This paper is in the following e-collection/theme issue:

Robots in Healthcare (79) Robotics in Rehabilitation (36) Technology in Physiotherapy (122) Telerehabilitation (141)

Published on in Vol 21, No 5 (2019): May

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/12708, first published .
The Role of Social Interactions in Motor Performance: Feasibility Study Toward Enhanced Motivation in Telerehabilitation

The Role of Social Interactions in Motor Performance: Feasibility Study Toward Enhanced Motivation in Telerehabilitation

The Role of Social Interactions in Motor Performance: Feasibility Study Toward Enhanced Motivation in Telerehabilitation

Authors of this article:

Roni Barak Ventura1 Author Orcid Image ;   Shinnosuke Nakayama1 Author Orcid Image ;   Preeti Raghavan2 Author Orcid Image ;   Oded Nov3 Author Orcid Image ;   Maurizio Porfiri1, 4 Author Orcid Image
Article Authors Cited by (11) Tweetations (2) Metrics

    Original Paper

    1Department of Mechanical and Aerospace Engineering, New York University Tandon School of Engineering, Brooklyn, NY, United States

    2Department of Rehabilitation Medicine, New York University School of Medicine, New York, NY, United States

    3Department of Technology Management and Innovation, New York University Tandon School of Engineering, Brooklyn, NY, United States

    4Department of Biomedical Engineering, New York University Tandon School of Engineering, Brooklyn, NY, United States

    Corresponding Author:

    Maurizio Porfiri, PhD

    Department of Mechanical and Aerospace Engineering

    New York University Tandon School of Engineering

    6 MetroTech Center

    Brooklyn, NY, 11201

    United States

    Phone: 1 646 997 3681

    Fax:1 646 997 3532

    Email: mporfiri@nyu.edu


    Background: Robot-mediated telerehabilitation has the potential to provide patient-tailored cost-effective rehabilitation. However, compliance with therapy can be a problem that undermines the prospective advantages of telerehabilitation technologies. Lack of motivation has been identified as a major factor that hampers compliance. Exploring various motivational interventions, the integration of citizen science activities in robotics-based rehabilitation has been shown to increase patients’ motivation to engage in otherwise tedious exercises by tapping into a vast array of intrinsic motivational drivers. Patient engagement can be further enhanced by the incorporation of social interactions.

    Objective: Herein, we explored the possibility of bolstering engagement in physical therapy by leveraging cooperation among users in an environmental citizen science project. Specifically, we studied how the integration of cooperation into citizen science influences user engagement, enjoyment, and motor performance. Furthermore, we investigated how the degree of interdependence among users, such that is imposed through independent or joint termination (JT), affects participation in citizen science-based telerehabilitation.

    Methods: We developed a Web-based citizen science platform in which users work in pairs to classify images collected by an aquatic robot in a polluted water canal. The classification was carried out by labeling objects that appear in the images and trashing irrelevant labels. The system was interfaced by a haptic device for fine motor rehabilitation. We recruited 120 healthy volunteers to operate the platform. Of these volunteers, 98 were cooperating in pairs, with 1 user tagging images and the other trashing labels. The other 22 volunteers performed both tasks alone. To vary the degree of interdependence within cooperation, we implemented independent and JTs.

    Results: We found that users’ engagement and motor performance are modulated by their assigned task and the degree of interdependence. Motor performance increased when users were subjected to independent termination (P=.02), yet enjoyment decreased when users were subjected to JT (P=.005). A significant interaction between the type of termination and the task was found to influence productivity (P<.001) as well as mean speed, peak speed, and path length of the controller (P=.01, P=.006, and P<.001, respectively).

    Conclusions: Depending on the type of termination, cooperation was not always positively associated with engagement, enjoyment, and motor performance. Therefore, enhancing user engagement, satisfaction, and motor performance through cooperative citizen science tasks relies on both the degree of interdependence among users and the perceived nature of the task. Cooperative citizen science may enhance motivation in robotics-based telerehabilitation, if designed attentively.

    J Med Internet Res 2019;21(5):e12708

    doi:10.2196/12708

    Keywords

    citizen science; social interactions; telerehabilitation; physical therapy 


    Background

    Debilitating neurological diseases such as stroke require intensive, repetitive, and high-frequency physical therapy for maximum recovery of motor function and self-reliance [1-3]. However, costly resources and limited rehabilitation personnel make rehabilitation unavailable to the majority of patients. Furthermore, disability often encumbers mobility, preventing patients from leaving their homes and frequenting the therapists’ office [4]. Therefore, reaching their full recovery potential is greatly contingent upon performing self-directed physical therapy with limited professional feedback.

    Robot-Mediated Telerehabilitation

    Several rehabilitation robots have been developed for delivery of exercise programs for the upper limb, including the MIT-Manus (Massachusetts Institute of Technology) [5,6], Gentle/S (University of Reading) [7], ArmIn (ETH Zurich) [8], and Mirror Image Movement Enabler (Stanford University) [9]. Ultimately, robotic rehabilitation devices aim to administer and monitor exercise for the arm with reproducible high-intensity and high-dosage sensorimotor therapy while collecting pertinent data for assessment by a medical professional [10,11].

    The major hurdles in the widespread adoption of these robotics-based technologies are costs and user-friendliness, whereby these devices often have prohibitive costs to the general public and require some form of technological proficiency that may be beyond the typical background of patients or even therapists [12,13]. To fill these gaps, several studies have explored the feasibility of delivering rehabilitation treatments using low-cost, off-the-shelf gaming systems such as the Microsoft Kinect and PlayStation EyeToy [14-17]. Gaming controllers are intuitive to users, easy to repurpose, and more affordable, thereby offering a promising means for accessible home-based telerehabilitation. Gaming controllers can also measure motor performance objectively, toward remote assessment of patient status and progress by physicians [18].

    As an example, the Novint Falcon can detect subtle differences in the kinematics of healthy and affected individuals [19] and evaluate patients’ motor learning as their rehabilitation progresses [20], through measurements of mean speed, peak speed, and path length traversed. These metrics have been previously used and upheld in robotic telerehabilitation of the upper limb to rapidly assess physical effort and movement accuracy and smoothness [21,22]. Other metrics can be used to assess patients’ motor performance, including range of motion [8,23], coordination [24,25], and amount of force exerted [14,26].

    Adherence to Rehabilitation Regimen

    Although the mechanical framework of telerehabilitation has been successfully implemented in homes, patients often fail to comply with their home-based physical therapy, primarily because of the lack of motivation [27,28]. Acknowledging that sustained engagement is a prerequisite for successful outcomes, a large body of research has studied motivational interventions through game designs toward overcoming noncompliance in telerehabilitation [29-33]. For example, in an experimental study, Colombo et al [31] simulated a video game experience by displaying performance scores to improve motivation and adherence to the physical regimen. Similarly, in the study by Nijenhuis [29], a motivational messaging system was introduced to encourage future engagement in training sessions. Other studies have considered the use of serious games that do not aim primarily at entertainment for enhancing the physical rehabilitation experience [34-37]. For instance, Jonsdottir et al [38] have demonstrated the feasibility and efficacy of Rehab@Home, a therapeutic framework that simulates daily life activities in a virtual environment using Kinect. This gaming system was shown to increase gross motor function and improve patients’ experience and perception of health in patients with multiple sclerosis.

    Overall, these studies have demonstrated that gamification increases engagement in rehabilitation exercises [37,39]. Yet, the full capacity of supplemental motivational interventions remains largely untapped as designers rarely emphasize the users’ intellect and interest to maintain prolonged engagement. Particularly in the context of rehabilitation, the age group of the majority of patients may not be conducive to the use of typical computer games that target young gamers [40,41]. Aiming to address the differential motivations of the elderly, Flores et al [27] pinpointed the gaming design criteria catering to the needs of both young and elderly users, which include (1) consideration of decreased sensorimotor abilities, (2) cognitively challenging elements, and (3) some degree of socialization.

