Physiological Signal Analysis and Stress Classification from VR Simulations Using Decision Tree Methods.

Extended-Reality Technologies: An Overview of Emerging Applications in Medical Education and Clinical Care.

Although the goal of rectal cancer treatment is to restore gastrointestinal continuity, some patients with rectal cancer develop a permanent stoma (PS) after sphincter-saving operations. Although many studies have identified the risk factors and causes of PS, few have precisely predicted the probability of PS formation before surgery. To validate whether an artificial intelligence model can accurately predict PS formation in patients with rectal cancer after sphincter-saving operations. Patients with rectal cancer who underwent a sphincter-saving operation at Taipei Medical University Hospital between January 1, 2012, and December 31, 2021, were retrospectively included in this study. A machine learning technique was used to predict whether a PS would form after a sphincter-saving operation. We included 19 routinely available preoperative variables in the artificial intelligence analysis. To evaluate the efficiency of the model, 6 performance metrics were utilized: accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiving operating characteristic curve. In our classification pipeline, the data were randomly divided into a training set (80% of the data) and a validation set (20% of the data). The artificial intelligence models were trained using the training dataset, and their performance was evaluated using the validation dataset. Synthetic minority oversampling was used to solve the data imbalance. A total of 428 patients were included, and the PS rate was 13.6% (58/428) in the training set. The logistic regression (LR), Gaussian Naïve Bayes (GNB), Extreme Gradient Boosting (XGB), Gradient Boosting (GB), random forest, decision tree and light gradient boosting machine (LightGBM) algorithms were employed. The accuracies of the logistic regression (LR), Gaussian Naïve Bayes (GNB), Extreme Gradient Boosting (XGB), Gradient Boosting (GB), random forest (RF), decision tree (DT) and light gradient boosting machine (LightGBM) models were 70%, 76%, 89%, 93%, 95%, 79% and 93%, respectively. The area under the receiving operating characteristic curve values were 0.79 for the LR model, 0.84 for the GNB, 0.95 for the XGB, 0.95 for the GB, 0.99 for the RF model, 0.79 for the DT model and 0.98 for the LightGBM model. The key predictors that were identified were the distance of the lesion from the anal verge, clinical N stage, age, sex, American Society of Anesthesiologists score, and preoperative albumin and carcinoembryonic antigen levels. Integration of artificial intelligence with available preoperative data can potentially predict stoma outcomes after sphincter-saving operations. Our model exhibited excellent predictive ability and can improve the process of obtaining informed consent.

Decision trees are a fundamental statistical learning tool for addressing classification and regression problems through a recursive partitioning approach that effectively accommodates numerical and categorical data [1, 2]. The Classification and regression tree (CART) algorithm underlies modern Boosting methodologies such as Gradient boosting machine (GBM), Extreme gradient boosting (XGBoost), and Light gradient boosting machine (LightGBM). However, the standard CART algorithm may require improvement due to its inability to learn from unlabeled data. This study proposes several modifications to incorporate test data into the training phase. Specifically, we introduce a method based on Graph-based semi-supervised learning called “Distance-based Weighting,” which calculates and removes irrelevant records from the training set to accelerate the training process and improve performance. We present Semi-supervised classification and regression tree (Semi-Cart), a new implementation of CART that constructs a decision tree using weighted training data. We evaluated its performance on thirteen datasets from various domains. Our results demonstrate that Semi-Cart outperforms standard CART methods and contributes to statistical learning.

