Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A Mobile-based Virtual Reality Speech Rehabilitation App for Patients With Aphasia After Stroke: Development and Pilot Usability Study

A Mobile-based Virtual Reality Speech Rehabilitation App for Patients With Aphasia After Stroke:... Background: Stroke has the highest disability-adjusted life-years lost in any disease, and approximately one-third of the patients get aphasia. Computers and tablets are innovative and aid in intensive treatments in speech rehabilitation for patients with aphasia. However, mechanical training limits the help to patients. Objective: This study aims to provide a framework for an integrated virtual reality (VR) app to provide speech rehabilitation for patients with aphasia. Methods: The content was generated through an in-depth literature review and discussion with experienced rehabilitation physicians and occupational therapists. We then conducted a 2-round Delphi study with 15 experts from hospitals and universities to rate the content using a 5-point Likert scale. The app was developed by an interdisciplinary team involving VR, medical science of rehabilitation, and therapeutic rehabilitation. Pilot usability testing of this novel app was conducted among 5 patients with aphasia, 5 healthy volunteers, 5 medical staff, and 2 VR experts. Results: We designed 4 modules of speech rehabilitation: oral expression, auditory comprehension, cognition, and comprehensive application. Our VR-based interactive and intelligent app was developed to provide an alternative option for patients with aphasia. Pilot usability testing revealed user satisfaction with the app. Conclusions: This study designed and tested a novel VR-based app for speech rehabilitation specifically adapted to patients with aphasia. This will guide other studies to develop a similar program or intelligent system in a clinical setting. (JMIR Serious Games 2022;10(2):e30196) doi: 10.2196/30196 KEYWORDS virtual reality; speech rehabilitation; stroke; app; Delphi https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 1 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al uncontrollable, and inaccessible environments in VR than in Introduction hospitals and will be able to try out new therapeutic strategies [12]. It has been identified that speech treatment based on VR Background is effective. The novel VR platform, EVA Park, is an online Aphasia is an acquired language impairment following acquired virtual island that contains various simulated locations, including brain injury (ABI) that affects some or all language modalities, houses, cafés, restaurants, health centers, hair salons, tropical including the expression and understanding of speech, reading, bars, and discos [13]. It is designed to enable patients with writing, and gestures [1]. ABI is a rapidly growing public health aphasia to communicate successfully with 1 or more problem resulting from traumatic brain injury, stroke, conversation partners via speech using a headset and microphone hypoxic-ischemic encephalopathy after cardiac arrest, and brain in real time [13]. Mirror neuron rehabilitation training software tumors [2]. Stroke leads to the highest disability-adjusted combined with a 4-channel VR panoramic helmet and Z-channel life-year loss in any disease, with over 2 million new cases independent training equipment, including training contents of annually in China [3]. It is estimated that there are 1.1 million nouns, verbs, phrases, and sentence listening and reading, is stroke-related deaths in China per year, and this number is effective for patients with Broca aphasia [14]. Additionally, due increasing [3]. Globally, the number of stroke deaths is projected to the VR system almost completely shielding interference from to rise to 7.8 million by 2030 [4]. the outside world, patients focus more on speech treatment [14]. However, these systems are semi-immersive VR environments Approximately one-third of stroke patients experience aphasia that require partners and the help of a therapist. [1]. Patients with aphasia have a higher risk of not returning to work than those without aphasia [5]. It is likely that an Objectives individual's inability to reenter the workforce poststroke is due This study aims to provide a framework for an integrated VR to the presence of aphasia [5]. The incidence of stroke in app and perform a preliminary test of its usability and safety. younger patients was considerably lower than that in the older Few apps are explicitly designed to meet the requirements of cohorts; however, it remains on the rise [6], and rehabilitation engagement, functionality, aesthetics, and information quality needs are worthy of attention [5]. It has been reported that 94% [11]. Therefore, we intended to add interactive and engaging of individuals are diagnosed with cognitive or communicative elements by providing patients with opportunities to interact deficits. However, only 45% were referred for speech-language with a virtual environment and practice speaking in real-life pathology services [7]. Computers and tablets have proposed scenarios. innovative and intensive treatments for patients with aphasia in language rehabilitation [8-10]. Naming abilities are improved Methods in patients who receive training, whereas no significant improvements have been shown in verbal communication skills Phase I [8,10]. A review concluded that many apps identified from the Google Play Store, Apple App Store, and web searches are Designing the Contents of VR-Based Language available to adults with communication disorders for Rehabilitation: a Delphi Study speech-language therapy. However, few have been designed to For the development of speech rehabilitation content based on specifically meet this vulnerable population's engagement, VR for patients with aphasia after stroke, a 2-round Delphi study functionality, aesthetics, and information quality [11]. was conducted (Figure 1). The Delphi technique is used to obtain Mechanical training limits the help to patients. Communication the most reliable consensus among a panel of experts using a activity in real-world settings is more effective for patients in series of questionnaires [15,16]. Experts can modify, add, or improving their communicative ability [1]. Neither smartphones delete content, as appropriate [17]. We listed the module name, nor computers can make a patient feel like they are in a real-life submodule name, specific contents, and a short description of environment. The development of virtual reality (VR) the training modes during each Delphi round. Between 2 rounds, technology has created interactive computer-generated worlds we revised the content based on expert feedback. In addition to through visual, listening, and touch simulations. This makes written descriptions, the drafts included pictures and visual patients more enthusiastic and more willing to speak. Patients images. will be much more easily accessible to dangerous, expensive, https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 2 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Figure 1. A 2-round Delphi study and development of the app. VR: virtual reality. we learned about existing computer- or tablet-based speech Ethics Approval rehabilitation systems. The Ethics Review Committee of Nursing and Behavioral In November 2020, we held group discussions with experienced Medicine Research, School of Nursing, Central South rehabilitation physicians and occupational therapists at the Brain University, approved the study (E202118; approval date: 21 Hospital of Hunan Province and the Hunan Cancer Hospital to April 2021). determine the relationship between brain regions and language Procedures and Participants production, the rehabilitation process of speech, and their opinions about current speech treatment. In December 2020, Before round 1 of the Delphi study, we conducted a literature we held an online meeting with 2 experts in the Department of review using the search terms “aphasia,” “speech,” “speech Computing at the Hong Kong Polytechnic University to generate treatment,” “language therapy,” “communication,” and ideas about potential content based on VR. Based on the “rehabilitation” via the Wan Fang Database, Chinese National information collected, we developed an initial draft of 20 Knowledge Infrastructure (CNKI), PubMed, and Web of Science submodules in 6 modules. We designed a questionnaire for in September and October 2020 to identify available content experts to review and use a 5-point Likert scale to evaluate and modes related to language rehabilitation. We developed the which part should be included. Each piece was described with content by learning from a model of cognitive communication a name, specific content, and a short description of the training competence [7]. It includes 7 functioning domains: individual, modes. contextual or