    Citizen Science and Telerehabilitation

    Following this line of work, we have previously demonstrated the potential utility of citizen science in increasing engagement and enjoyment in rehabilitation through the systematic interaction of environmental citizen science and robotics-based low-cost telerehabilitation technologies [19,42,43]. Citizen science projects address a wide range of scientific fields of inquiry. For example, on Stardust@Home, users can review images of an aerogel that was sent to outer space and flag traces of interstellar dust trapped in it [44]. In a different project, Foldit, volunteers fold virtual proteins and produce novel models of protein structures [45]. Citizen science could be a passive undertaking whereby citizen scientists lend computational power while their computers are idle [46,47]. The activities are not restricted to desktop computers and may also take place outdoors, where volunteers report of animal sightings or record air and water quality using their mobile phones [46,47].

    In general, in citizen science projects, members of the public execute scientific tasks in authentic research projects, led by professional scientists and otherwise [46,47]. The contribution of volunteers typically involves data collection or data analysis and does not require specific expertise [47,48]. The motivation ascribed to citizen science projects is majorly intrinsic as participation is intellectually stimulating and promotes learning [48,49]. Moreover, the completion of individual tasks requires a small time commitment, allowing users to contribute at their own pace. Therefore, citizen science inherently satisfies the criteria identified by Flores et al [27], with the exception of social interaction, which must be separately addressed through new design interventions.

    Social Interactions as a Motivational Driver

    Personal and social support, whether provided by practitioners, family, or friends, has been found to increase patients’ motivation toward performing exercise at home and to improve their mental well-being [50,51]. In the context of telerehabilitation, socially assistive robots, affording social interaction with patients while relaying treatments, were created [52]. The mere interaction with inanimate socially assistive robots was demonstrated to increase patients’ engagement in therapy and alleviate their feelings of stress and depression [52-54]. Building on this evidence, interhuman social interactions were introduced between the patient and their practitioner [52,55,56] and subsequently extended to include interactions with relatives and friends, and even strangers [30,52,57-59]. In all cases, patients expressed a strong preference to perform exercise with another person rather than alone and with a human partner rather than a virtual one [30,57,60]. Moreover, social interactions were demonstrated to improve motor performance [61]. Yet, the context in which social interactions were studied is largely limited to games.

    Whether social interactions could benefit or hamper the success of citizen science–based rehabilitation treatments remains elusive. It is known that social presence alone should enhance user engagement and prolong participation in Web-based platforms through social comparison [62-65]. However, working in a team may also lead to the opposite outcome, whereby users could reduce their participation in an activity because of diffusion of responsibility, a sociopsychological phenomenon observed when an individual is less likely to assume responsibility of action in the presence of other individuals [66]. Diffusion of responsibility is moderated by several factors, including anonymity [67], group size [68], and division of labor [69], which can all be found in citizen science [70]. As a result, it is difficult to predict whether including social elements in citizen science–based rehabilitation could produce the sought motivational factor advocated in the study by Flores et al [27] or, instead, produce an adverse social phenomenon through diffusion of responsibility.

    Objectives

    In this study, we sought to fill this gap in knowledge by examining the influence of computer-mediated cooperation on the engagement and motor performance of users involved in a rehabilitation exercise that integrated environmental citizen science and robotics-based technologies. We hypothesized that introducing cooperative tasks into citizen science would motivate users to extend their contribution by increasing the amount of scientific data they collect or analyze (productivity) and the time they spend performing the scientific task (persistence). This hypothesis rests on previous evidence that both of these measures are positively associated with motivation in goal-related activities [71]. In addition to increasing engagement, we expected that the integration of cooperation would improve users’ motor performance, reflected by their exertion of higher levels of physical effort. Finally, we hypothesized that the level of improvement in engagement and motor performance would be modulated by varying the degree of independence between paired users, whereby strengthening the interdependence between them would mitigate diffusion of responsibility.

    To test our hypotheses, we created a novel, dedicated interface for Brooklyn Atlantis—a local citizen science project for environmental monitoring of the highly polluted Gowanus canal, located in Brooklyn, New York [72]. Our system enabled users to analyze pictures of the canal taken by an aquatic robot using a low-cost haptic controller, whose potential use in rehabilitation treatments on patients has been previously demonstrated by our group and other researchers [19]. Using the system, a pair of volunteers was presented with a list of descriptive keywords that may describe the objects in images. The volunteers sorted a list of labels together where one user allocates labels that describe objects in the image while the other discards irrelevant labels. Here, we report results for the effect of using the platform and the collaborative procedure on healthy people.


    Hardware and Software

    All activities were performed using the Novint Falcon game controller, a low-cost haptic controller capable for use in 3 dimensions (Figure 1). The Novint Falcon offers translational hand movement with 3 degrees of freedom: left-right (x-axis), up-down (y-axis), and push-pull (z-axis). This device was demonstrated to provide effective fine-motor hand rehabilitation [73]. The system, developed using Unity 3D (Unity Technologies), displayed a 360° image of the Gowanus Canal on a computer screen (Figure 2). To explore the image, users pressed the middle button on the controller continuously and moved the controller in the general direction they wanted to rotate the view. A reproduction of the Novint Falcon interface was continuously shown on the screen, as a reference for the function of each button (Figure 2).

    Movement was implemented in spherical coordinates, whereby motion of the controller along the x-axis (Figure 1) translated into azimuthal rotation (turning right or left in Figure 2) and motion of the controller along the y-axis (Figure 1) translated into elevational rotation (turning up or down in Figure 2). As the motion along the push-pull axis did not convey a meaningful function (zoom was not offered), a highly resistive force was applied in this direction to prevent motion. In addition, visual feedback was added to the system, such that the image of the Gowanus Canal would fade in response to motion along the z-axis, either pushing or pulling. The deviation from the z-axis was further conveyed through a black circle and radiating cone, portraying where the user is facing and how far off the axis they are located (Figure 2).

    As part of the citizen science image classification project, 2 tasks were implemented in the system. The first task consisted of tagging objects observed in the images using labels from a list, located on the right of the 360° image (green panel in Figure 2). The second task entailed eliminating labels from the list that were not in the image by allocating them to the trash bin, located on the right of the list of labels (yellow panel in Figure 2). When 2 users performed the tasks together, each user had independent control over exploration of the 360° image. Social cues were conveyed through the system as cooperating users could view their partner’s actions in real time. That is, the user assigned with the tagging task could see a list of the eliminated labels forming below the trash bin (Figure 2). Similarly, the user assigned with the trashing task could see the labels assigned to the image (Figure 2). The presence of a peer was further made evident by highlighting a label in red, indicating that it was selected by the peer. An illustration of the setup is depicted in Figure 3.

    Users were able to select labels by pressing the right or left button on the controller, depending on their dexterity. Once a button was pressed, the label was tethered to the cursor and effectively dragged by it. To deselect a label, or to release it at a desired location, users pressed the controller button again. Once assigned, labels were replaced by others from a predetermined sequence of 49 labels, all of which were previously contributed by citizen scientists in Brooklyn Atlantis in our previous research [19,74-77]. To maintain fluidity of the image classification process and avoid oversaturation of images with tags, an image was replaced by another image after 5 tags were assigned. A tag counter was displayed on the screen to inform the user of the number of tags assigned to the current image (Figure 2).

    Users were able to terminate the activity by pressing the red Quit button on the screen (Figure 2). To simulate the different levels of interdependence in the cooperation between individuals, 2 types of termination were considered in the experiments: independent termination (IT), whereby users could continue contributing to the project even after their peer had quit, and joint termination (JT) whereby termination by 1 user ceased the session for the other user too.

