LightPRA: A Lightweight Temporal Convolutional Network for Automatic Physical Rehabilitation Exercise Assessment.

Research evidence shows that physical rehabilitation exercises prescribed by medical experts can assist in restoring physical function, improving life quality, and promoting independence for physically disabled individuals. In response to the absence of immediate expert feedback on performed actions, developing a Human Action Evaluation (HAE) system emerges as a valuable automated solution, addressing the need for accurate assessment of exercises and guidance during physical rehabilitation. Previous HAE systems developed for the rehabilitation exercises have focused on developing models that utilize skeleton data as input to compute a quality score for each action performed by the patient. However, existing studies have focused on improving scoring performance while often overlooking computational efficiency. In this research, we propose LightPRA (Light Physical Rehabilitation Assessment) system, an innovative architectural solution based on a Temporal Convolutional Network (TCN), which harnesses the capabilities of dilated causal Convolutional Neural Networks (CNNs). This approach efficiently captures complex temporal features and characteristics of the skeleton data with lower computational complexity, making it suitable for real-time feedback provided on resource-constrained devices such as Internet of Things (IoT) devices and Edge computing frameworks. Through empirical analysis performed on the University of Idaho-Physical Rehabilitation Movement Data (UI-PRMD) and KInematic assessment of MOvement for remote monitoring of physical REhabilitation (KIMORE) datasets, our proposed LightPRA model demonstrates superior performance over several state-of-the-art approaches such as Spatial-Temporal Graph Convolutional Network (STGCN) and Long Short-Term Memory (LSTM)-based models in scoring human activity performance, while exhibiting lower computational cost and complexity.

A Multi-Activity Fusion Approach for Gender Recognition based on Human Activity.

Due to its advantages in numerous industries, including healthcare, sports, rehabilitation, and wearable electronics, gender recognition has garnered a lot of attention in the last ten years. The gender recognition method described in this study uses a wearable sensor device with inertial measurement units to record a variety of activities. The system consists of five sensors that are mounted to the upper and lower bodies while performing seven standing, walking, and climbing exercises that are meant to replicate daily activity. To create a model for gender recognition, we carried out an extensive study based on supervised machine learning. This study identifies a collection of sensor locations and behaviours to better precisely classify gender. Gender classification based on single activity was performed using Random Forest Classifier (RFC) and Support Vector Machines (SVM). Maximum accuracy of 92.06% was gained using Random Forest Classifier for the sensor located at the ankle when walking. Multi-activity based gender classification outperformed former by achieving an accuracy of 94.13% using RFC. This was for the activity combination of Romberg test eyes open, Single leg stance eyes open and Staircase up and down.

Application of a Comprehensive Evaluation Framework to COVID-19 Studies: Systematic Review of Translational Aspects of Artificial Intelligence in Health Care.

Background: Despite immense progress in artificial intelligence (AI) models, there has been limited deployment in health care environments. The gap between potential and actual AI applications is likely due to the lack of translatability between controlled research environments (where these models are developed) and clinical environments for which the AI tools are ultimately intended. Objective: We previously developed the Translational Evaluation of Healthcare AI (TEHAI) framework to assess the translational value of AI models and to support successful transition to health care environments. In this study, we applied the TEHAI framework to the COVID-19 literature in order to assess how well translational topics are covered. Methods: A systematic literature search for COVID-19 AI studies published between December 2019 and December 2020 resulted in 3830 records. A subset of 102 (2.7%) papers that passed the inclusion criteria was sampled for full review. The papers were assessed for translational value and descriptive data collected by 9 reviewers (each study was assessed by 2 reviewers). Evaluation scores and extracted data were compared by a third reviewer for resolution of discrepancies. The review process was conducted on the Covidence software platform. Results: We observed a significant trend for studies to attain high scores for technical capability but low scores for the areas essential for clinical translatability. Specific questions regarding external model validation, safety, nonmaleficence, and service adoption received failed scores in most studies. Conclusions: Using TEHAI, we identified notable gaps in how well translational topics of AI models are covered in the COVID-19 clinical sphere. These gaps in areas crucial for clinical translatability could, and should, be considered already at the model development stage to increase translatability into real COVID-19 health care environments.

Artificial Intelligence for skeleton-based physical rehabilitation action evaluation: A systematic review.

Performing prescribed physical exercises during home-based rehabilitation programs plays an important role in regaining muscle strength and improving balance for people with different physical disabilities. However, patients attending these programs are not able to assess their action performance in the absence of a medical expert. Recently, vision-based sensors have been deployed in the activity monitoring domain. They are capable of capturing accurate skeleton data. Furthermore, there have been significant advancements in Computer Vision (CV) and Deep Learning (DL) methodologies. These factors have promoted the solutions for designing automatic patient's activity monitoring models. Then, improving such systems' performance to assist patients and physiotherapists has attracted wide interest of the research community. This paper provides a comprehensive and up-to-date literature review on different stages of skeleton data acquisition processes for the aim of physio exercise monitoring. Then, the previously reported Artificial Intelligence (AI) - based methodologies for skeleton data analysis will be reviewed. In particular, feature learning from skeleton data, evaluation, and feedback generation for the purpose of rehabilitation monitoring will be studied. Furthermore, the associated challenges to these processes will be reviewed. Finally, the paper puts forward several suggestions for future research directions in this area.

Multiomics, virtual reality and artificial intelligence in heart failure.

Aim: Multiomics delivers more biological insight than targeted investigations. We applied multiomics to patients with heart failure (HF) and reduced ejection fraction (HFrEF), with machine learning applied to advanced ECG (AECG) and echocardiography artificial intelligence (Echo AI). Patients & methods: In total, 46 patients with HFrEF and 20 controls underwent metabolomic profiling, including liquid/gas chromatography-mass spectrometry and solid-phase microextraction volatilomics in plasma and urine. HFrEF was defined using left ventricular (LV) global longitudinal strain, EF and N-terminal pro hormone BNP. AECG and Echo AI were performed over 5 min, with a subset of patients undergoing a virtual reality mental stress test. Results: A-ECG had similar diagnostic accuracy as N-terminal pro hormone BNP for HFrEF (area under the curve = 0.95, 95% CI: 0.85-0.99), and correlated with global longitudinal strain (r = -0.77, p < 0.0001), while Echo AI-generated measurements correlated well with manually measured LV end diastolic volume r = 0.77, LV end systolic volume r = 0.8, LVEF r = 0.71, indexed left atrium volume r = 0.71 and indexed LV mass r = 0.6, p < 0.005. AI-LVEF and other HFrEF biomarkers had a similar discrimination for HFrEF (area under the curve AI-LVEF = 0.88; 95% CI: -0.03 to 0.15; p = 0.19). Virtual reality mental stress test elicited arrhythmic biomarkers on AECG and indicated blunted autonomic responsiveness (alpha 2 of RR interval variability, p = 1 × 10-4) in HFrEF. Conclusion: Multiomics-related machine learning shows promise for the assessment of HF.

Lay abstract Multiomics is the integration of multiple sources of health information, for example, genomic, metabolite, etc. This delivers more insight than targeted single investigations and provides an ability to perceive subtle individual differences between people. In this study we applied multiomics to patients with heart failure (HF) using DNA sequencing, metabolomics and machine learning applied to ECG echocardiography. We demonstrated significant differences between subsets of patients with HF using these methods. We also showed that machine learning has significant diagnostic potential in identifying HF patients more efficiently than manual or conventional techniques.