environmental, cognitive, communication, physical/sensory, emotional/psychosocial, and communication The panel consisted of experts with both theory-based and competence. This model provides evidence that spans the fields practice-based backgrounds to obtain a variety of insights from of speech-language pathology, psychology, neuroscience, researchers. Experts with a theory-based background all rehabilitation, and education, and concerns the complex interplay published papers in the speech treatment or rehabilitation fields. between cognitive, communicative, emotional, and physical Experts with a practice-based background were selected based factors. Rehabilitation platforms, such as Constant Therapy [18] on their practical experience and publications. Additionally, we and Aphasia Therapy Online [19], were retrieved. In addition, approached our network and asked responding experts to provide the names of essential experts in this field, such as a snowball https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 3 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al sampling technique. The experts who participated in the first Data Analysis Delphi round were also approached in the second round. In Microsoft Excel 2003 and IBM SPSS Statistics (version 26.0) total, 21 experts were invited by email, WeChat, or an in-person were used for data analysis and processing. The responses of meeting to participate in the 2 rounds of the Delphi study. It the experts were independently analyzed by 2 researchers, and contained detailed information about the goal, study procedure, the data were double-entered to minimize typing errors. We first-round questionnaire, experts’ demographic information, listed and combined similar recommendations. The mean, SD, and their judgment and familiarity with the corresponding field. and coefficient of variation (CV) were calculated. The CV is Of the 21 experts, 15 (71.4% response rate) responded to the defined as the SD divided by the mean used to describe the invitation. relative dispersion degree of the items' importance evaluation by experts [20,21]. A lower CV value represents a higher degree First Round of coordination among the experts' opinions [21]. After Questionnaire calculating the mean (SD) and CV that the contents should include, we removed all capabilities for which the mean score The first-round questionnaire was composed based on the was less than 4.0 and the CV was over 0.2. Consensus among contents identified by the literature review and group the experts was evaluated using the Kendall coefficient W test. discussions. This resulted in a questionnaire consisting of 6 It refers to the level of intraexpert agreement for all indicators structured modules and 20 submodules. Experts were invited [22]. Statistical significance was considered at a 2-tailed P value to rate these contents on a 5-point Likert scale ranging from 1 of <.05 [21]. Cs represents the experts' familiarity with the (not at all important) to 5 (extremely important). They were research field, and Ca represents the judgment criteria the asked to modify, add, or delete content, as appropriate, and to experts are based on [21]. The defined values are listed in Tables provide suggestions they supposed were reasonable. The 1 and 2. Cr represents the authority coefficient of experts, which experts’ demographic information, as well as their judgment is the mean of the sum of familiar Cs and Ca [23]. We then and familiarity with the corresponding field, was required. We produced the round 2 questionnaire. Complete data were reminded those who did not complete the survey following the collected from 15 (71.4%) of the 21 panel members, and the initial invitation at 2-3-week intervals by WeChat or email. results were used to revise the contents and establish a second questionnaire for round 2 of the Delphi exercise. Table 1. The value of judgment (relevance) criteria. Criterion Influence degree Great Medium Small Theory analysis 0.5 0.4 0.3 Working experience 0.3 0.2 0.1 Referring to literatures 0.15 0.1 0.05 Self-intuition 0.05 0.1 0.15 Table 2. The degree of familiarity with content. Familiarity degree Very familiar Familiar General Unfamiliar Very unfamiliar Self-evaluation 0.9 0.7 0.5 0.3 0.1 Data Analysis Second Round Analysis following round 2 aimed to identify any consensus on Questionnaire the contents and determine whether an additional round was needed. Microsoft Excel 2003 and IBM SPSS Statistics (version In February 2021, we held an online meeting with 2 experts in 26.0) were used for data analysis and processing. We predefined the VR field to discuss the content based on VR before the a mean score of no less than 4, a CV no more than 0.2, and a second round. Experts who completed the first-round 2-tailed P value of the Kendall coefficient W test of no more questionnaire were invited to participate in round 2. Again, we than .05 among experts that the contents should be included. reminded those who did not complete the survey following the The mean, SD, and CV values were calculated for each part. initial invitation at 2-3-week intervals by WeChat or email. The The Delphi survey was completed when all the items in the experts rated each revised section again. They were asked to questionnaire met the aforementioned criteria. rate how much they agreed that each element could be used and to comment on each part. Additional advice about the content that they assumed reasonable was required. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 4 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al and mobile phones makes it a portable and cheap VR equipment Phase II that most patients can afford. The software can be used to Development of the VR-Based Speech Rehabilitation receive language rehabilitation whenever and wherever possible. App The entire VR scene was developed and constructed using the In November 2020 and February 2021, we held online meetings Unity game engine, and Adobe Photoshop was used for 2D with 2 experts in the VR field to discuss this project. A variety assets. Unity is a bridge connecting artificial intelligence (AI) of devices and components can deliver a VR experience, with platforms, VR scenes, and back-end data storage. It uploads the the main categories including smartphone VR headsets, tethered designed text content to the Baidu AI platform and obtains a personal computer (PC)-based VR headsets (eg, HTC Vive), or synthesized voice. The synthesized voice is used in the VR stand-alone VR headsets (eg, Oculus Quest). HTC Vive and scene as a voice instruction to guide the user. Cardboard VR Oculus Quest create a more immersive experience and provide delivers pictures, words, videos, and synthesized voices to the user with a stronger sense of presence than smartphone VR patients that substitute therapists in traditional treatment owing to resolution, frame rate, and sufficient input mechanisms methods. The patients will see the designed questions and [24]. To provide an alternative option for patients under relevant materials in the VR scene and will record their voice significant financial pressure, we chose to use smartphone VR answers using a mobile phone microphone or register their head and developed a mobile-based VR speech rehabilitation app. rotation as an input for selecting answers in multiselection Smartphone VR headsets delivered a VR experience through a questions. Unity would then record their responses according smartphone fitted on a headset, which was as simple as the to their head rotation or send the recorded voice to the Baidu original Google Cardboard. These types of VR apps are AI platform to get the recognized text back. The results will be affordable and cost-effective [24]. Moreover, patients can also stored in an Excel sheet and sent to the therapist's email address use existing smartphones. with the patient’s name. Figure 2 elaborates on the detailed VR-based aphasia therapy process. The VR scene was deployed on mobile phones using the Google Cardboard VR plug-in. Combining Google Cardboard equipment Figure 2. Detailed VR-based aphasia therapy process. AI: artificial intelligence; VR: virtual reality. of the mobile-based VR speech rehabilitation app? All Pilot Application and Usability Testing of the App interviews were recorded using a digital voice recorder and We conducted feasibility exercises of the app on 5 patients with transcribed verbatim within 24 hours. The researcher explored aphasia, 5 healthy volunteers, 5 medical staff, and 2 VR experts participants’ feelings about the app by asking questions in a between 25 and 65 years of age (mean 39.53 years, SD 16.54 private environment, and they were encouraged to put forward years). All 5 patients with aphasia were hospitalized for speech new questions and content. rehabilitation for 2 weeks, and all of them had ischemic stroke. The aphasia quotient (AQ) of the Chinese version of the Western Results Aphasia Battery ranged from 58 to 75 after 