    ‎
    Figure 1. The Novint Falcon with the designated axes of motion.
    View this figure
    ‎
    Figure 2. A screenshot of the user interface. On the left of the screenshot is a 360° image of the Gowanus canal. The user’s cursor is placing the label “Crane” onto the image, while a tag containing the word “Buoy” has already been placed. A reproduction of the Novint Falcon controller with a description of the function of each button is located on the upper right corner of the image. In the green panel, a counter of the number of labels that are yet to be assigned to the current image is displayed at the top. Below the counter, there is a list of 10 labels. The label “Crane” is highlighted in red as it is currently selected by the user. Below the list of labels is a visual feedback that represents deviation from the z-axis. A Quit button is situated at the bottom of the green panel. In the yellow panel, there is a garbage bin for eliminating labels that do not describe objects in the current image. The labels below it, “Robot” and “Person”, have been eliminated by the user.
    View this figure
    ‎
    Figure 3. Schematic of two cooperating users classifying images remotely from two different computers in separate rooms.
    View this figure

    Experimental Procedure

    This study was carried out in compliance with the institutional review board (IRB) at New York University (IRB FY2016-184). Overall, 120 members of the university community, with a mean age of 27.36 (SD 8.28) years, were recruited and subjected to one of 3 conditions (Table 1): IT (50 volunteers), JT (48 volunteers), and control (22 volunteers). Although control subjects performed both tagging and trashing, volunteers in IT and JT cooperatively carried out the activity such that half of the volunteers (25 and 24 volunteers in IT and JT, respectively) performed only tagging and the other half (25 and 24 volunteers in IT and JT, respectively) performed only trashing. In the control condition, the volunteer was able to withdraw from the activity at any time.

    Table 1. A summary of the experimental conditions tested.
    Condition and task assignmentCooperationNumber of volunteers
    Control

    		Tagging and trashingAbsent22
    Independent termination

    		TaggingPresent25

    		TrashingPresent25
    Joint termination

    		TaggingPresent24

    		TrashingPresent24

    Recruitment and experimental procedures were standardized through scripts and a PowerPoint presentation. We recruited volunteers in public spaces on campus. During recruitment, we verbally introduced potential participants on campus to the notion of citizen science following a script. Once recruited, paired volunteers were brought into 2 separate private rooms to simulate Web-based cooperation as envisioned in future application within robotics-based telerehabilitation. They did not know who their peer was.

    All participants were subjected to the same experimental protocol. Before beginning the experiment, participants were given an overview of the Gowanus Canal and Brooklyn Atlantis using a PowerPoint presentation. Through the presentation, within cooperative conditions, IT and JT, participants were notified that they will be working together with a peer and were instructed to complete their assigned task only. They were informed that they may withdraw at any point they would like and whether their withdrawal will terminate their peer’s participation (JT) or not (IT). Upon signing a consent form, the participants underwent a tutorial teaching them how to use the Novint Falcon and the system. After the tutorial, they were connected with their peer and began carrying out their tasks. Users who were subjected to the control condition carried out both tasks, tagging and trashing. Users who were subjected to cooperative conditions were randomly assigned to one of the 2 tasks. They performed the exercise until they pressed the quit button. After quitting, the participants rated their experience on a 7-point Likert scale in response to the statements “I enjoyed this activity” and “This activity was fun.” Once the volunteers submitted their answers, the experiment was concluded.

    Data Collection and Analysis

    Data Acquisition

    For each user, 3 datasets were created. The first dataset documented information on tag allocation, including tag content, time of allocation, and allocating user identity number. The second dataset recorded users’ scores of enjoyment. The third dataset consisted of the Novint Falcon controller position in 3D space, recorded at a sampling rate of 60 positions per second. The collected data were used to quantify user engagement, enjoyment, and motor performance.

    Data Processing

    User engagement was evaluated through their productivity and persistence [71]. Productivity was measured as the number of labels processed by the user. Persistence was measured as the time spent performing the activity [71]. Users’ enjoyment was evaluated from surveys. Interrater reliability was validated using the Cronbach alpha [78]. Enjoyment was scored by averaging the ratings on the multiple questions for each user, linearly scaling between 0 (Likert scale 1) and 1 (Likert scale 7), and normalizing using an arcsine transformation by considering the proportional nature of the variable [79].

    The trajectory of the controller was examined from the recording of consecutive points in space over time. A total of 3 motion metrics were evaluated from the trajectory, namely, the controller’s mean speed, peak speed, and path length. For each trial, the instantaneous speed was estimated using a backward Euler scheme on the sampled positions from the haptic device. The mean speed was computed by averaging instantaneous values over the whole trajectory and the peak speed as the maximum value among the 90th percentile from the trajectory [80]. The path length was measured as the sum of distances between pairs of consecutive data points.

    Statistical Analysis

    The influence of cooperation on engagement, enjoyment, and motor performance was investigated by fitting each variable into a generalized linear-mixed effects model [81], specifying condition (3 levels: control, IT, and JT) as an independent variable and both pair identity and task assignment (tagging and trashing) as random effects (R lme4 package version 1.1-15 [82]). To improve the normality of the model residual, we specified a Gaussian family with a log link for persistence and enjoyment, a Poisson family with a log link for productivity, and a gamma family with a log link for motor performance. The significance of the influence of conditions was tested using a likelihood ratio test, comparing the model against a null model in the absence of the condition as the independent variable. When significant effect was found, post hoc analysis was performed using the Dunnett test (R multcomp package version 1.4-8 [83]).

    Next, we evaluated the influence of the modality through which cooperation was implemented on engagement, enjoyment, and motor performance. Specifically, we fitted each variable into a generalized linear mixed-effects model, specifying condition (2 levels: IT and JT), task assignment (2 levels: tagging and trashing), and the interaction between them as independent variables, and pair identity as a random effect. The same error family as the previous model was used for the corresponding variable. To test the significance of the interaction term, the full model was tested against a null model without the interaction using a likelihood ratio test. In case a significant interaction was found, the difference between tasks was further examined within each condition, by specifying task as an independent variable and pair identity as a random effect. In case the interaction was not significant, we removed the interaction from the full model, and the effects of condition and task were tested using a likelihood ratio test, individually, comparing against a null model.

    Although not part of our original hypotheses, we also tested the influence of social presence on individual speed performance. In each pair of cooperating peers in IT, the trajectory of the more persistence was partitioned into 2 parts, before and after their peer had quit. Users’ mean and peak speeds were fitted into separate generalized linear mixed-effects models, specifying the time partition (2 levels: before and after) as an independent variable and user identity as a random effect. A gamma family with log link was specified to normalize the model residual. Users’ speeds before peer withdrawal were compared with their speeds following peer withdrawal using a likelihood ratio test, comparing the model against a null model in the absence of the time partition as the independent variable.

    For all statistical tests, we set the level of significance at alpha=.05.


    Influence of Cooperation on Engagement

    On average, users processed (tagged or trashed) a mean of 46.35 (SD 3.16) labels, spending 16.37 (SD 0.63) min. Neither productivity nor persistence were found to differ among conditions (χ22=0.1 P=.92 and χ22<0.1; P=.79, respectively; Figure 4). However, the level of enjoyment was found to vary among conditions (χ22=10.5; P=.005; Figure 4), with JT users rating the activity significantly lower than control users (z=3.25; P=.002). By contrast, IT users did not rate the activity significantly different from control users (z=1.94; P=.08).

