models in parallel with more specific activity classes in each dataset. The shared tasks between parallel models relieve the complexity for a single one. There are six additional parameters for accelerometer characteristics, which were calculated from three axes accelerations as input features to improve the ML's consciousness. Once all models were trained, the system was ready to receive the input accelerations and activated the logical flow to manage link operation between these ML models for output predictions. Random Forest (RF) had the highest potential among the ML classification algorithms after the validation. In the experiment, the comparison between the PTLE model and the regular ML model were carried out with real-time data from an M5stickC wearable device on the user's chest to the trained models on PC. The result showed the advancement of the proposed method in term of precision, recall, F1-score with an overall accuracy of 98% in the real-time test. The accelerations from the wearable device were sent to ML models via Wi-Fi with Message Queue Telemetry Transport (MQTT) broker, and the activity predictions were transferred to the cloud for the family members or doctor care based on Internet of Things (IoT) communication.

Wrapper-based deep feature optimization for activity recognition in the wearable sensor networks of healthcare systems

The Human Activity Recognition (HAR) problem leverages pattern recognition to classify physical human activities as they are captured by several sensor modalities. Remote monitoring of an individual's activities has gained importance due to the reduction in travel and physical activities during the pandemic. Research on HAR enables one person to either remotely monitor or recognize another person's activity via the ubiquitous mobile device or by using sensor-based Internet of Things (IoT). Our proposed work focuses on the accurate classification of daily human activities from both accelerometer and gyroscope sensor data after converting into spectrogram images. The feature extraction process follows by leveraging the pre-trained weights of two popular and efficient transfer learning convolutional neural network models. Finally, a wrapper-based feature selection method has been employed for selecting the optimal feature subset that both reduces the training time and improves the final classification performance. The proposed HAR model has been tested on the three benchmark datasets namely, HARTH, KU-HAR and HuGaDB and has achieved 88.89%, 97.97% and 93.82% respectively on these datasets. It is to be noted that the proposed HAR model achieves an improvement of about 21%, 20% and 6% in the overall classification accuracies while utilizing only 52%, 45% and 60% of the original feature set for HuGaDB, KU-HAR and HARTH datasets respectively. This proves the effectiveness of our proposed wrapper-based feature selection HAR methodology.

Comparing Handcrafted Features and Deep Neural Representations for Domain Generalization in Human Activity Recognition

Human Activity Recognition (HAR) has been studied extensively, yet current approaches are not capable of generalizing across different domains (i.e., subjects, devices, or datasets) with acceptable performance. This lack of generalization hinders the applicability of these models in real-world environments. As deep neural networks are becoming increasingly popular in recent work, there is a need for an explicit comparison between handcrafted and deep representations in Out-of-Distribution (OOD) settings. This paper compares both approaches in multiple domains using homogenized public datasets. First, we compare several metrics to validate three different OOD settings. In our main experiments, we then verify that even though deep learning initially outperforms models with handcrafted features, the situation is reversed as the distance from the training distribution increases. These findings support the hypothesis that handcrafted features may generalize better across specific domains.

Behavior-based user authentication on mobile devices in various usage contexts

Reliable and non-intrusive user identification and authentication on mobile devices, such as smartphones, are topical tasks today. The majority of state-of-the-art solutions in this domain are based on “device unlock” scenario—checking of information (authentication factors) provided by the user for unlocking a smartphone. As such factors, we may use either single strong authentication factor, for example, password or PIN, or several “weaker” factors, such as tokens, biometrics, or geolocation data. However, these solutions require additional actions from a user, for example, password typing or taking a fingerprint, that may be inappropriate for on-the-fly authentication. In addition, biometric-based user authentication systems tend to be prone to presentation attack (spoofing) and typically perform well in fixed positions only, such as still standing or sitting.

We propose BehaviorID solution that is passwordless (transparent) user-adaptive context-dependent authentication method. The feature of BehaviorID is usage of new “device lock” scenario—smartphone is stayed unlocked and can be fast locked if non-owner’s actions are detected. This is achieved by tracking of user’s behavior with embedded sensors after triggering events, such as actions in banking apps, e-mails, and social services. The advanced adaptive recurrent neural network (A-RNN) is used for accurate estimation and adaptation of behavioral patterns to a new usage context. Thus, proposed BehaviorID solution allows reliable user authentication in various usage contexts by preserving low battery consumption.

Performance evaluation of both state-of-the-art and proposed solutions in various usage contexts proved the effectiveness of BehaviorID in real situations. Proposed solution allows reducing error levels up to three times in comparison with modern Abuhamad’s solutions (Abuhamad et al., IEEE Internet Things J 7(6):5008–5020, 2020) (about $$0.3\%$$ 0.3 %  false acceptance rate (FAR) and $$1.3\%$$ 1.3 %  false rejection rate (FRR)) by preserving high robustness to spoofing attack ($$2.5\%$$ 2.5 %  spoof acceptance rate (SAR)). In addition, BehaviorID showed low drift of error level in case of long-term usage in contrast to modern solutions. This makes the proposed BehaviorID solution an attractive candidate for next-generation behavior-based user authentication systems on mobile devices.

Estimation of Head Accelerations in Crashes Using Neural Networks and Sensors Embedded in the Protective Helmet

Traumatic Brain Injuries (TBIs) are one of the most frequent and severe outcomes of a Powered Two-Wheeler (PTW) crash. Early diagnosis and treatment can greatly reduce permanent consequences. Despite the fact that devices to track head kinematics have been developed for sports applications, they all have limitations, which hamper their use in everyday road applications. In this study, a new technical solution based on accelerometers integrated in a motorcycle helmet is presented, and the related methodology to estimate linear and rotational acceleration of the head with deep Artificial Neural Networks (dANNs) is developed. A finite element model of helmet coupled with a Hybrid III head model was used to generate data needed for the neural network training. Input data to the dANN model were time signals of (virtual) accelerometers placed on the inner surface of the helmet shell, while the output data were the components of linear and rotational head accelerations. The network was capable of estimating, with good accuracy, time patterns of the acceleration components in all impact conditions that require medical treatment. The correlation between the reference and estimated values was high for all parameters and for both linear and rotational acceleration, with coefficients of determination (R2) ranging from 0.91 to 0.97.

Daily motionless activities: A dataset with accelerometer, magnetometer, gyroscope, environment, and GPS data

The dataset presented in this paper presents a dataset related to three motionless activities, including driving, watching TV, and sleeping. During these activities, the mobile device may be positioned in different locations, including the pants pockets, in a wristband, over the bedside table, on a table, inside the car, or on other furniture, for the acquisition of accelerometer, magnetometer, gyroscope, GPS, and microphone data. The data was collected by 25 individuals (15 men and 10 women) in different environments in Covilhã and Fundão municipalities (Portugal). The dataset includes the sensors' captures related to a minimum of 2000 captures for each motionless activity, which corresponds to 2.8 h (approximately) for each one. This dataset includes 8.4 h (approximately) of captures for further analysis with data processing techniques, and machine learning methods. It will be useful for the complementary creation of a robust method for the identification of these type of activities.