2 weeks of face-to-face training. This feasibility exercise allowed us to Results of the Delphi Study better understand and potentially reduce the likelihood of Expert Panel adverse effects in patients with aphasia who may have been Table 3 lists the main characteristics of the 15 (71.4%) of 21 more sensitive to sensory effects (eg, motion sickness and experts who participated in the first and second Delphi rounds. discomfort). Interviews were the measurements used in the The 15 experts’ academic background included the medical experiments. The outline of interview questions after the science of rehabilitation or rehabilitation therapy. Of them, 14 participants’ experience with the mobile-based VR speech (93.3%) were from 3A hospitals, 5 (33.3%) of them took a rehabilitation app is as follows: (1) What groups do you think position in the university, and 2 (13.3%) of them were this mobile-based VR speech rehabilitation app is suitable for? professors. Detailed demographic characteristics of the Delphi (2) How does the mobile-based VR speech rehabilitation app panel are presented in Table 3. The reliability of the experts is differ from face-to-face language rehabilitation? (3) What would presented in Table 4. The mean value of the expert authority encourage or hinder app adoption after using the mobile-based coefficient (Cr) was 0.84. The authority of the experts was high; VR speech rehabilitation app? (4) What is your overall rating thus, the results of this study are trustworthy. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 5 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Table 3. Characterization of experts participating in the Delphi panel (N=15). Characteristic Participants, n (%) Gender Male 2 (13.3) Female 13 (86.7) Age (years) <35 6 (40.0) ≥35 9 (60.0) Education University diploma 9 (60.0) Master’s degree 4 (26.7) Doctoral degree 2 (13.3) Academic background Medical science of rehabilitation 5 (33.3) Rehabilitation Therapeutic 10 (66.7) Working experience (years) 5-9 5 (33.3) 10-14 5 (33.3) ≥15 5 (33.3) Main area of the work role Clinical 10 (66.7) Clinical and research 5 (33.3) Professional title Intermediate 12 (80.0) Deputy senior 1 (6.7) Senior 2 (13.3) https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 6 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al a b c Table 4. Reliability of experts (Ca , Cs , and Cr ). Expert number Criterion score Theoretical analysis Working experience Referring to literature Self-intuition Ca Cs Cr 1 0.4 0.3 0.1 0.1 0.9 0.7 0.8 2 0.5 0.3 0.1 0.1 1 0.7 0.85 3 0.5 0.3 0.05 0.05 0.9 0.5 0.7 4 0.4 0.3 0.15 0.1 0.95 0.9 0.925 5 0.5 0.3 0.1 0.05 0.95 0.9 0.925 6 0.4 0.3 0.1 0.05 0.85 0.9 0.875 7 0.4 0.3 0.15 0.1 0.95 0.7 0.825 8 0.4 0.3 0.15 0.1 0.95 0.7 0.825 9 0.4 0.3 0.1 0.1 0.9 0.7 0.8 10 0.4 0.3 0.1 0.1 0.9 0.9 0.9 11 0.4 0.2 0.05 0.1 0.75 0.7 0.725 12 0.5 0.2 0.05 0.1 0.85 0.7 0.775 13 0.4 0.3 0.1 0.1 0.9 0.9 0.9 14 0.4 0.3 0.1 0.1 0.9 0.9 0.9 15 0.5 0.3 0.15 0.1 1.05 0.9 0.975 Ca: judgment criteria the experts are based on. Cs: experts' familiarity with the research field. Cr: authority coefficient of experts. therapy was combined with oral expression therapy based on First Round the advice of 2 panel members. Furthermore, 2 panel members The Kendall coefficient of concordance (W) was 0.204 (P<.001). suggested that single characters could be deleted because they Of the 6 modules and 20 submodules rated by the panel are not as easy to express as phrases, so we deleted those. Some members in the first round (Table 5), the CV of 2 submodules respondents suggested additional executive functioning, reading, (syllabic and naming with verbs, nouns, and adjectives) and the and writing training. However, this app is meant to improve CV of the arithmetic therapy module that did not meet the patients' oral expression abilities, so reading and writing are criteria were removed (Table 6). In addition, 2 experts beyond our consideration, in particular, executive functioning recommended to delete the arithmetic therapy module. Pictures and writing. Due to smartphone VR and technological and videos describing daily life were added based on the limitations, we were unable to add writing training to VR. The recommendations of 3 panel members. In addition, naming revised content contained 4 modules and 17 submodules. Table 5. Modules in round 1. Module Task 1 Oral expression therapy 2 Auditory comprehension therapy 3 Cognition therapy 4 Naming therapy 5 Arithmetic therapy 6 Comprehensive application https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 7 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Table 6. Scores in round 1. Tasks Mean (SD) CV Oral expression therapy Read and repeat task: syllabic 4.13 (0.99) 0.24 Read and repeat task: single characters, including the 150 most commonly used Chinese characters 4.60 (0.91) 0.20 Read and repeat task: some phrases, including categories of digital, fruit, animals, vegetables, transportation, kitchen 5 (0) 0.00 supplies, daily necessities, body parts, food, address, location, and sports Read and repeat task: sentences, including proverbs and daily expressions 4.87 (0.35) 0.07 Answering some questions related to everyday life 4.80 (0.56) 0.12 Auditory comprehension therapy Listening and matching: included digital, color, pictures, and words 4.93 (0.26) 0.05 Yes-no questions 4.93 (0.26) 0.05 Listening to the passage and answering questions 4.73 (0.46) 0.10 Cognition therapy Attention training 4.27 (0.80) 0.19 Memory training 4.13 (0.74) 0.18 Reasoning and problem solving 4.47 (0.83) 0.19 Naming therapy Picture naming 4.93 (0.26) 0.05 Naming with verbs, nouns, and adjectives 4.33 (0.90) 0.21 Naming a list of items with the exact nature 4.93 (0.26) 0.05 Arithmetic therapy Addition, subtraction, multiplication, and division 4.27 (1.16) 0.27 Comprehensive application Supermarket task: A computer-generated virtual customer purchases goods. The patient has to pick items, count the 4.80 (0.56) 0.12 total amount of money that the virtual customer has selected, and return the change. Interview task: The patient plays the role of a job seeker. A computer-generated virtual interviewer asks questions about 4.73 (0.59) 0.13 the patient’s personal information (eg, name, gender, age, nationality, education level, birth date, height, weight, spe- cialty, address, and family members) and assesses their language skills by reading text and describing pictures. Bedroom task: The patient plays the role of a mother. By communicating with the computer-generated virtual daughter, 4.80 (0.56) 0.12 the patient selects clothes for the virtual daughter according to the weather conditions and discusses breakfast and what to choose for dinner and the kind of transportation to use. Ordering task: The patient acts like a customer to order in a virtual restaurant; communicates with a virtual server syn- 4.80 (0.41) 0.09 thesized by the computer; requests a certain number of dishes, desserts, cakes, and drinks; and completes the payment. Park task: By buying tickets, paying, and asking for directions in the virtual ticket office, together with a computer- 4.80 (0.41) 0.09 generated virtual friend, the patient enters a virtual park. The patient and the virtual friend communicate with each other about the scene. Likert scale: 1, not important; 2, somewhat important; 3, moderately important; 4 = important; and 5, very important. CV: coefficient of variation. CV>0.2. Deleted on the recommendations of the 2 panel members. suggested that patients should complete answering within the Second Round allotted time. Some respondents suggested additional executive The results of the second round are presented in Tables 7 and functioning, reading, and writing training. However, this app 8 and Figure 3. All mean scores for importance were above is meant to improve patients' oral expression abilities, so reading 4.00, and the CVs were less than 0.20. The Kendall coefficient and writing are beyond our consideration. The revised content of concordance (W) was 0.335 (P<.001). Finally, a consensus contained 4 modules and 17 submodules. was reached on 4 modules and 17 submodules. One expert https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 8 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Table 7. Modules in round 2. Module Task 1 Oral expression therapy 2 Auditory comprehension therapy 3 Cognition therapy 4 Comprehensive application Table 8. Scores in round 2. Tasks Mean (SD) CV Oral expression therapy Read and repeat task: some phrases, including categories of digital, fruit, animals, vegetables, transportation, kitchen 4.93 (0.26) 0.05 supplies, daily necessities, body parts, food, address, location, action, and sports Read and repeat task: sentences, including proverbs and daily expressions 5.00 (0) 0 Answering some questions related to everyday life 4.87 (0.35) 0.07 Naming task: picture naming 5.00 (0) 0 Naming task: naming list of items with