Rachel L. Elkin, Jeff M. Beaubien, Nathaniel Damaghi, Todd P. Chang, David O. Kessler
{"title":"Dynamic Cognitive Load Assessment in Virtual Reality","authors":"Rachel L. Elkin, Jeff M. Beaubien, Nathaniel Damaghi, Todd P. Chang, David O. Kessler","doi":"10.1177/10468781241248821","DOIUrl":null,"url":null,"abstract":"BackgroundRecent advances in non-invasive physiologic monitoring leverage machine learning to provide unobtrusive, real-time assessments of a learner’s cognitive load (CL) as they engage in specific tasks. However, the performance characteristics of these novel composite physiologic CL measures are incompletely understood.ObjectivesWe aimed to 1) explore the feasibility of measuring CL in real time using physiologically-derived inputs; 2) evaluate the performance characteristics of a novel composite CL measure during simulated virtual reality resuscitations; and 3) understand how this measure compares to traditional, self-reported measures of CL.MethodsNovice (PGY1-2 pediatric residents) and expert (pediatric emergency medicine fellows and attendings) participants completed four virtual reality simulations as team leader. The scenario content (status epilepticus versus anaphylaxis) and level of distraction (high versus low) were manipulated. Cognitive load was measured in all participants using electroencephalography and electrocardiography data (“real-time CL”) as well as through self-report (NASA-TLX). Scenario performance also was measured.ResultsComplete data were available for 6 experts and 6 novices. Experts generally had lower CL than novices on both measures. Both measures localized the most significant differences between groups to the anaphylaxis scenarios (real-time CL [low-distraction] Cohen’s d -1.33 [95% CI -.2.56, -0.03] and self-reported CL [high-distraction] Cohen’s d -1.41 [95% CI -2.67, -0.10]). No consistent differences were seen with respect to level of distraction. Performance was similar between the two groups, though both exhibited fewer errors over time (F<jats:sub>(3,48)</jats:sub> = 5.75, p = .002).ConclusionIt is feasible to unobtrusively measure cognitive load in real time during virtual reality simulations. There was convergence between the two CL measures: in both, experts had lower CL than novices, with the most significant effect size differences in the more challenging anaphylaxis scenarios.","PeriodicalId":47521,"journal":{"name":"SIMULATION & GAMING","volume":"94 1","pages":""},"PeriodicalIF":1.5000,"publicationDate":"2024-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"SIMULATION & GAMING","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1177/10468781241248821","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"EDUCATION & EDUCATIONAL RESEARCH","Score":null,"Total":0}
引用次数: 0
Abstract
BackgroundRecent advances in non-invasive physiologic monitoring leverage machine learning to provide unobtrusive, real-time assessments of a learner’s cognitive load (CL) as they engage in specific tasks. However, the performance characteristics of these novel composite physiologic CL measures are incompletely understood.ObjectivesWe aimed to 1) explore the feasibility of measuring CL in real time using physiologically-derived inputs; 2) evaluate the performance characteristics of a novel composite CL measure during simulated virtual reality resuscitations; and 3) understand how this measure compares to traditional, self-reported measures of CL.MethodsNovice (PGY1-2 pediatric residents) and expert (pediatric emergency medicine fellows and attendings) participants completed four virtual reality simulations as team leader. The scenario content (status epilepticus versus anaphylaxis) and level of distraction (high versus low) were manipulated. Cognitive load was measured in all participants using electroencephalography and electrocardiography data (“real-time CL”) as well as through self-report (NASA-TLX). Scenario performance also was measured.ResultsComplete data were available for 6 experts and 6 novices. Experts generally had lower CL than novices on both measures. Both measures localized the most significant differences between groups to the anaphylaxis scenarios (real-time CL [low-distraction] Cohen’s d -1.33 [95% CI -.2.56, -0.03] and self-reported CL [high-distraction] Cohen’s d -1.41 [95% CI -2.67, -0.10]). No consistent differences were seen with respect to level of distraction. Performance was similar between the two groups, though both exhibited fewer errors over time (F(3,48) = 5.75, p = .002).ConclusionIt is feasible to unobtrusively measure cognitive load in real time during virtual reality simulations. There was convergence between the two CL measures: in both, experts had lower CL than novices, with the most significant effect size differences in the more challenging anaphylaxis scenarios.
期刊介绍:
Simulation & Gaming: An International Journal of Theory, Practice and Research contains articles examining academic and applied issues in the expanding fields of simulation, computerized simulation, gaming, modeling, play, role-play, debriefing, game design, experiential learning, and related methodologies. The broad scope and interdisciplinary nature of Simulation & Gaming are demonstrated by the wide variety of interests and disciplines of its readers, contributors, and editorial board members. Areas include: sociology, decision making, psychology, language training, cognition, learning theory, management, educational technologies, negotiation, peace and conflict studies, economics, international studies, research methodology.