    ‎
    Figure 4. Engagement of users in the activity. A) number of labels processed by participants in each condition. B) rate of enjoyment for each condition. The vertical lines represent standard errors. *: statistically different means among conditions. $: statistically different means among tasks.
    View this figure

    Influence of Cooperation on Motor Performance

    With regard to motor performance, the mean speed did not differ among conditions (χ22=2.6; P=.26). Contrarily, we determined a significant variation of peak speed among conditions (χ22=7.7; P=.02; Figure 5). Although post hoc comparisons failed to identify a significant difference between JT and control conditions (z=0.11; P=.99), we registered a significant difference between IT users and control users (z=2.44; P=.02). The path length was not significantly different among conditions (χ22=4.1; P=.12).

    Influence of Modality on Engagement

    Testing for the influence of cooperation modality on productivity, we found a significant interaction between condition and task (χ22=43.1; P<.001). When investigating the effect of task assignment in each condition, we found a significant difference in productivity between the tasks in both conditions, IT (χ21=265.7; P<.001) and JT (χ21=25.5; P<.001). We failed to identify a significant interaction between condition and task with regard to persistence (χ21=3.3; P=.06). An interaction between condition and task was not found in enjoyment as well (χ21=1.1; P=.29). Enjoyment was significantly different between the tasks (χ21=7.7; P=.005), whereas condition failed to reach significance (χ21=2.8; P=.09).

    Influence of Modality on Motor Performance

    With regard to the performance metrics, a significant interaction between condition and task was found to influence path length (χ21=6.3; P=.01), mean speed (χ21=7.4; P=.006), and peak speed (χ21=25.8; P<.001; Figure 5). For path length, we found a significant difference between task assignment in IT condition (χ21=11.3; P<.001), whereas we did not find one in JT condition (χ21<0.1; P=.90). For mean speed, a significant difference between tasks was found in IT condition (χ21=6.2; P=.01) but not in JT condition (χ21=2.4; P=.12). Finally, for peak speed, a significant difference was found between tasks in both IT and JT conditions (χ21=16.1; P<.001 and χ21=11.0; P<.001, respectively).

    In IT, volunteers significantly reduced their mean and peak speeds following their peer’s quitting (z=2.97; P=.002 and z=3.30; P<.001, respectively; Figure 6).

    ‎
    Figure 5. Motor metrics. A) mean speed in each condition, B) peak speed for each condition, C) path length traversed by the controller in each conditions. The vertical lines represent standard errors. * represents statistically different means among conditions. $ represents statistically different means among tasks.
    View this figure
    ‎
    Figure 6. Differences in mean and peak speeds of the more persistent users in condition IT, before and after their peer has withdrawn. The vertical lines represent standard errors. * represents statistically different means among conditions.
    View this figure

    Principal Findings

    Citizen science is an effective means for improving rehabilitation treatments. Patients undergoing physical rehabilitation have shown a strong preference toward exercise embedded with citizen science and were more likely to repeat it at the cost of their time commitment [19]. Although social interactions hold potential to further increase patients’ engagement in rehabilitation [30,57], the modality in which they are framed could widely shape the outcomes of the treatment [84]. In this study, we attempted to elucidate the influence of computer-mediated cooperation on motor performance during a citizen science activity, mediated by a low-cost haptic device.

    We designed a series of experiments simulating an authentic telerehabilitation setting, where participants remotely cooperated in the analysis of environmental images using low-cost haptic devices. From survey instruments and direct measurements of motor activities, we sought to quantify the potential effect of cooperation in telerehabilitation. Our results indicate that 2 elements, interdependence and task assignment, can influence the effects of cooperation on engagement and motor performance, thereby offering a potential means for improving rehabilitation treatments. However, the way in which these 2 variables interact to influence the response of the subjects may challenge one’s intuition.

    In partial contrast with our hypotheses, we did not find that cooperative division of labor, such that each user is assigned to a different task toward a shared goal [85], is always conducive to higher engagement. Although we did not register a difference in the level of engagement between control users and users who cooperated via IT, we found an expected reduction in enjoyment for users who cooperated via JT. In our experimental design, JT was implemented to promote interdependence between users, which we had initially identified as a key factor to mitigate diffusion of responsibility and coerce users to persist for a longer period of time [86,87]. However, it is likely that JT was accompanied by other confounding factors, contributing to a reduction in enjoyment.

    It is tenable that we inadvertently introduced random termination in the trials where users faced uncertainty with regard to the timing of termination, as the other player could terminate the task at any time. This uncertainty about the horizon of the relationship with the other player was shown in previous research to impact the degree of cooperativeness of players negatively [88,89], where the more termination becomes likely, the less cooperation has been observed [90]. In game theory, random termination has been shown to discount the payoff of players’ actions [88,89] such that players would try to avoid losses and become less cooperative. Such weakening of cooperation also results in lower levels of enjoyment and satisfaction [91]. Similar dynamics likely emerged in the proposed citizen science–based telerehabilitation activity, thereby calling for future research to explore alternative strategies that could promote interdependence between users. For instance, we could attempt at a priori identifying a predetermined length for the trials, by matching patients undergoing a similar rehabilitation therapy.

    Our findings also provided insight into the role of social interactions on motor performance. Measuring relevant kinematic variables is central to the notion of telerehabilitation, whereby supplying care providers with clinical information will enable them to track patients’ status and adjust their rehabilitation program remotely and efficiently [18]. Ideally, care providers could also infer abnormal, compensatory movement from the data and instruct patients to correct it [18]. Mean speed, peak speed, and path length traversed [19,92] have been used in robotic telerehabilitation of the upper limb as indicators of motion quality [14,21,22]. In addition to the evaluation of motor performance, the physical effort exerted by an individual can also be linked to their motivation to perform the exercise task. Investing greater effort to complete a challenging task often leads to self-determined behavior, resulting in a sense of competence [93] and an increase in intrinsic motivation [94].

    Although we expected that users would improve their motor performance because of cooperation, we found a modest reduction in motor performance similar to the discussed reduction in enjoyment. More specifically, motor performance of users cooperating via JT was similar to control with regard for all the selected metrics, whereas IT resulted in higher values of the peak speed relative to control. It is possible that random termination could explain the observed difference, where a user would not invest the same effort when faced with the potential that his/her work could be vanished because of exogenous termination of the activity by the peer. Future research should seek to explore alternative modalities to favor cooperation, without challenging enjoyment and effort that are key to the success of rehabilitation. The notion of setting intervals for the exercises could be a viable approach, whereby it could mitigate the harmful effects of random termination, while leveraging the beneficial role of cooperation. In fact, analyzing time variations of motor performance of users who cooperated via IT, we discovered that the speed of the more persistent users in IT significantly reduced following their peer’s withdrawal. This confirms the intuition that social interaction should prompt individuals to exert more effort in their task, thereby calling for future studies to engineer social interactions toward improving recovery and patients’ self-perception of physical capacity [95,96].

    In addition to the type of termination, we found that task assignment can modulate user engagement. We found that cooperating users in IT condition assigned with tagging were more engaged than their peers who were assigned with trashing. The difference in engagement may be attributed to the perceived nature of the task a user has been assigned to. In fact, engagement is positively associated with the identifiability of the share an individual contributes to the groupwork [87,97], that is, individuals whose contribution is more valued and recognized by group peers are more likely to be motivated to perform their task. Conversely, individuals whose contribution is less important and recognized by others in the group are less motivated to perform their task. Therefore, the dissimilarity between tagging and trashing tasks could lead to unequal levels of engagement, with tagging users being more engaged than trashing users. This observation calls for further studies in which targeted design interventions will be explored to investigate differences between cooperation and collaboration, where individuals are assigned to the same task. It is tenable that cooperating individuals should perform better when assigned with a common group task rather than interdependent tasks [84]. Working in a collaborative setting where they fulfill identical functions jointly in support of a shared goal [85] could bolster team cohesion and lead to even higher performance and satisfaction among users [98].