the exact nature 4.93 (0.26) 0.05 Describing pictures and videos pertaining to daily life 4.67 (0.49) 0.10 Auditory comprehension therapy Listening and matching: included digital, color, pictures, and words 5.00 (0) 0 Yes-no questions 4.93 (0.26) 0.05 Listening to the passage and answering questions 4.80 (0.41) 0.09 Cognition therapy Attention training 4.27 (0.70) 0.16 Memory training 4.13 (0.74) 0.18 Reasoning and problem solving 4.20 (0.77) 0.18 Comprehensive application Supermarket task: A computer-generated virtual customer purchases goods. The patient has to pick items, count the 4.80 (0.56) 0.12 total amount of money that the virtual customer has selected, and return the change. Interview task: The patient plays the role of a job seeker. A computer-generated virtual interviewer asks questions 4.80 (0.56) 0.12 about the patient’s personal information (eg, name, gender, age, nationality, education level, birth date, height, weight, specialty, address, and family members) and assesses their language skills by reading text and describing pictures. Bedroom task: The patient plays the role of a mother. By communicating with the computer-generated virtual 4.87 (0.52) 0.11 daughter, the patient selects clothes for the virtual daughter according to the weather conditions and discusses breakfast and what to choose for dinner and the kind of transportation to use. Ordering task: The patient acts like a customer to order in a virtual restaurant; communicates with a virtual server 4.87 (0.52) 0.11 synthesized by the computer; requests a certain number of dishes, desserts, cakes, and drinks; and completes the payment. Park task: By buying tickets, paying, and asking for directions in the virtual ticket office, together with a computer- 4.53 (0.64) 0.14 generated virtual friend, the patient enters a virtual park. The patient and the virtual friend communicate with each other about the scene. CV: coefficient of variation. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 9 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Figure 3. Consensus on 4 modules. In response to question 1, “What groups do you think this Usability Testing of the App mobile-based VR speech rehabilitation app is suitable for?”, 12 All 17 (100%) participants could open the app. A user-centered (70.6%) of 17 participants suggested that the user group could design philosophy was implemented to create additional include patients who were discharged or currently hospitalized. interactive and customizable features that indicated a high Participants stated that with this app, they could correct and degree of usability for patients with aphasia. Participants' refine their language skills, and 5 (29.4%) of 17 participants comments were collected and descriptively analyzed based on recommended that the user group should include people, the interview outline. The 4 responses were classified based on especially those who are under significant financial pressure their feedback. Figure 4 shows the functionality of the VR-based and do not have easy access to medical treatment but would aphasia therapy app. prefer an opportunity to learn. Figure 4. Functionality of the VR-based aphasia therapy app. VR: virtual reality. In response to question 2, “How does the mobile-based VR difficulty of going to the hospital and is more convenient as speech rehabilitation app differ from face-to-face language they do not have to worry about weather and traffic. In addition, training?”, all 17 (100%) participants mentioned the advantages 15 (88.2%) of 17 participants reported a high level of immersion of in-home apps. Participants stated that this app reduces the and engagement and thought that patients could thoroughly https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 10 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al enjoy themselves when using this app without judgment from in patients with poststroke aphasia influences language others. Additionally, there is no need to worry about rehabilitation outcomes [33]. self-abasement because of the inability to speak fluently. Language rehabilitation combined with cognitive rehabilitation Moreover, 4 (23.5%) of 17 participants critiqued the can improve patients' communication skills [34]. The cognitive effectiveness of this app because patients could not receive domain includes processing, attention, working memory, social immediate feedback, and the efficiency of using this app could cognition, reasoning, and problem-solving [7]. Therefore, to not be guaranteed. improve the patients’ oral expression abilities, we designed In response to question 3, “After using the mobile-based VR cognitive training pertaining to reasoning, memory, attention, speech rehabilitation app, what would encourage or hinder your visuospatial perception, and logical thinking. A range of brain adoption of the app?”, 15 (88.2%) of 17 participants mentioned regions are related to cognition, and these nonlanguage domains that the design and arrangement met patients’ daily play a role in the ability of patients with chronic aphasia [29]. communication needs; they felt more secure because the training To emphasize the communication demands of an individual's content was designed by a professional rehabilitation team. In life, the contextual domain is placed at the top of the cognitive addition, 15 (88.2%) participants mentioned that the app was communication competence model [7]. A Cochrane review easy to use, and 3 (17.6%) reported low levels of motion concluded that communication activity in real-world settings, sickness. or functional communication, could improve communicative In response to question 4, “What is your overall rating of the ability [1]. Communication practices should occur in an mobile-based VR speech rehabilitation app?”, all 17 (100%) ecological context. Therefore, we designed 5 virtual scenario participants indicated that the app was novel, and 5 (29.4%) of tasks focusing on everyday communication activities and 17 participants gave some recommendations, such as considering pragmatic conversational skills. Patients may receive language different training difficulties and increased functionality. All practice opportunities in natural context communication settings patients expressed interest in language rehabilitation using the in VR. These tasks are a comprehensive practice for the patients, new technology. and different short communication dialogues of everyday activities and cognitive exercises were integrated, including Discussion oral expression, naming, calculation, attention, memory, and reasoning tasks. VR technology is a promising rehabilitation Principal Findings tool and may be a useful alternative to conventional training [35], but the scene in VR is inaccurate. However, we could This study developed the first accessible and cost-effective create almost everything in VR, and patients experience it as mobile-based VR speech rehabilitation app in China. Through an actual situation, which makes it easy to try new therapeutic a 2-round Delphi study, panel members reached a consensus strategies. All learning in VR can be transferred to the real world on 4 modules and 17 submodules that benefit patients in gaining [12]. With clinician shortages and a higher prevalence of essential daily communication language skills and various skills aphasia, VR is not meant to replace skilled therapists but to ease in everyday life. Patients get opportunities to interact with the clinicians' burdens. virtual environment and can practice at any time. A review systematically identified and evaluated a series of mobile apps Limitations for speech-language therapy and found a lack of interactive and This study had some limitations. First, a range of brain regions engaging elements in the apps, which failed to make patients are related to attention and executive functioning, and these self-manage [11]. The content we designed based on VR nonlanguage domains play a role in the abilities of patients with emphasizes user experience, engagement, and visual appeal to chronic aphasia [29]. Owing to the limitations of intelligent VR, improve patient adherence. We developed real-life scenarios we cannot design some daily instructions (eg, washing clothes, and gaming factors for the VR environment. Patients get cutting vegetables, cooking). VR could be used to provide an opportunities to interact with a virtual environment to practice enriched environment where patients could master skills related language in natural context communication settings in VR. We to daily life that cannot be finished in the hospital. Learning in believe that the rehabilitation content will provide patients with VR can then be transferred to real-life situations. Second, a strong basis for returning to society and work. dialects vary across different areas of China. However, we did We developed the content by learning from a model of cognitive not involve dialects from