A Feasibility Study of a Wearable Real-Time Notification System for Self-Awareness of Body-Rocking Behavior

Improving Performance and Quantifying Uncertainty of Body-Rocking Detection Using Bayesian Neural Networks

Body-rocking is an undesired stereotypical motor movement performed by some individuals, and its detection is essential for self-awareness and habit change. We envision a pipeline that includes inertial wearable sensors and a real-time detection system for notifying the user so that they are aware of their body-rocking behavior. For this task, similarities of body rocking to other non-related repetitive activities may cause false detections which prevent continuous engagement, leading to alarm fatigue. We present a pipeline using Bayesian Neural Networks with uncertainty quantification for jointly reducing false positives and providing accurate detection. We show that increasing model capacity does not consistently yield higher performance by itself, while pairing it with the Bayesian approach does yield significant improvements. Disparities in uncertainty quantification are better quantified by calibrating them using deep neural networks. We show that the calibrated probabilities are effective quality indicators of reliable predictions. Altogether, we show that our approach provides additional insights on the role of Bayesian techniques in deep learning as well as aids in accurate body-rocking detection, improving our prior work on this subject.

Augmented, Mixed, and Virtual Reality-Based Head-Mounted Devices for Medical Education: Systematic Review

BACKGROUND

Augmented reality (AR), mixed reality (MR), and virtual reality (VR), realized as head-mounted devices (HMDs), may open up new ways of teaching medical content for low-resource settings. The advantages are that HMDs enable repeated practice without adverse effects on the patient in various medical disciplines; may introduce new ways to learn complex medical content; and may alleviate financial, ethical, and supervisory constraints on the use of traditional medical learning materials, like cadavers and other skills lab equipment.

OBJECTIVE

We examine the effectiveness of AR, MR, and VR HMDs for medical education, whereby we aim to incorporate a global health perspective comprising low- and middle-income countries (LMICs).

METHODS

We conducted a systematic review according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) and Cochrane guidelines. Seven medical databases (PubMed, Cochrane Library, Web of Science, Science Direct, PsycINFO, Education Resources Information Centre, and Google Scholar) were searched for peer-reviewed publications from January 1, 2014, to May 31, 2019. An extensive search was carried out to examine relevant literature guided by three concepts of extended reality (XR), which comprises the concepts of AR, MR, and VR, and the concepts of medicine and education. It included health professionals who took part in an HMD intervention that was compared to another teaching or learning method and evaluated with regard to its effectiveness. Quality and risk of bias were assessed with the Medical Education Research Study Quality Instrument, the Newcastle-Ottawa Scale-Education, and A Cochrane Risk of Bias Assessment Tool for Non-Randomized Studies of Interventions. We extracted relevant data and aggregated the data according to the main outcomes of this review (knowledge, skills, and XR HMD).

RESULTS

A total of 27 studies comprising 956 study participants were included. The participants included all types of health care professionals, especially medical students (n=573, 59.9%) and residents (n=289, 30.2%). AR and VR implemented with HMDs were most often used for training in the fields of surgery (n=13, 48%) and anatomy (n=4, 15%). A range of study designs were used, and quantitative methods were clearly dominant (n=21, 78%). Training with AR- and VR-based HMDs was perceived as salient, motivating, and engaging. In the majority of studies (n=17, 63%), HMD-based interventions were found to be effective. A small number of included studies (n=4, 15%) indicated that HMDs were effective for certain aspects of medical skills and knowledge learning and training, while other studies suggested that HMDs were only viable as an additional teaching tool (n=4, 15%). Only 2 (7%) studies found no effectiveness in the use of HMDs.

CONCLUSIONS

The majority of included studies suggested that XR-based HMDs have beneficial effects for medical education, whereby only a minority of studies were from LMICs. Nevertheless, as most studies showed at least noninferior results when compared to conventional teaching and training, the results of this review suggest applicability and potential effectiveness in LMICs. Overall, users demonstrated greater enthusiasm and enjoyment in learning with XR-based HMDs. It has to be noted that many HMD-based interventions were small-scale and conducted as short-term pilots. To generate relevant evidence in the future, it is key to rigorously evaluate XR-based HMDs with AR and VR implementations, particularly in LMICs, to better understand the strengths and shortcomings of HMDs for medical education.

How We Found Our IMU: Guidelines to IMU Selection and a Comparison of Seven IMUs for Pervasive Healthcare Applications

Inertial measurement units (IMUs) are commonly used for localization or movement tracking in pervasive healthcare-related studies, and gait analysis is one of the most often studied topics using IMUs. The increasing variety of commercially available IMU devices offers convenience by combining the sensor modalities and simplifies the data collection procedures. However, selecting the most suitable IMU device for a certain use case is increasingly challenging. In this study, guidelines for IMU selection are proposed. In particular, seven IMUs were compared in terms of their specifications, data collection procedures, and raw data quality. Data collected from the IMUs were then analyzed by a gait analysis algorithm. The difference in accuracy of the calculated gait parameters between the IMUs could be used to retrace the issues in raw data, such as acceleration range or sensor calibration. Based on our algorithm, we were able to identify the best-suited IMUs for our needs. This study provides an overview of how to select the IMUs based on the area of study with concrete examples, and gives insights into the features of seven commercial IMUs using real data.