Purpose This study explores the potential of virtual reality (VR) to enhance construction safety training by eliciting critical emotional responses, particularly in novices lacking firsthand experience with hazardous scenarios. By targeting emotions like fear and distress, VR aims to improve hazard recognition and foster risk-averse behaviors, addressing gaps in traditional training methods. Design/methodology/approach The study involved 55 construction management students using a VR simulation integrated with haptic feedback to replicate accident scenarios. Emotional responses were measured through galvanic skin response (GSR) metrics, pre- and post-experience self-reported questionnaires and semi-structured interviews. Participants navigated a virtual construction site, identified hazards and experienced a simulated accident to assess emotional and behavioral impacts. Findings The VR simulation significantly increased emotional arousal, with GSR data showing a strong effect size and significant increases in distress, guilt, fear and shame. Participants reported heightened empathy, guilt and responsibility during post-VR interviews. These findings suggest that VR effectively engages users emotionally, bridging the experiential gap in traditional training by replicating hazardous scenarios with high realism and impact. Practical implications The study highlights VR's potential as a scalable, immersive training tool for eliciting emotional and physiological precursors associated with risk-averse decision-making in the construction industry. By evoking anticipatory emotions theorized to influence risk perception, VR may help prepare novices to engage more cautiously with hazardous scenarios, though the link between these precursors and actual safety behavior requires field validation. Originality/value This research provides empirical evidence of VR's capability to elicit emotional responses that influence decision-making in hazardous environments. It offers valuable insights for developing innovative, emotionally engaging safety training programs for the construction sector.

Driving simulation and training is one of the most typical applications of virtual reality (VR). This paper compares a VR based driving simulator to a desktop based one in terms of the level of user-felt presence and induced arousal. In particular, as the right level of arousal is important in learning and training, we focus on whether there is any difference in the level of arousal induced when different amounts of sensory feedback are given. An experiment was conducted in which subjects experienced two different driving simulators, VR-based (high amount of sensory stimulation) and desktop-based (low), under three different driving conditions (nominal/calm driving, driving on the curve, and driving through a collision accident). The level of presence was measured using a questionnaire, and arousal, indirectly through the galvanic skin response (GSR) and heart rate. Assuming that those physiological signals reflected the level of arousal of the user, our experimental results showed that while the level of presence was higher with the VR-based set up, no significant differences were found for arousal between the two system set ups. Arousal was dependent only on the task conditions. For one, this is contrary to several previous works [Meehan et al. 1998][Brogni et al. 2006] that had identified physiological signals (such as heart rate and GSR), as a reliable measure of presence. Even though VR simulators may not generate a differentiated amount of arousal, the training may still be effective for physical/spatial tasks (such as driving) due to other factors such as a large field of view and stereoscopy. However, for training of logical tasks, the cost to benefit ratio of VR based simulators may be too high.

The increasing fame of biofeedback game has brought convenience to human life. More and more people rely on biofeedback game as an alternative medical treatment to overcome stress problems. Practitioners, meditators, healthcare clinicians and patients are among the most common people who use this biofeedback method to handle their stress in a better way. Basically, biofeedback game integrates biofeedback device in their therapy. Different biofeedback devices have unique way to patterned user’s physiological response towards the game. Biofeedback game implements the virtual game to collaborate with these biofeedback devices. Global issues such as the Electric Vehicles (EVs) issue which are part of the green technology solution are rarely found in the existing biofeedback virtual games themes. Therefore this project, GreenVec Game for Skin Conductivity Level (SCL) Biofeedback Performance Simulator using Galvanic Skin Response (GSR) Sensor or mainly known as Green Vehicle Biofeedback Game is a biofeedback game-based application that able to measure SCL of the player with the integration of GSR. The skin conductivity level corresponds to the player’s awareness level about the green Electric Vehicles (EVs) issues. The biofeedback game will also create a short report on the player’s skin conductivity level based on the GSR sensor data and the game data by the end of the game. From the end result, it shows that when the SCL is high, the awareness on the electric vehicles issues is high. While the SCL is low, the awareness on the electric vehicles issues is low.

The concrete structure under constant loading generates the creep strain, which directly influences the durability of concrete, especially self-compacting concrete (SCC). This study improved the fib MC2010 model to predict both basic creep and total creep of SCC based on the expanded laboratory database of SCC with a total of 281 data sets of creep function. The improved model, namely the fib-SCC model, exhibited the best efficiency in creep prediction of SCC compared to other models, such as the fib MC2010 model and the other available models of SCC. Additionally, new methods were introduced to predict the creep of SCC based on the machine learning methods, such as decision tree, random forest, extra trees, gradient boosting, and extreme gradient boosting. The results indicated that the gradient boosting and XGBoost methods were more prone than decision tree, random forest, and extra trees methods in overfitting the total creep database of SCC.