different regions of China to provide communication competence [7]. Language processing is the options for those who have difficulty understanding and core of cognition and requires the participation of other expressing Mandarin. There are nonlanguage abilities that would nonlinguistic cognitive functions [25,26]. Left hemispheric be worthwhile to consider in aphasia rehabilitation, such as stroke leading to impairments in language processing often prosody and emotional tone in utterances. However, it is difficult affects other cognitive functions, such as executive function, to implement these in the app. In addition, the usability of the attention, visuospatial perception, logical thinking, and memory app was assessed by only 17 individuals, which may have biased [27,28], which play an essential role in aphasia recovery and the usability results. In the future, we will design a more rehabilitation [29-31]. Patients' language and nonlinguistic significant multicenter trial with a longer-term follow-up. cognitive functions cannot be separated owing to complex Additionally, this app is meant to improve patients' oral neurobiological networks [32]. Nonlinguistic cognitive function expression abilities, so reading and writing are beyond our https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 11 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al consideration. In particular, owing to the current limitations of significant potential to make our app an in-home app in the technology, we cannot plug writing training into VR. future, contributing to the automation of rehabilitation administration. However, further studies are needed to evaluate Conclusion the feasibility and efficacy of our app. We hope that our research We conducted a Delphi study and developed a mobile-based provides guidelines and references for others in the medical VR app. This study constitutes a step toward the development field. of a combination of health and VR. We believe there is Acknowledgments We gratefully thank the experts for completing the study surveys. This research was supported by the Fundamental Research Funds for the Central Universities of Central South University (project no. 2020zzts847) and the Hunan Provincial Health Commission (project nos. 2020SK51104 and 202114021494). Authors' Contributions All authors listed in this publication conducted the research. XB was responsible for all stages of the project. XB, PHFN, QC, YT, RF, QT, QC, SL, ASKC, and XL participated in the design, analysis, and writing of the article. PHFN and QC provided expertise in the development of the app. PHFN, QC, and ASKC revised the article. All authors approved the final draft. Conflicts of Interest None declared. References 1. Brady M, Kelly H, Godwin J, Enderby P, Campbell P. Speech and language therapy for aphasia following stroke. Cochrane DB Syst Rev 2020 Jun 01:2016. [doi: 10.1002/14651858.cd000425.pub4] 2. Menon DK, Bryant C. Time for change in acquired brain injury. Lancet Neurol 2019 Jan;18(1):28. [doi: 10.1016/s1474-4422(18)30463-0] 3. Wu S, Wu B, Liu M, Chen Z, Wang W, Anderson CS, et al. Stroke in China: advances and challenges in epidemiology, prevention, and management. Lancet Neurol 2019 Apr;18(4):394-405. [doi: 10.1016/s1474-4422(18)30500-3] 4. Strong K, Mathers C, Bonita R. Preventing stroke: saving lives around the world. Lancet Neurol 2007 Feb;6(2):182-187. [doi: 10.1016/s1474-4422(07)70031-5] 5. Graham JR, Pereira S, Teasell R. Aphasia and return to work in younger stroke survivors. Aphasiology 2011 May 25;25(8):952-960. [doi: 10.1080/02687038.2011.563861] 6. Béjot Y, Bailly H, Durier J, Giroud M. Epidemiology of stroke in Europe and trends for the 21st century. Presse Med 2016 Dec;45(12 Pt 2):e391-e398. [doi: 10.1016/j.lpm.2016.10.003] [Medline: 27816343] 7. MacDonald S. Introducing the model of cognitive-communication competence: a model to guide evidence-based communication interventions after brain injury. Brain Inj 2017 Oct 24;31(13-14):1760-1780. [doi: 10.1080/02699052.2017.1379613] [Medline: 29064304] 8. Lavoie M, Macoir J, Bier N. Effectiveness of technologies in the treatment of post-stroke anomia: a systematic review. J Commun Disord 2017 Jan;65:43-53. [doi: 10.1016/j.jcomdis.2017.01.001] [Medline: 28171741] 9. Kurland J, Liu A, Stokes P. Effects of a tablet-based home practice program with telepractice on treatment outcomes in chronic aphasia. J Speech Lang Hear Res 2018 May 17;61(5):1140-1156. [doi: 10.1044/2018_jslhr-l-17-0277] 10. Lavoie M, Routhier S, Légaré A, Macoir J. Treatment of verb anomia in aphasia: efficacy of self-administered therapy using a smart tablet. Neurocase 2016 May 26;22(1):109-118. [doi: 10.1080/13554794.2015.1051055] [Medline: 26007615] 11. Vaezipour A, Campbell J, Theodoros D, Russell T. Mobile apps for speech-language therapy in adults with communication disorders: review of content and quality. JMIR Mhealth Uhealth 2020 Oct 29;8(10):e18858. [doi: 10.2196/18858] 12. Freeman D, Reeve S, Robinson A, Ehlers A, Clark D, Spanlang B, et al. Virtual reality in the assessment, understanding, and treatment of mental health disorders. Psychol Med 2017 Mar 22;47(14):2393-2400. [doi: 10.1017/s003329171700040x] 13. Marshall J, Booth T, Devane N, Galliers J, Greenwood H, Hilari K, et al. Evaluating the benefits of aphasia intervention delivered in virtual reality: results of a quasi-randomised study. PLoS One 2016 Aug 12;11(8):e0160381 [FREE Full text] [doi: 10.1371/journal.pone.0160381] [Medline: 27518188] 14. Zhang Y, Chen P, Li X. Clinical research on therapeutic effect of virtual reality technology on Broca aphasia patients. 2009 Presented at: Proceedings of the International Conference on Information Technology; 2017; Ioannina, Epirus, Greece p. 134. [doi: 10.1109/incit.2017.8257880] 15. Keeney S, Hasson F, McKenna HP. A critical review of the Delphi technique as a research methodology for nursing. Int J Nurs Stud 2001 Apr;38(2):195-200. [doi: 10.1016/s0020-7489(00)00044-4] 16. Powell C. The Delphi technique: myths and realities. J Adv Nurs 2003 Feb;41(4):376-382. [doi: 10.1046/j.1365-2648.2003.02537.x] [Medline: 12581103] https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 12 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al 17. Mullen PM. Delphi: myths and reality. J Health Organ Manag 2003 Feb;17(1):37-52. [doi: 10.1108/14777260310469319] [Medline: 12800279] 18. Brain Health, for Life. URL: https://constanttherapyhealth.com/ [accessed 2022-04-01] 19. Therapy Menu. URL: http://www.aphasiatherapyonline.com/main.html [accessed 2022-04-01] 20. Reed GF, Lynn F, Meade BD. Use of coefficient of variation in assessing variability of quantitative assays. Clin Diagn Lab Immunol 2002 Nov;9(6):1235-1239 [FREE Full text] [doi: 10.1128/cdli.9.6.1235-1239.2002] [Medline: 12414755] 21. Chen Y, Li Z, Peng J, Shen L, Shi J. Development of a care bundle for stroke survivors with psychological symptoms: evidence summary and Delphi study. Evid Based Complement Alternat Med 2020 Jun 30;2020:7836024 [FREE Full text] [doi: 10.1155/2020/7836024] [Medline: 32714421] 22. Xing Q, Zhang M, Zhao F, Zhou Y, Mo Y, Yuan L. The development of a standardized framework for primary nurse specialists in diabetes care in China: a Delphi study. J Nurs Res 2019 Dec;27(6):e53. [doi: 10.1097/jnr.0000000000000330] [Medline: 31162295] 23. Ren H, Liu C, Wang R, Zhang M, Ma F, Li R, et al. Core competencies required for gastroenterology nursing specialists in China. Gastroenterol Nurs 2019;42(2):169-178. [doi: 10.1097/sga.0000000000000392] 24. Shen J, Xiang H, Luna J, Grishchenko A, Patterson J, Strouse RV, et al. Virtual reality-based executive function rehabilitation system for children with traumatic brain injury: design and usability study. JMIR Serious Games 2020 Aug 25;8(3):e16947 [FREE Full text] [doi: 10.2196/16947] [Medline: 32447275] 25. Ardila A, Rubio-Bruno S. Aphasia from the inside: the cognitive world of the aphasic patient. Appl Neuropsychol Adult 2018 May 23;25(5):434-440. [doi: 10.1080/23279095.2017.1323753] [Medline: 28535071] 26. Peach R. Cognitive approaches to aphasia treatment: application of the cognition of language to aphasia intervention. Semin Speech Lang 2017 Feb 15;38(1):3-4. [doi: 10.1055/s-0036-1597259] [Medline: 28201832] 27. Murray LL. Attention and other cognitive deficits in aphasia: presence and relation to language and communication measures. Am J Speech Lang Pathol 2012 May;21(2):S51-S64. [doi: 10.1044/1058-0360(2012/11-0067)] 28. Yao J, Liu X, Liu Q, Wang J, Ye N, Lu X, et al. Characteristics of non-linguistic cognitive impairment in post-stroke aphasia patients. Front Neurol 2020 Sep 30;11:1038 [FREE Full text] [doi: 10.3389/fneur.2020.01038] [Medline: 33117251] 29. Schumacher R, Halai AD, Lambon Ralph MA. Assessing and mapping language, attention and executive multidimensional deficits in stroke aphasia. Brain 2019 Oct 01;142(10):3202-3216 [FREE Full text] [doi: 10.1093/brain/awz258] [Medline: 