    Evidence shows that group cohesiveness can be improved with increasingly overt sharing of information, leading to greater engagement in group tasks [99]. Recently, we showed that the mere presentation of social foot prints, digital cues that suggest the presence of other Web users can be used to increase the amount and duration of physical exercise during citizen science activities [64]. In a different study, using a virtual peer operating in open- and closed-loop paradigms, we demonstrated that bidirectional flow of social information can substantially increase user contribution to a citizen science project [75]. Seemingly, as more social presence is conveyed between Web users, an increasingly trusting climate is created, conducive to cooperation [100]. In future studies, one could explore how sharing of personal information such as age, location, and interests can impact trust and engagement in Web-based citizen science telerehabilitation [101].

    Citizen science contributes to a sense of community in cooperative telerehabilitation through the introduction of scientific content. Unlike gaming-based motivational interventions, this study capitalizes on human intellect as an intrinsic motivator. Previously, we had shown that users prefer to perform an exercise associated with scientific content [19,92]. Beyond the intellectual stimulus, citizen science adds virtue and a sense of contribution to the activity, which is not found in the majority of serious games. Web-based social platforms where individuals share personal values often create communities with which contributors can identify [48]. In telerehabilitation in particular, patients can benefit from such an environment that could alleviate the isolation many of them experience [102,103].

    Limitations

    Although our work brings forward evidence in favor of the use cooperative citizen science in rehabilitation, it comes with a number of limitations. First, the difference in engagement that we observed among conditions was moderate. It is possible that because citizen science is inherently engaging [92], additional motivational interventions such as social interactions offer only a weak enhancement to engagement, limited by a ceiling effect. To surpass such a ceiling effect, it would be beneficial to explore the role of cooperation in citizen science–based rehabilitation in a longitudinal study, where persistence is measured by the frequency patients choose to engage in exercise and productivity is measured as aggregated contribution.

    Second, we studied motor performance using the Novint Falcon, a haptic device which is no longer being produced. However, the fine-motor tasks imparted by the Novint Falcon can be achieved using other haptic devices. For instance, surgical delta robots such as the Force Dimension and Phantom offer movement with 6 degrees of freedom and can apply a comparable amount of force feedback [104-106].

    Third, in this study, we recruited healthy subjects from New York University campuses. Our findings may be narrowly generalizable as the sample consists of healthy individuals from the Brooklyn area with access to high education. Although we drew our sample from different programs of the university, the volunteers may have distinctly different interests and motivations from the typical patient undergoing rehabilitation. For example, the participants in this study may have greater interest in science or in restoration of the Gowanus Canal than the average person. Offering a wider range of citizen science projects to choose from based on personal interests might further improve enjoyment, engagement, and motor performance. Future research with patients from diverse backgrounds in a clinical setting will elucidate effects of this work on clinical outcomes.

    Conclusions

    We offer evidence for the utility of cooperation in improving engagement in citizen science–based telerehabilitation. Citizen science can offer intellectual stimulus and a community for patients to engage with and relate to. It attends the needs of more patients, including those who are less interested in traditional gaming [27,37], thereby extending the benefits of adherence to home-based physical therapy to a larger population. The value of this study can be expanded to other domains that rely on user participation and engagement, including Web-based consumer platforms [107], social networks [108], crowdsourcing efforts [109], and general game design [91]. Ultimately, we anticipate our approach will be translated into low-cost technology for telerehabilitation and help patients reach their full potential recovery.

    Acknowledgments

    This study was supported by the National Science Foundation under award number CBET-1604355. RBV would like to acknowledge the Mitsui USA Foundation for their generous support. All the authors would like to thank M Nadini for helping conduct experiments and S Richmond and TJ Tolbert for software development.

    Authors' Contributions

    RBV, ON, and MP conceived and designed the experiments. RBV performed the experiments. RBV, SN ON, and MP analyzed the data. RBV, SN, ON, and MP wrote the manuscript. RBV, SN, PR, ON, and MP reviewed the manuscript.

    Conflicts of Interest

    None declared.