Can virtual reality simulation be used for advanced bariatric surgical training?

With the rapid development of artificial intelligence, human-computer interaction, pattern recognition and other technologies, emotion recognition has become a hot topic in this field. Traditional emotional recognition studies mostly use voice features and facial expression image features for recognition, but the external expression features of these emotions are easily subject to subjective control of the human body. However, physiological information is closely related to the cerebral cortex and nerve center of human body, which is objective and authentic. In this paper, four kinds of chaotic characteristic parameters were extracted from ECG(Electrocardiogram), SC(Skin Conductance) and RSP(Respiration), including complexity, box dimension, approximate entropy and information entropy. Three kinds of emotions (Joy, Anger and Sadness) were identified by C4.5 decision tree algorithm. The results of the study show that this method is feasible for emotion recognition. Using C# programming language, Visual Studio integrated development environment (IDE), SQL Server database and other tools, a emotional recognition platform based on multi-physiological information was established, which can extract 12 chaotic characteristic parameters from the collected ECG, SC and RSP. Joy, anger and sadness were recognized through the C4.5 decision tree classifier algorithm, and finally save the information to the local database. This platform includes user login, volunteer management, administrator management, data center and other functional modules to ensure the security and information integrity of the platform. The verification experiment was carried out on the completed platform(In this paper, omit), which proved the effectiveness and practicability of the platform.

For situations where the data are drawn from reasonably homogeneous populations, traditional methods such as multiple regression typically yield insightful analyses. For situations where the data are drawn from more heterogeneous populations, decision tree approaches, such as Classification and Regression Trees (CART) and Generalized, Unbiased, Interaction, Detection, and Estimation (GUIDE), are more likely to recognize idiosyncratic subpopulations and interactions automatically. In contrast to CART, however, GUIDE yields models with better predictive performance for each subpopulation. This article extends the idea of GUIDE to handle analysis of covariance-type problems. This article compares GUIDE modeling to various decision tree methods and to multiple regression. The article identifies and discusses the relative advantages and disadvantages of multiple regression, CART, and GUIDE. GUIDE produces quality or reliability models that exhibit greater predictive accuracy than multiple regression or CART for complex, highly diverse populations. Also, GUIDE is readily applicable to many other areas, such as repairability and maintainability settings involving both qualitative and quantitative variables. A small case study of an engineered wood product, medium-density fiberboard, is presented to illustrate the application of GUIDE. Accepted in 2005 for a special issue on Reliability co-edited by Hoang Pham, Rutgers University; Dong Ho Park, Hallym University, Korea; and Richard Cassady, University of Arkansas.

Minimally invasive laparoscopic surgery requires intensive training due to challenges such as loss of haptic feedback and depth perception. Traditional training methods include box trainers (BT), which offer realistic haptic feedback but lack objective performance assessment, and virtual reality (VR) simulators, which provide automated feedback but lack haptic feedback. This review, conducted at the Barts Cancer Institute, Queen Mary University, examines the learning outcomes of VRsimulators with haptic feedback compared to BT. A systematic review and meta-analysis was conducted following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines from December 2023 to April 2024. Research databases, such as PubMed, EMBASE, CINAHL, and Web of Science, were searched for randomized controlled trials (RCTs) comparing VR simulators with haptic feedback to BT in training medical students. Seven RCTs met the inclusion criteria, and four were included in the meta-analysis. The primary outcomes were learning curves and learning effectswhile secondary outcomes included skill transfer to a surgical environment. The review analysed data from 125 participants across the studies. Results indicated that BTs demonstrated a superior learning curve, with participants achieving proficiency faster than those using VR. Both simulators showed significant learning effects; however, BTs resulted in greater improvements across more performance parameters. Regarding skill transfer to surgical environments, there was no significant difference between the two groups, suggesting both approaches effectively support surgical skill transfer. Overall, BT has a more effective learning curve and marginally better performance in skill acquisition. While VR with haptic feedback offers enhanced realism, it does not fully replicate the natural haptic feedback provided by BT. Further studies are needed to improve VR haptic feedback and its integration into training programs to enhance learning outcomes.

