Contents lists available at ScienceDirect International Journal of Medical Informatics journal homepage: www.elsevier.com/locate/ijmedinf Review article Vision-based serious games and virtual reality systems for motor rehabilitation: A review geared toward a research methodology Ines Ayed a,c , Adel Ghazel c , Antoni Jaume-i-Capó a, ⁎ , Gabriel Moyà-Alcover a , Javier Varona a , Pau Martínez-Bueso b a Unitat de Gràfics i Visió per Computador i Inteligència Artificial, Departament de Ciències Matemàtiques i Informàtica, Universitat de les Illes Balears, Spain b Research Group on Evidence, Lifestyles and Health, Department of Nursing and Physiotherapy, University of the Balearic Islands, Spain c GresCom Lab, Ecole Supérieure des Communications de Tunis, Université de Carthage, Tunis ARTICLE INFO Keywords: Vision-based sensors Clinical trials Motor rehabilitation Serious games Research methodology Systematic literature review ABSTRACT Background: Nowadays, information technologies are being widely adopted to promote healthcare and re- habilitation. Owing to their affordability and use of hand-free controllers, vision-based systems have gradually been integrated into motor rehabilitation programs and have greatly drawn the interest of healthcare practi- tioners and the research community. Many studies have illustrated the effectiveness of these systems in re- habilitation. However, the report and design aspects of the reported clinical trials were disregarded. Objective: In this paper, we present a systematic literature review of the use of vision-based serious games and virtual reality systems in motor rehabilitation programs. We aim to propose a research methodology that en- gineers can use to improve the designing and reporting processes of their clinical trials. Methods: We conducted a review of published studies that entail clinical experiments. Searches were performed using Web of Science and Medline (PubMed) electronic databases, and selected studies were assessed using the Downs and Black Checklist and then analyzed according to specific research questions. Results: We identified 86 studies and our findings indicate that the number of studies in this field is increasing, with Korea and USA in the lead. We found that Kinect, EyeToy system, and GestureTek IREX are the most commonly used technologies in studying the effects of vision-based serious games and virtual reality systems on rehabilitation. Findings also suggest that cerebral palsy and stroke patients are the main target groups, with a particular interest on the elderly patients in this target population. The findings indicate that most of the studies focused on postural control and upper extremity exercises and used different measurements during assessment. Conclusions: Although the research community’s interest in this area is growing, many clinical trials lack suf- ficient clarity in many aspects and are not standardized. Some recommendations have been made throughout the article. 1. Introduction Vision-based interaction sensors (VBI) [1] can visually capture in- formation on the performance of users’ actions. Different studies have validated their effectiveness in rehabilitation purposes such as postural control [2], clinical functional analysis and rehabilitation [3], gait re- training [4], activities of daily living rehabilitation [5], and coaching of the elderly population [6]. Visual information on the performance of patients’ actions is the preferred data collection method for two rea- sons: first, because motor rehabilitation consists of body movements that can be recorded; and second, because vision capture technology is non-invasive and can be used by patients who have difficulty holding physical devices. VBI systems aim to provide reliable computer methods to detect and analyze human movements. As new vision-based tech- nologies emerged, diverse disciplines’ interest in this area of research grew rapidly. This growth of interest has made the interpretation of results and the extraction of broader principles from existing work more challenging. One possible solution to this challenge is to adopt an evi- dence-based paradigm. To understand the role of evidence, we need to recognize that across diverse study disciplines, there is a common need for methods that allow objective and consistent aggregation of out- comes in multiple empirical studies [7]. In this context, evidence is defined as the synthesis of the best scientific studies on a specific topic or research question. The primary method of synthesis is a systematic https://doi.org/10.1016/j.ijmedinf.2019.06.016 Received 7 September 2018; Received in revised form 19 May 2019; Accepted 17 June 2019 ⁎ Corresponding author. E-mail address: antoni.jaume@uib.es (A. Jaume-i-Capó). International Journal of Medical Informatics 130 (2019) 103909 1386-5056/ © 2019 Elsevier B.V. All rights reserved. T