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.

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.

HARTH: A Human Activity Recognition Dataset for Machine Learning.

Existing accelerometer-based human activity recognition (HAR) benchmark datasets that were recorded during free living suffer from non-fixed sensor placement, the usage of only one sensor, and unreliable annotations. We make two contributions in this work. First, we present the publicly available Human Activity Recognition Trondheim dataset (HARTH). Twenty-two participants were recorded for 90 to 120 min during their regular working hours using two three-axial accelerometers, attached to the thigh and lower back, and a chest-mounted camera. Experts annotated the data independently using the camera's video signal and achieved high inter-rater agreement (Fleiss' Kappa =0.96). They labeled twelve activities. The second contribution of this paper is the training of seven different baseline machine learning models for HAR on our dataset. We used a support vector machine, k-nearest neighbor, random forest, extreme gradient boost, convolutional neural network, bidirectional long short-term memory, and convolutional neural network with multi-resolution blocks. The support vector machine achieved the best results with an F1-score of 0.81 (standard deviation: ±0.18), recall of 0.85±0.13, and precision of 0.79±0.22 in a leave-one-subject-out cross-validation. Our highly professional recordings and annotations provide a promising benchmark dataset for researchers to develop innovative machine learning approaches for precise HAR in free living.

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.

Learning Then, Learning Now, and Every Second in Between: Lifelong Learning With a Simulated Humanoid Robot.

Long-term human-robot interaction requires the continuous acquisition of knowledge. This ability is referred to as lifelong learning (LL). LL is a long-standing challenge in machine learning due to catastrophic forgetting, which states that continuously learning from novel experiences leads to a decrease in the performance of previously acquired knowledge. Two recently published LL approaches are the Growing Dual-Memory (GDM) and the Self-organizing Incremental Neural Network+ (SOINN+). Both are growing neural networks that create new neurons in response to novel sensory experiences. The latter approach shows state-of-the-art clustering performance on sequentially available data with low memory requirements regarding the number of nodes. However, classification capabilities are not investigated. Two novel contributions are made in our research paper: (I) An extended SOINN+ approach, called associative SOINN+ (A-SOINN+), is proposed. It adopts two main properties of the GDM model to facilitate classification. (II) A new LL object recognition dataset (v-NICO-World-LL) is presented. It is recorded in a nearly photorealistic virtual environment, where a virtual humanoid robot manipulates 100 different objects belonging to 10 classes. Real-world and artificially created background images, grouped into four different complexity levels, are utilized. The A-SOINN+ reaches similar state-of-the-art classification accuracy results as the best GDM architecture of this work and consists of 30 to 350 times fewer neurons, evaluated on two LL object recognition datasets, the novel v-NICO-World-LL and the well-known CORe50. Furthermore, we observe an approximately 268 times lower training time. These reduced numbers result in lower memory and computational requirements, indicating higher suitability for autonomous social robots with low computational resources to facilitate a more efficient LL during long-term human-robot interactions.