Development of three-dimensional visualisation technology of the aerodynamic environment in a greenhouse using CFD and VR technology, Part 2: Development of an educational VR simulator

Modeling of necking area reduction of carbon steel in hydrogen environment using machine learning approach

Retrospective cohort study. This study aimed to develop machine-learning algorithms to predict ambulation outcomes following surgery for spinal metastasis. Postoperative ambulation status following spinal metastasis surgery is currently difficult to predict. The improved ability to predict this important postoperative outcome would facilitate management decision-making and help in determining realistic treatment goals. This retrospective study included patients who underwent spinal metastasis at a university-based medical center in Thailand between January 2009 and November 2021. Collected data included preoperative parameters and ambulatory status 90 and 180 days following surgery. Thirteen machine-learning algorithms, namely, artificial neural network, logistic regression, CatBoost classifier, linear discriminant analysis, extreme gradient boosting, extra trees classifier, random forest classifier, gradient boosting classifier, light gradient boosting machine, naïve Bayes, K-neighbor classifier, Ada boost classifier, and decision tree classifier were developed to predict ambulatory status 90 and 180 days following surgery. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and F1-score. In total, 167 patients were enrolled. The number of patients classified as ambulatory 90 and 180 days following surgery was 140 (81.9%) and 137 (82.0%), respectively. The extreme gradient boosting algorithm was found to most accurately predict 180-day ambulatory outcome (AUC, 0.85; F1-score, 0.90), and the decision tree algorithm most accurately predicted 90-day ambulatory outcome (AUC, 0.94; F1-score, 0.88). Machine-learning algorithms were effective in predicting ambulatory status following surgery for spinal metastasis. Based on our data, the extreme gradient boosting and decision tree best predicted postoperative ambulatory status 180 and 90 days after spinal metastasis surgery, respectively.

BackgroundExtended reality, which encompasses virtual reality (VR), augmented reality (AR), and mixed reality (MR), is increasingly used in medical education. Studies assessing the effectiveness of these new educational modalities should measure relevant outcomes using outcome measurement tools with validity evidence.ObjectiveOur aim is to determine the choice of outcomes, measurement instruments, and the use of measurement instruments with validity evidence in randomized controlled trials (RCTs) on the effectiveness of VR, AR, and MR in medical student education.MethodsWe conducted a systematic mapping review. We searched 7 major bibliographic databases from January 1990 to April 2020, and 2 reviewers screened the citations and extracted data independently from the included studies. We report our findings in line with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines.ResultsOf the 126 retrieved RCTs, 115 (91.3%) were on VR and 11 (8.7%) were on AR. No RCT on MR in medical student education was found. Of the 115 studies on VR, 64 (55.6%) were on VR simulators, 30 (26.1%) on screen-based VR, 9 (7.8%) on VR patient simulations, and 12 (10.4%) on VR serious games. Most studies reported only a single outcome and immediate postintervention assessment data. Skills outcome was the most common outcome reported in studies on VR simulators (97%), VR patient simulations (100%), and AR (73%). Knowledge was the most common outcome reported in studies on screen-based VR (80%) and VR serious games (58%). Less common outcomes included participants’ attitudes, satisfaction, cognitive or mental load, learning efficacy, engagement or self-efficacy beliefs, emotional state, competency developed, and patient outcomes. At least one form of validity evidence was found in approximately half of the studies on VR simulators (55%), VR patient simulations (56%), VR serious games (58%), and AR (55%) and in a quarter of the studies on screen-based VR (27%). Most studies used assessment methods that were implemented in a nondigital format, such as paper-based written exercises or in-person assessments where examiners observed performance (72%).ConclusionsRCTs on VR and AR in medical education report a restricted range of outcomes, mostly skills and knowledge. The studies largely report immediate postintervention outcome data and use assessment methods that are in a nondigital format. Future RCTs should include a broader set of outcomes, report on the validity evidence of the measurement instruments used, and explore the use of assessments that are implemented digitally.