    1. Bütefisch C, Hummelsheim H, Denzler P, Mauritz KH. Repetitive training of isolated movements improves the outcome of motor rehabilitation of the centrally paretic hand. J Neurol Sci 1995 May;130(1):59-68. [CrossRef] [Medline]
    2. Harris JE, Eng JJ, Miller WC, Dawson AS. A self-administered Graded Repetitive Arm Supplementary Program (GRASP) improves arm function during inpatient stroke rehabilitation: a multi-site randomized controlled trial. Stroke 2009 Jun;40(6):2123-2128. [CrossRef] [Medline]
    3. Volpe BT, Lynch D, Rykman-Berland A, Ferraro M, Galgano M, Hogan N, et al. Intensive sensorimotor arm training mediated by therapist or robot improves hemiparesis in patients with chronic stroke. Neurorehabil Neural Repair 2008;22(3):305-310 [FREE Full text] [CrossRef] [Medline]
    4. Kroll T, Jones GC, Kehn M, Neri MT. Barriers and strategies affecting the utilisation of primary preventive services for people with physical disabilities: a qualitative inquiry. Health Soc Care Community 2006 Jul;14(4):284-293. [CrossRef] [Medline]
    5. Krebs HI, Ferraro M, Buerger SP, Newbery MJ, Makiyama A, Sandmann M, et al. Rehabilitation robotics: pilot trial of a spatial extension for MIT-Manus. J Neuroeng Rehabil 2004 Oct 26;1(1):5 [FREE Full text] [CrossRef] [Medline]
    6. Fasoli SE, Krebs HI, Stein J, Frontera WR, Hogan N. Effects of robotic therapy on motor impairment and recovery in chronic stroke. Arch Phys Med Rehabil 2003 Apr;84(4):477-482. [CrossRef] [Medline]
    7. Coote S, Murphy B, Harwin W, Stokes E. The effect of the GENTLE/s robot-mediated therapy system on arm function after stroke. Clin Rehabil 2008 May;22(5):395-405. [CrossRef] [Medline]
    8. Nef T, Mihelj M, Colombo G, Riener R. ARMin - Robot for rehabilitation of the upper extremities. In: Proceedings of the 2006 IEEE International Conference on Robotics and Automation. 2006 Presented at: IEEE International Conference on Robotics and Automation; May 15-19, 2006; Orlando, Florida, USA. [CrossRef]
    9. Lum PS, Burgar CG, Loos MV, Shor PC, Majmundar M, Yap R. MIME robotic device for upper-limb neurorehabilitation in subacute stroke subjects: a follow-up study. J Rehabil Res Dev 2006;43(5):631-642. [CrossRef] [Medline]
    10. Laut J, Porfiri M, Raghavan P. The present and future of robotic technology in rehabilitation. Curr Phys Med Rehabil Rep 2016 Dec;4(4):312-319 [FREE Full text] [CrossRef] [Medline]
    11. Krebs H, Dipietro L, Levy-Tzedek S, Fasoli S, Rykman-Berland A, Zipse J, et al. A paradigm shift for rehabilitation robotics. IEEE Eng Med Biol Mag 2008 Jul;27(4):61-70 [FREE Full text] [CrossRef] [Medline]
    12. Romer GR, Stuyt HJ, Peters A. Cost-savings and economic benefits due to the Assistive Robotic Manipulator (ARM). 2005 Presented at: International Conference on Rehabilitation Robotics; 28 June-1 July, 2005; Chicago, Illinois, USA. [CrossRef]
    13. Wootton R, Hebert MA. What constitutes success in telehealth? J Telemed Telecare 2001;7(Suppl 2):3-7. [CrossRef] [Medline]
    14. Reinkensmeyer DJ, Pang CT, Nessler JA, Painter CC. Web-based telerehabilitation for the upper extremity after stroke. IEEE Trans Neural Syst Rehabil Eng 2002 Jun;10(2):102-108. [CrossRef] [Medline]
    15. Lange B, Chang C, Suma E, Newman B, Rizzo AS, Bolas M. Development and evaluation of low cost game-based balance rehabilitation tool using the Microsoft Kinect sensor. Conf Proc IEEE Eng Med Biol Soc 2011;2011:1831-1834. [CrossRef] [Medline]
    16. Weiss PL, Rand D, Katz N, Kizony R. Video capture virtual reality as a flexible and effective rehabilitation tool. J Neuroeng Rehabil 2004 Dec 20;1(1):12 [FREE Full text] [CrossRef] [Medline]
    17. Rand D, Kizony R, Weiss PTL. The Sony PlayStation II EyeToy: low-cost virtual reality for use in rehabilitation. J Neurol Phys Ther 2008 Dec;32(4):155-163. [CrossRef] [Medline]
    18. Perry JC, Ruiz-Ruano JA, Keller T. Telerehabilitation: toward a cost-efficient platform for post-stroke neurorehabilitation. IEEE Int Conf Rehabil Robot 2011;2011:5975413. [CrossRef] [Medline]
    19. Laut J, Cappa F, Nov O, Porfiri M. Increasing patient engagement in rehabilitation exercises using computer-based citizen science. PLoS One 2015;10(3):e0117013 [FREE Full text] [CrossRef] [Medline]
    20. Scalona E, Martelli F, Del PZ, Palermo E, Rossi S. A novel protocol for the evaluation of motor learning in 3D reching tasks using Novint Falcon. 2018 Presented at: IEEE RAS/EMBS International Conference on Biomedical Robotics and Biomechatronics; August 27-29, 2018; Enschede, Netherlands. [CrossRef]
    21. Bosecker C, Dipietro L, Volpe B, Krebs HI. Kinematic robot-based evaluation scales and clinical counterparts to measure upper limb motor performance in patients with chronic stroke. Neurorehabil Neural Repair 2010 Jan;24(1):62-69 [FREE Full text] [CrossRef] [Medline]
    22. Zollo L, Rossini L, Bravi M, Magrone G, Sterzi S, Guglielmelli E. Quantitative evaluation of upper-limb motor control in robot-aided rehabilitation. Med Biol Eng Comput 2011 Oct;49(10):1131-1144. [CrossRef] [Medline]
    23. Colombo R, Pisano F, Micera S, Mazzone A, Delconte C, Carrozza MC, et al. Robotic techniques for upper limb evaluation and rehabilitation of stroke patients. IEEE Trans Neural Syst Rehabil Eng 2005 Sep;13(3):311-324. [CrossRef] [Medline]
    24. Raghavan P, Santello M, Gordon AM, Krakauer JW. Compensatory motor control after stroke: an alternative joint strategy for object-dependent shaping of hand posture. J Neurophysiol 2010 Jun;103(6):3034-3043 [FREE Full text] [CrossRef] [Medline]
    25. Squeri V, Zenzeri J, Morasso P, Basteris A. Integrating proprioceptive assessment with proprioceptive training of stroke patients. IEEE Int Conf Rehabil Robot 2011;2011:5975500. [CrossRef] [Medline]
    26. Boissy P, Bourbonnais D, Carlotti MM, Gravel D, Arsenault BA. Maximal grip force in chronic stroke subjects and its relationship to global upper extremity function. Clin Rehabil 1999 Aug;13(4):354-362. [CrossRef] [Medline]
    27. Flores E, Tobon G, Cavallaro E, Cavallaro F, Perry J, Keller T. Improving patient motivation in game development for motor deficit rehabilitation. 2008 Presented at: International Conference on Advances in Computer Entertainment Technology; December 3-5, 2008; Yokohama, Japan p. 381-384. [CrossRef]
    28. Conraads VM, Deaton C, Piotrowicz E, Santaularia N, Tierney S, Piepoli MF, et al. Adherence of heart failure patients to exercise: barriers and possible solutions: a position statement of the Study Group on Exercise Training in Heart Failure of the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail 2012 May;14(5):451-458 [FREE Full text] [CrossRef] [Medline]
    29. Nijenhuis SM, Prange GB, Amirabdollahian F, Sale P, Infarinato F, Nasr N, et al. Feasibility study into self-administered training at home using an arm and hand device with motivational gaming environment in chronic stroke. J Neuroeng Rehabil 2015 Oct 09;12:89 [FREE Full text] [CrossRef] [Medline]
    30. Novak D, Nagle A, Keller U, Riener R. Increasing motivation in robot-aided arm rehabilitation with competitive and cooperative gameplay. J Neuroeng Rehabil 2014 Apr 16;11:64 [FREE Full text] [CrossRef] [Medline]
    31. Colombo R, Pisano F, Mazzone A, Delconte C, Micera S, Carrozza MC, et al. Design strategies to improve patient motivation during robot-aided rehabilitation. J Neuroeng Rehabil 2007 Feb 19;4:3 [FREE Full text] [CrossRef] [Medline]
    32. Zimmerli L, Jacky M, Lünenburger L, Riener R, Bolliger M. Increasing patient engagement during virtual reality-based motor rehabilitation. Arch Phys Med Rehabil 2013 Sep;94(9):1737-1746. [CrossRef] [Medline]
    33. Vong SK, Cheing GL, Chan F, So EM, Chan CC. Motivational enhancement therapy in addition to physical therapy improves motivational factors and treatment outcomes in people with low back pain: a randomized controlled trial. Arch Phys Med Rehabil 2011 Feb;92(2):176-183. [CrossRef] [Medline]