31504247] 30. El Hachioui H, Visch-Brink EG, Lingsma HF, van de Sandt-Koenderman MWME, Dippel DWJ, Koudstaal PJ, et al. Nonlinguistic cognitive impairment in poststroke aphasia: a prospective study. Neurorehabil Neural Repair 2014 Nov 08;28(3):273-281. [doi: 10.1177/1545968313508467] [Medline: 24213963] 31. Brownsett SLE, Warren JE, Geranmayeh F, Woodhead Z, Leech R, Wise RJS. Cognitive control and its impact on recovery from aphasic stroke. Brain 2014 Jan;137(Pt 1):242-254 [FREE Full text] [doi: 10.1093/brain/awt289] [Medline: 24163248] 32. Hagoort P. The core and beyond in the language-ready brain. Neurosci Biobehav Rev 2017 Oct;81(Pt B):194-204. [doi: 10.1016/j.neubiorev.2017.01.048] [Medline: 28193452] 33. Gilmore N, Meier EL, Johnson JP, Kiran S. Nonlinguistic cognitive factors predict treatment-induced recovery in chronic poststroke aphasia. Arch Phys Med Rehabil 2019 Jul;100(7):1251-1258 [FREE Full text] [doi: 10.1016/j.apmr.2018.12.024] [Medline: 30639272] 34. Zakariás L, Keresztes A, Marton K, Wartenburger I. Positive effects of a computerised working memory and executive function training on sentence comprehension in aphasia. Neuropsychol Rehabil 2018 Apr 21;28(3):369-386. [doi: 10.1080/09602011.2016.1159579] [Medline: 26999324] 35. Dockx K, Bekkers E, Van DBV. Virtual reality for rehabilitation in Parkinson's disease. Cochrane DB Syst Rev 2016:CD010760. [doi: 10.1002/14651858.cd010760] Abbreviations ABI: acquired brain injury AI: artificial intelligence CV: coefficient of variation VR: virtual reality https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 13 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Edited by N Zary; submitted 23.05.21; peer-reviewed by T Ong, G Cárdenas-López; comments to author 03.12.21; revised version received 28.02.22; accepted 05.03.22; published 07.04.22 Please cite as: Bu X, Ng PHF, Tong Y, Chen PQ, Fan R, Tang Q, Cheng Q, Li S, Cheng ASK, Liu X A Mobile-based Virtual Reality Speech Rehabilitation App for Patients With Aphasia After Stroke: Development and Pilot Usability Study JMIR Serious Games 2022;10(2):e30196 URL: https://games.jmir.org/2022/2/e30196 doi: 10.2196/30196 PMID: ©Xiaofan Bu, Peter HF Ng, Ying Tong, Peter Q Chen, Rongrong Fan, Qingping Tang, Qinqin Cheng, Shuangshuang Li, Andy SK Cheng, Xiangyu Liu. Originally published in JMIR Serious Games (https://games.jmir.org), 07.04.2022. 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 JMIR Serious Games, is properly cited. The complete bibliographic information, a link to the original publication on https://games.jmir.org, as well as this copyright and license information must be included. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 14 (page number not for citation purposes) XSL FO RenderX http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JMIR Serious Games JMIR Publications

A Mobile-based Virtual Reality Speech Rehabilitation App for Patients With Aphasia After Stroke: Development and Pilot Usability Study

Loading next page...
 
/lp/jmir-publications/a-mobile-based-virtual-reality-speech-rehabilitation-app-for-patients-WGivoJUxSO

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
JMIR Publications
Copyright
Copyright © The Author(s). Licensed under Creative Commons Attribution cc-by 4.0
ISSN
2291-9279
DOI
10.2196/30196
Publisher site
See Article on Publisher Site

Abstract

Background: Stroke has the highest disability-adjusted life-years lost in any disease, and approximately one-third of the patients get aphasia. Computers and tablets are innovative and aid in intensive treatments in speech rehabilitation for patients with aphasia. However, mechanical training limits the help to patients. Objective: This study aims to provide a framework for an integrated virtual reality (VR) app to provide speech rehabilitation for patients with aphasia. Methods: The content was generated through an in-depth literature review and discussion with experienced rehabilitation physicians and occupational therapists. We then conducted a 2-round Delphi study with 15 experts from hospitals and universities to rate the content using a 5-point Likert scale. The app was developed by an interdisciplinary team involving VR, medical science of rehabilitation, and therapeutic rehabilitation. Pilot usability testing of this novel app was conducted among 5 patients with aphasia, 5 healthy volunteers, 5 medical staff, and 2 VR experts. Results: We designed 4 modules of speech rehabilitation: oral expression, auditory comprehension, cognition, and comprehensive application. Our VR-based interactive and intelligent app was developed to provide an alternative option for patients with aphasia. Pilot usability testing revealed user satisfaction with the app. Conclusions: This study designed and tested a novel VR-based app for speech rehabilitation specifically adapted to patients with aphasia. This will guide other studies to develop a similar program or intelligent system in a clinical setting. (JMIR Serious Games 2022;10(2):e30196) doi: 10.2196/30196 KEYWORDS virtual reality; speech rehabilitation; stroke; app; Delphi https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 1 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al uncontrollable, and inaccessible environments in VR than in Introduction hospitals and will be able to try out new therapeutic strategies [12]. It has been identified that speech treatment based on VR Background is effective. The novel VR platform, EVA Park, is an online Aphasia is an acquired language impairment following acquired virtual island that contains various simulated locations, including brain injury (ABI) that affects some or all language modalities, houses, cafés, restaurants, health centers, hair salons, tropical including the expression and understanding of speech, reading, bars, and discos [13]. It is designed to enable patients with writing, and gestures [1]. ABI is a rapidly growing public health aphasia to communicate successfully with 1 or more problem resulting from traumatic brain injury, stroke, conversation partners via speech using a headset and microphone hypoxic-ischemic encephalopathy after cardiac arrest, and brain in real time [13]. Mirror neuron rehabilitation training software tumors [2]. Stroke leads to the highest disability-adjusted combined with a 4-channel VR panoramic helmet and Z-channel life-year loss in any disease, with over 2 million new cases independent training equipment, including training contents of annually in China [3]. It is estimated that there are 1.1 million nouns, verbs, phrases, and sentence listening and reading, is stroke-related deaths in China per year, and this number is effective for patients with Broca aphasia [14]. Additionally, due increasing [3]. Globally, the number of stroke deaths is projected to the VR system almost completely shielding interference from to rise to 7.8 million by 2030 [4]. the outside world, patients focus more on speech treatment [14]. However, these systems are semi-immersive VR environments Approximately one-third of stroke patients experience aphasia that require partners and the help of a therapist. [1]. Patients with aphasia have a higher risk of not returning to work than those without aphasia [5]. It is likely that an Objectives individual's inability to reenter the workforce poststroke is due This study aims to provide a framework for an integrated VR to the presence of aphasia [5]. The incidence of stroke in app and perform a preliminary test of its usability and safety. younger patients was considerably lower than that in the older Few apps are explicitly designed to meet the requirements of cohorts; however, it remains on the rise [6], and rehabilitation engagement, functionality, aesthetics, and information quality needs are worthy of attention [5]. It has been reported that 94% [11]. Therefore, we intended to add interactive and engaging of individuals are diagnosed with cognitive or communicative elements by providing patients with opportunities to interact deficits. However, only 45% were referred for speech-language with a virtual environment and practice speaking in real-life pathology services [7]. Computers and tablets have proposed scenarios. innovative and intensive treatments for patients with aphasia in language rehabilitation [8-10]. Naming abilities are improved Methods in patients who receive training, whereas no significant improvements have been shown in verbal communication skills Phase I [8,10]. A review concluded that many apps identified from the Google Play Store, Apple App Store, and web searches are Designing the Contents of VR-Based Language available to adults with communication disorders for Rehabilitation: a Delphi Study speech-language therapy. However, few have been designed to For the development of speech rehabilitation content based on specifically meet this vulnerable population's engagement, VR for patients with aphasia after stroke, a 2-round