    34. Laver KE, Lange B, George S, Deutsch JE, Saposnik G, Crotty M. Virtual reality for stroke rehabilitation. Cochrane Database Syst Rev 2017 Dec 20;11:CD008349. [CrossRef] [Medline]
    35. Taylor MJ, Griffin M. The use of gaming technology for rehabilitation in people with multiple sclerosis. Mult Scler 2015 Apr;21(4):355-371. [CrossRef] [Medline]
    36. Burke JW, McNeill M, Charles D, Morrow P, Crosbie J, McDonough S. Serious games for upper limb rehabilitation following stroke. : IEEE; 2009 Presented at: Conference in Games and Virtual Worlds for Serious Applications; March 23-24, 2009; Coventry, United Kingdom p. 103-110. [CrossRef]
    37. Rego P, Moreira P, Reis L. Serious games for rehabilitation: a survey and a classification towards a taxonomy. : IEEE; 2010 Presented at: Iberian Conference on Information Systems and Technologies; June 16-19, 2010; Santiago de Compostela, Spain   URL: https://ieeexplore.ieee.org/document/5556674
    38. Jonsdottir J, Bertoni R, Lawo M, Montesano A, Bowman T, Gabrielli S. Serious games for arm rehabilitation of persons with multiple sclerosis. A randomized controlled pilot study. Mult Scler Relat Disord 2018 Jan;19:25-29. [CrossRef] [Medline]
    39. Burke JW, McNeill MD, Charles DK, Morrow PJ, Crosbie JH, McDonough SM. Optimising engagement for stroke rehabilitation using serious games. Vis Comput 2009 Aug 27;25(12):1085-1099. [CrossRef]
    40. Awad M, Ferguson S, Craig C. Designing games for older adults: an affordance based approach. : IEEE; 2014 Presented at: IEEE International Conference on Serious Games and Applications for Health (SeGAH); May 14-16, 2014; Rio de Janeiro, Brazil. [CrossRef]
    41. Garcia MJ, Felix NK, Lawrence E. Serious games to improve the physical health of the elderly: a categorization scheme. CENTRIC 2011; 2011 Presented at: The International Conference on Advances in Human-oriented and Personalized Mechanisms, Technologies, and Services (CENTRIC); October 23-28, 2011; Barcelona, Spain p. 64-71   URL: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.454.9722&rep=rep1&type=pdf
    42. Palermo E, Laut J, Nov O, Cappa P, Porfiri M. A natural user interface to integrate citizen science and physical exercise. PLoS One 2017;12(2):e0172587 [FREE Full text] [CrossRef] [Medline]
    43. Laut J, Henry E, Nov O, Porfiri M. Development of a mechatronics-based citizen science platform for aquatic environmental monitoring. IEEE/ASME Trans. Mechatron 2014 Oct;19(5):1541-1551. [CrossRef]
    44. Westphal A, Butterworth A, Snead C, Craig N, Anderson D, Jones S, et al. National Aeronautics and Space Administration. 2005. Stardust@ home: a massively distributed public search for interstellar dust in the stardust interstellar dust collector   URL: https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20050180792.pdf [accessed 2018-11-06] [WebCite Cache]
    45. Cooper S, Khatib F, Treuille A, Barbero J, Lee J, Beenen M, et al. Predicting protein structures with a multiplayer online game. Nature 2010 Aug 05;466(7307):756-760 [FREE Full text] [CrossRef] [Medline]
    46. Silvertown J. A new dawn for citizen science. Trends Ecol Evol 2009 Sep;24(9):467-471. [CrossRef] [Medline]
    47. Haklay M. Citizen Science and Policy: A European Perspective. Washington, DC: The Woodrow Wilson Center, Commons Lab; 2015.
    48. Nov O, Arazy O, Anderson D. Dusting for science: motivation and participation of digital citizen science volunteers. 2011 Presented at: iConference 2011; February 08-11, 2011; Seattle, Washington, USA p. 68-74. [CrossRef]
    49. Nov O, Arazy O, Anderson D. Crowdsourcing for science: understanding and enhancing SciSourcing contribution. 2010 Presented at: 2010 Workshop on the Changing Dynamics of Scientific Collaborations; February 7, 2010; Savannah, Georgia, USA   URL: http://faculty.poly.edu/~onov/Nov%20Arazy%20Anderson%20CSCW%202010%20workshop.pdf
    50. Lewinter M, Mikkelsen S. Patients' experience of rehabilitation after stroke. Disabil Rehabil 2009 Jul 28;17(1):3-9. [CrossRef]
    51. Friedland J, McColl M. Social support intervention after stroke: results of a randomized trial. Arch Phys Med Rehabil 1992 Jun;73(6):573-581. [Medline]
    52. Tapus A, Ţăpuş C, Matarić MJ. User—robot personality matching and assistive robot behavior adaptation for post-stroke rehabilitation therapy. Intel Serv Robotics 2008 Feb 13;1(2):169-183. [CrossRef]
    53. Tapus A, Mataric M, Scassellati B. Socially assistive robotics [Grand Challenges of Robotics]. IEEE Robotics and Automation 2007 Mar;14(1):35-42. [CrossRef]
    54. Matarić MJ, Eriksson J, Feil-Seifer DJ, Winstein CJ. Socially assistive robotics for post-stroke rehabilitation. J Neuroeng Rehabil 2007 Feb 19;4:5 [FREE Full text] [CrossRef] [Medline]
    55. Johnson MJ, Feng X, Johnson LM, Ramachandran B, Winters JM, Kosasih JB. Robotic systems that rehabilitate as well as motivate: three strategies for motivating impaired arm use. : IEEE; 2006 Presented at: IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics; February 20-22, 2006; Pisa, Italy. [CrossRef]
    56. Carignan CR, Krebs HI. Telerehabilitation robotics: bright lights, big future? J Rehabil Res Dev 2006;43(5):695-710. [CrossRef]
    57. Gorsic M, Novak D. Design and pilot evaluation of competitive and cooperative exercise games for arm rehabilitation at home. : IEEE; 2016 Dec Presented at: IEEE Engineering in Medicine and Biology Society (EMBC); August 16-20, 2016; Orlando, Florida, USA p. 4690-4694. [CrossRef]
    58. Baur K, Schättin A, de Bruin ED, Riener R, Duarte JE, Wolf P. Trends in robot-assisted and virtual reality-assisted neuromuscular therapy: a systematic review of health-related multiplayer games. J Neuroeng Rehabil 2018 Nov 19;15(1):107 [FREE Full text] [CrossRef] [Medline]
    59. Dithmer M, Rasmussen JO, Grönvall E, Spindler H, Hansen J, Nielsen G, et al. "The Heart Game": using gamification as part of a telerehabilitation program for heart patients. Games Health J 2016 Feb;5(1):27-33 [FREE Full text] [CrossRef] [Medline]
    60. Weibel D, Wissmath B, Habegger S, Steiner Y, Groner R. Playing online games against computer- vs human-controlled opponents: effects on presence, flow, and enjoyment. Comput Human Behav 2008 Sep;24(5):2274-2291. [CrossRef]
    61. Johnson MJ, Loureiro RC, Harwin WS. Collaborative tele-rehabilitation and robot-mediated therapy for stroke rehabilitation at home or clinic. Intel Serv Robotics 2008 Jan 23;1(2):109-121. [CrossRef]
    62. Shen J. Social comparison, social presence, and enjoyment in the acceptance of social shopping websites. Journal Electron Commer Re 2012;13(3):198-212 [FREE Full text]
    63. Festinger L. A theory of social comparison processes. Hum Relat 2016 Apr 22;7(2):117-140. [CrossRef]
    64. Wu J, Wang S, Tsai H. Falling in love with online games: the uses and gratifications perspective. Comput Human Behav 2010;26(6):1862-1871. [CrossRef]
    65. Nakayama S, Tolbert T, Nov O, Porfiri M. Social information as a means to enhance engagement in citizen science‐based telerehabilitation. J Assoc Inf Sci Technol 2018 Dec 23;70(6):587-595. [CrossRef]
    66. Darley JM, Latané B. Bystander intervention in emergencies: diffusion of responsibility. J Pers Soc Psychol 1968 Apr;8(4):377-383. [Medline]
    67. Schwartz SH, Gottlieb A. Bystander anonymity and reactions to emergencies. J Pers Soc Psychol 1980 Sep;39(3):418-430. [Medline]
    68. Freeman S, Walker MR, Borden R, Latané B. Diffusion of responsibility and restaurant tipping: cheaper by the bunch. Pers Soc Psychol Bull 1975;1(4):584-587. [CrossRef]
    69. Kelman HG. Violence without moral restraint: reflections on the dehumanization of victims and victimizers. J Soc Issues 1973;29(4):25-61. [CrossRef]
    70. Nov O, Arazy O, Anderson D. Technology-mediated citizen science participation: a motivational model. 2011 Presented at: International Conference on Weblogs and Social Media; July 17-21, 2011; Barcelona, Catalonia, Spain   URL: http://faculty.poly.edu/~onov/Nov_Arazy_Anderson_Citizen_Science_ICWSM_2011.pdf
    71. Touré-Tillery M, Fishbach A. How to measure motivation: a guide for the experimental social psychologist. Soc Personal Psychol Compass 2014 Jul 02;8(7):328-341. [CrossRef]
    72. Laut J, Porfiri M, Nov O. Brooklyn Atlantis: a robotic platform for environmental monitoring with community participation. 2014 Presented at: International Conference of Control, Dynamic Systems, and Robotics; May 15-16, 2014; Ottawa, Ontario, Canada p. 961-968   URL: https://avestia.com/CDSR2014_Proceedings/papers/96.pdf
    73. Lange B, Flynn S, Rizzo A. Initial usability assessment of off-the-shelf video game consoles for clinical game-based motor rehabilitation. Phys Ther Rev 2013 Jul 19;14(5):355-363. [CrossRef]
    74. Laut J, Cappa F, Nov O, Porfiri M. Increasing citizen science contribution using a virtual peer. J Assoc Inf Sci Technol 2016 Aug 16;68(3):583-593. [CrossRef]
    75. Nov O, Laut J, Porfiri M. Using targeted design interventions to encourage extra-role crowdsourcing behavior. J Assn Inf Sci Tec 2015 May 05;67(2):483-489. [CrossRef]
    76. Laut J, Cappa F, Nov O, Porfiri M. A model for citizen scientist contribution in an image tagging task. 2016 Presented at: Dynamic Systems and Control Conference; October 12-14, 2016; Minneapolis, Minnesota, USA. [CrossRef]
    77. Nakayama S, Diner D, Holland JG, Bloch G, Porfiri M, Nov O. The influence of social information and self-expertise on emergent task allocation in virtual groups. Front Ecol Evol 2018 Feb 21;6:-. [CrossRef]
    78. Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika 1951 Sep;16(3):297-334. [CrossRef]
    79. Warton DI, Hui FK. The arcsine is asinine: the analysis of proportions in ecology. Ecology 2011 Jan;92(1):3-10. [Medline]
    80. Kaushik K, Sharifi E, Young S. Computing performance measures with national performance management research data set. Transport Res Rec 2015 Nov;2529:10-26. [CrossRef]
    81. Zuur A, Ieno E, Walker N, Saveliev A, Smith G. Mixed Effects Models and Extensions in Ecology With R. New York, New York, USA: Springer; 2019.
    82. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Soft 2015;67(1):-. [CrossRef]
    83. Bretz F, Hothorn T, Westfall P. Multiple Comparisons Using R. Boca Raton, Florida, USA: Chapman And Hall/CRC; 2019.
    84. Jones SC, Vroom VH. Division of labor and performance under cooperative and competitive conditions. J Abnorm Soc Psychol 1964;68(3):313-320. [CrossRef]
    85. Roschelle J, Teasley S. The construction of shared knowledge in collaborative problem solving. Comput Support Collab Learn 1995:69-97 [FREE Full text] [CrossRef]
    86. Johnson D, Johnson R, Holubec E. The New Circles of Learning: Cooperation in the Classroom and School. Minnesota, USA: Association for Supervision & Curriculum Development; 1994.
    87. Wagner JA. Studies of individualism-collectivism: effects on cooperation in groups. Acad Manag J 1995 Feb 01;38(1):152-173. [CrossRef]
    88. Pevnitskaya S, Ryvkin D. Environmental context and termination uncertainty in games with a dynamic public bad. Envir Dev Econ 2012 Nov 27;18(01):27-49. [CrossRef]
    89. Camerer C. Behavioral Game Theory: Experiments in Strategic Interaction. Princeton, New Jersey, USA: Princeton University Press; 2019.
    90. Bó PD. Cooperation under the shadow of the future: experimental evidence from infinitely repeated games. Am Econ Rev 2005 Nov;95(5):1591-1604. [CrossRef]
    91. Sweetser P, Wyeth P. GameFlow: a model for evaluating player enjoyment in games. Comput Entertain 2005 Jul 01;3(3):1-24. [CrossRef]
    92. Cappa P, Clerico A, Nov O, Porfiri M. Can force feedback and science learning enhance the effectiveness of neuro-rehabilitation? An experimental study on using a low-cost 3D joystick and a virtual visit to a zoo. PLoS One 2013;8(12):e83945 [FREE Full text] [CrossRef] [Medline]
    93. Liukkonen J, Barkoukis V, Watt A, Jaakkola T. Motivational climate and students’ emotional experiences and effort in physical education. J Educ Res 2010 Jun 16;103(5):295-308. [CrossRef]
    94. Deci EL, Ryan RM. The "What" and "Why" of goal pursuits: human needs and the self-determination of behavior. Psychol Inq 2000 Oct;11(4):227-268. [CrossRef]
    95. Ouellette MM, LeBrasseur NK, Bean JF, Phillips E, Stein J, Frontera WR, et al. High-intensity resistance training improves muscle strength, self-reported function, and disability in long-term stroke survivors. Stroke 2004 Jun;35(6):1404-1409. [CrossRef] [Medline]
    96. Weiss A, Suzuki T, Bean J, Fielding RA. High intensity strength training improves strength and functional performance after stroke. Am J Phys Med Rehabil 2000;79(4):369-376. [Medline]
    97. Shaw DM. Size of share in task and motivation in work groups. Sociometry 1960 Jun;23(2):203. [CrossRef]
    98. Van den Bossche P, Gijselaers W, Segers M, Kirschner P. Social and cognitive factors driving teamwork in collaborative learning environments. Small Group Res 2006;37(5):490-521. [CrossRef]
    99. Mesmer-Magnus JR, Dechurch LA. Information sharing and team performance: a meta-analysis. J Appl Psychol 2009 Mar;94(2):535-546. [CrossRef] [Medline]
    100. Friedman B, Khan PH, Howe DC. Trust online. Commun ACM 2000;43(12):34-40. [CrossRef]
    101. Robert LP, You S. Are you satisfied yet? Shared leadership, individual trust, autonomy, and satisfaction in virtual teams. J Assoc Inf Sci Technol 2017 Nov 23;69(4):503-513. [CrossRef]
    102. Ijsselsteijn W, Nap HH, de Kort Y, Poels K. Digital game design for elderly users. 2007 Presented at: FuturePlay; November 14-17, 2007; Toronto, Canada p. 17-22. [CrossRef]
    103. Eggermont S, Vandebosch H, Steyaert S. Towards the desired future of the elderly and ICT: policy recommendations based on a dialogue with senior citizens. Poiesis Prax 2006 May 25;4(3):199-217. [CrossRef]
    104. Brokaw EB, Lum PS, Cooper RA, Brewer BR. Using the Kinect to limit abnormal kinematics and compensation strategies during therapy with end effector robots. 2013 Jun Presented at: IEEE International Conference on Rehabilitation Robotics (ICORR); June 24-26, 2013; Seattle, Washington, USA p. 6650384. [CrossRef]
    105. Massie TH, Salisbury JK. The PHANTOM haptic interface: a device for probing virtual objects. 1994 Presented at: Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems; November, 1994; Chicago, Illinois, USA   URL: https://alliance.seas.upenn.edu/~medesign/wiki/uploads/Courses/Massie94-DSC-Phantom.pdf
    106. Helmer P, Conti F, Grange S, Rouller P. Google Patents. 2012. Haptic device gravity compensation   URL: https://patents.google.com/patent/US20100019890 [accessed 2019-04-24] [WebCite Cache]
    107. Füller J, Bartl M, Ernst H, Mühlbacher H. Community based innovation: how to integrate members of virtual communities into new product development. Electron Commer Res 2006 Jan;6(1):57-73. [CrossRef]
    108. Sparrowe RT, Liden RC, Wayne SJ, Kraimer ML. Social networks and the performance of individuals and groups. Acad Manag J 2001 Apr 01;44(2):316-325. [CrossRef]
    109. Kittur A, Nickerson JV, Bernstein M, Gerber E, Shaw A, Zimmerman J, et al. The future of crowd work. In: Proceedings of the 2013 conference on Computer supported cooperative work. 2016 Dec Presented at: Conference on Computer supported cooperative work; February 23-27, 2013; San Antonio, Texas, USA p. 4690-4694   URL: http://europepmc.org/abstract/MED/28269319 [CrossRef]

    ‎
    IRB: institutional review board
    IT: independent termination
    JT: joint termination

    Edited by G Eysenbach; submitted 06.11.18; peer-reviewed by S Rossi, C Nebeker, L Hassan; comments to author 30.01.19; revised version received 12.02.19; accepted 17.02.19; published 15.05.19

    Copyright

    ©Roni Barak Ventura, Shinnosuke Nakayama, Preeti Raghavan, Oded Nov, Maurizio Porfiri. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 15.05.2019.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.