Delphi study functionality, aesthetics, and information quality [11]. was conducted (Figure 1). The Delphi technique is used to obtain Mechanical training limits the help to patients. Communication the most reliable consensus among a panel of experts using a activity in real-world settings is more effective for patients in series of questionnaires [15,16]. Experts can modify, add, or improving their communicative ability [1]. Neither smartphones delete content, as appropriate [17]. We listed the module name, nor computers can make a patient feel like they are in a real-life submodule name, specific contents, and a short description of environment. The development of virtual reality (VR) the training modes during each Delphi round. Between 2 rounds, technology has created interactive computer-generated worlds we revised the content based on expert feedback. In addition to through visual, listening, and touch simulations. This makes written descriptions, the drafts included pictures and visual patients more enthusiastic and more willing to speak. Patients images. will be much more easily accessible to dangerous, expensive, https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 2 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Figure 1. A 2-round Delphi study and development of the app. VR: virtual reality. we learned about existing computer- or tablet-based speech Ethics Approval rehabilitation systems. The Ethics Review Committee of Nursing and Behavioral In November 2020, we held group discussions with experienced Medicine Research, School of Nursing, Central South rehabilitation physicians and occupational therapists at the Brain University, approved the study (E202118; approval date: 21 Hospital of Hunan Province and the Hunan Cancer Hospital to April 2021). determine the relationship between brain regions and language Procedures and Participants production, the rehabilitation process of speech, and their opinions about current speech treatment. In December 2020, Before round 1 of the Delphi study, we conducted a literature we held an online meeting with 2 experts in the Department of review using the search terms “aphasia,” “speech,” “speech Computing at the Hong Kong Polytechnic University to generate treatment,” “language therapy,” “communication,” and ideas about potential content based on VR. Based on the “rehabilitation” via the Wan Fang Database, Chinese National information collected, we developed an initial draft of 20 Knowledge Infrastructure (CNKI), PubMed, and Web of Science submodules in 6 modules. We designed a questionnaire for in September and October 2020 to identify available content experts to review and use a 5-point Likert scale to evaluate and modes related to language rehabilitation. We developed the which part should be included. Each piece was described with content by learning from a model of cognitive communication a name, specific content, and a short description of the training competence [7]. It includes 7 functioning domains: individual, modes. contextual or environmental, cognitive, communication, physical/sensory, emotional/psychosocial, and communication The panel consisted of experts with both theory-based and competence. This model provides evidence that spans the fields practice-based backgrounds to obtain a variety of insights from of speech-language pathology, psychology, neuroscience, researchers. Experts with a theory-based background all rehabilitation, and education, and concerns the complex interplay published papers in the speech treatment or rehabilitation fields. between cognitive, communicative, emotional, and physical Experts with a practice-based background were selected based factors. Rehabilitation platforms, such as Constant Therapy [18] on their practical experience and publications. Additionally, we and Aphasia Therapy Online [19], were retrieved. In addition, approached our network and asked responding experts to provide the names of essential experts in this field, such as a snowball https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 3 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al sampling technique. The experts who participated in the first Data Analysis Delphi round were also approached in the second round. In Microsoft Excel 2003 and IBM SPSS Statistics (version 26.0) total, 21 experts were invited by email, WeChat, or an in-person were used for data analysis and processing. The responses of meeting to participate in the 2 rounds of the Delphi study. It the experts were independently analyzed by 2 researchers, and contained detailed information about the goal, study procedure, the data were double-entered to minimize typing errors. We first-round questionnaire, experts’ demographic information, listed and combined similar recommendations. The mean, SD, and their judgment and familiarity with the corresponding field. and coefficient of variation (CV) were calculated. The CV is Of the 21 experts, 15 (71.4% response rate) responded to the defined as the SD divided by the mean used to describe the invitation. relative dispersion degree of the items' importance evaluation by experts [20,21]. A lower CV value represents a higher degree First Round of coordination among the experts' opinions [21]. After Questionnaire calculating the mean (SD) and CV that the contents should include, we removed all capabilities for which the mean score The first-round questionnaire was composed based on the was less than 4.0 and the CV was over 0.2. Consensus among contents identified by the literature review and group the experts was evaluated using the Kendall coefficient W test. discussions. This resulted in a questionnaire consisting of 6 It refers to the level of intraexpert agreement for all indicators structured modules and 20 submodules. Experts were invited [22]. Statistical significance was considered at a 2-tailed P value to rate these contents on a 5-point Likert scale ranging from 1 of <.05 [21]. Cs represents the experts' familiarity with the (not at all important) to 5 (extremely important). They were research field, and Ca represents the judgment criteria the asked to modify, add, or delete content, as appropriate, and to experts are based on [21]. The defined values are listed in Tables provide suggestions they supposed were reasonable. The 1 and 2. Cr represents the authority coefficient of experts, which experts’ demographic information, as well as their judgment is the mean of the sum of familiar Cs and Ca [23]. We then and familiarity with the corresponding field, was required. We produced the round 2 questionnaire. Complete data were reminded those who did not complete the survey following the collected from 15 (71.4%) of the 21 panel members, and the initial invitation at 2-3-week intervals by WeChat or email. results were used to revise the contents and establish a second questionnaire for round 2 of the Delphi exercise. Table 1. The value of judgment (relevance) criteria. Criterion Influence degree Great Medium Small Theory analysis 0.5 0.4 0.3 Working experience 0.3 0.2 0.1 Referring to literatures 0.15 0.1 0.05 Self-intuition 0.05 0.1 0.15 Table 2. The degree of familiarity with content. Familiarity degree Very familiar Familiar General Unfamiliar Very unfamiliar Self-evaluation 0.9 0.7 0.5 0.3 0.1 Data Analysis Second Round Analysis following round 2 aimed to identify any consensus on Questionnaire the contents and determine whether an additional round was needed. Microsoft Excel 2003 and IBM SPSS Statistics (version In February 2021, we held an online meeting with 2 experts in 26.0) were used for data analysis and processing. We predefined the VR field to discuss the content based on VR before the a mean score of no less than 4, a CV no more than 0.2, and a second round. Experts who completed the first-round 2-tailed P value of the Kendall coefficient W test of no more questionnaire were invited to participate in round 2. Again, we than .05 among experts that the contents should be included. reminded those who did not complete the survey following the The mean, SD, and CV values were calculated for each part. initial invitation at 2-3-week intervals by WeChat or email. The The Delphi survey was completed when all the items in the experts rated each revised section again. They were asked to questionnaire met the aforementioned criteria. rate how much they agreed that each element could be used and to comment on each part. Additional advice about the content that they assumed reasonable was required. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 4 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al and mobile phones makes it a portable and cheap VR equipment Phase II that most patients can afford. The software can be used to Development of the VR-Based Speech Rehabilitation receive language rehabilitation whenever and wherever possible. App The entire VR scene was developed and constructed using the In November 2020 and February 2021, we held online meetings Unity game engine, and Adobe Photoshop was used for 2D with 2 experts in the VR field to discuss this project. A variety assets. Unity is a bridge connecting artificial intelligence (AI) of devices and components can deliver a VR experience, with platforms, VR scenes, and back-end data storage. It uploads the the main categories including smartphone VR headsets, tethered designed text content to the Baidu AI platform and obtains a personal computer (PC)-based VR headsets (eg, HTC Vive), or synthesized voice. The synthesized voice is used in the VR stand-alone VR headsets (eg, Oculus Quest). HTC Vive and scene as a voice instruction to guide the user. Cardboard VR Oculus Quest create a more immersive experience and provide delivers pictures, words, videos, and synthesized voices to the user with a stronger sense of presence than smartphone VR patients that substitute therapists in traditional treatment owing to resolution, frame rate, and sufficient input mechanisms methods. The patients will see the designed questions and [24]. To provide an alternative option for patients under relevant materials in the VR scene and will record their voice significant financial pressure, we chose to use smartphone VR answers using a mobile phone microphone or register their head and developed a mobile-based VR speech rehabilitation app. rotation as an input for selecting answers in multiselection Smartphone VR headsets delivered a VR experience through a questions. Unity would then record their responses according smartphone fitted on a headset, which was as simple as the to their head rotation or send the recorded voice to the Baidu original Google Cardboard. These types of VR apps are AI platform to get the recognized text back. The results will be affordable and cost-effective [24]. Moreover, patients can also stored in an Excel sheet and sent to the therapist's email address use existing smartphones. with the patient’s name. Figure 2 elaborates on the detailed VR-based aphasia therapy process. The VR scene was deployed on mobile phones using the Google Cardboard VR plug-in. Combining Google Cardboard equipment Figure 2. Detailed VR-based aphasia therapy process. AI: artificial intelligence; VR: virtual reality. of the mobile-based VR speech rehabilitation app? All Pilot Application and Usability Testing of the App interviews were recorded using a digital voice recorder and We conducted feasibility exercises of the app on 5 patients with transcribed verbatim within 24 hours. The researcher explored aphasia, 5 healthy volunteers, 5 medical staff, and 2 VR experts participants’ feelings about the app by asking questions in a between 25 and 65 years of age (mean 39.53 years, SD 16.54 private environment, and they were encouraged to put forward years). All 5 patients with aphasia were hospitalized for speech new questions and content. rehabilitation for 2 weeks, and all of them had ischemic stroke. The aphasia quotient (AQ) of the Chinese version of the Western Results Aphasia Battery ranged from 58 to 75 after 2 weeks of face-to-face training. This feasibility exercise allowed us to Results of the Delphi Study better understand and potentially reduce the likelihood of Expert Panel adverse effects in patients with aphasia who may have been Table 3 lists the main characteristics of the 15 (71.4%) of 21 more sensitive to sensory effects (eg, motion sickness and experts who participated in the first and second Delphi rounds. discomfort). Interviews were the measurements used in the The 15 experts’ academic background included the medical experiments. The outline of interview questions after the science of rehabilitation or rehabilitation therapy. Of them, 14 participants’ experience with the mobile-based VR speech (93.3%) were from 3A hospitals, 5 (33.3%) of them took a rehabilitation app is as follows: (1) What groups do you think position in the university, and 2 (13.3%) of them were this mobile-based VR speech rehabilitation app is suitable for? professors. Detailed demographic characteristics of the Delphi (2) How does the mobile-based VR speech rehabilitation app panel are presented in Table 3. The reliability of the experts is differ from face-to-face language rehabilitation? (3) What would presented in Table 4. The mean value of the expert authority encourage or hinder app adoption after using the mobile-based coefficient (Cr) was 0.84. The authority of the experts was high; VR speech rehabilitation app? (4) What is your overall rating thus, the results of this study are trustworthy. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 5 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Table 3. Characterization of experts participating in the Delphi panel (N=15). Characteristic Participants, n (%) Gender Male 2 (13.3) Female 13 (86.7) Age (years) <35 6 (40.0) ≥35 9 (60.0) Education University diploma 9 (60.0) Master’s degree 4 (26.7) Doctoral degree 2 (13.3) Academic background Medical science of rehabilitation 5 (33.3) Rehabilitation Therapeutic 10 (66.7) Working experience (years) 5-9 5 (33.3) 10-14 5 (33.3) ≥15 5 (33.3) Main area of the work role Clinical 10 (66.7) Clinical and research 5 (33.3) Professional title Intermediate 12 (80.0) Deputy senior 1 (6.7) Senior 2 (13.3) https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 6 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al a b c Table 4. Reliability of experts (Ca , Cs , and Cr ). Expert number Criterion score Theoretical analysis Working experience Referring to literature Self-intuition Ca Cs Cr 1 0.4 0.3 0.1 0.1 0.9 0.7 0.8 2 0.5 0.3 0.1 0.1 1 0.7 0.85 3 0.5 0.3 0.05 0.05 0.9 0.5 0.7 4 0.4 0.3 0.15 0.1 0.95 0.9 0.925 5 0.5 0.3 0.1 0.05 0.95 0.9 0.925 6 0.4 0.3 0.1 0.05 0.85 0.9 0.875 7 0.4 0.3 0.15 0.1 0.95 0.7 0.825 8 0.4 0.3 0.15 0.1 0.95 0.7 0.825 9 0.4 0.3 0.1 0.1 0.9 0.7 0.8 10 0.4 0.3 0.1 0.1 0.9 0.9 0.9 11 0.4 0.2 0.05 0.1 0.75 0.7 0.725 12 0.5 0.2 0.05 0.1 0.85 0.7 0.775 13 0.4 0.3 0.1 0.1 0.9 0.9 0.9 14 0.4 0.3 0.1 0.1 0.9 0.9 0.9 15 0.5 0.3 0.15 0.1 1.05 0.9 0.975 Ca: judgment criteria the experts are based on. Cs: experts' familiarity with the research field. Cr: authority coefficient of experts. therapy was combined with oral expression therapy based on First Round the advice of 2 panel members. Furthermore, 2 panel members The Kendall coefficient of concordance (W) was 0.204 (P<.001). suggested that single characters could be deleted because they Of the 6 modules and 20 submodules rated by the panel are not as easy to express as phrases, so we deleted those. Some members in the first round (Table 5), the CV of 2 submodules respondents suggested additional executive functioning, reading, (syllabic and naming with verbs, nouns, and adjectives) and the and writing training. However, this app is meant to improve CV of the arithmetic therapy module that did not meet the patients' oral expression abilities, so reading and writing are criteria were removed (Table 6). In addition, 2 experts beyond our consideration, in particular, executive functioning recommended to delete the arithmetic therapy module. Pictures and writing. Due to smartphone VR and technological and videos describing daily life were added based on the limitations, we were unable to add writing training to VR. The recommendations of 3 panel members. In addition, naming revised content contained 4 modules and 17 submodules. Table 5. Modules in round 1. Module Task 1 Oral expression therapy 2 Auditory comprehension therapy 3 Cognition therapy 4 Naming therapy 5 Arithmetic therapy 6 Comprehensive application https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 7 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Table 6. Scores in round 1. Tasks Mean (SD) CV Oral expression therapy Read and repeat task: syllabic 4.13 (0.99) 0.24 Read and repeat task: single characters, including the 150 most commonly used Chinese characters 4.60 (0.91) 0.20 Read and repeat task: some phrases, including categories of digital, fruit, animals, vegetables, transportation, kitchen 5 (0) 0.00 supplies, daily necessities, body parts, food, address, location, and sports Read and repeat task: sentences, including proverbs and daily expressions 4.87 (0.35) 0.07 Answering some questions related to everyday life 4.80 (0.56) 0.12 Auditory comprehension therapy Listening and matching: included digital, color, pictures, and words 4.93 (0.26) 0.05 Yes-no questions 4.93 (0.26) 0.05 Listening to the passage and answering questions 4.73 (0.46) 0.10 Cognition therapy Attention training 4.27 (0.80) 0.19 Memory training 4.13 (0.74) 0.18 Reasoning and problem solving 4.47 (0.83) 0.19 Naming therapy Picture naming 4.93 (0.26) 0.05 Naming with verbs, nouns, and adjectives 4.33 (0.90) 0.21 Naming a list of items with the exact nature 4.93 (0.26) 0.05 Arithmetic therapy Addition, subtraction, multiplication, and division 4.27 (1.16) 0.27 Comprehensive application Supermarket task: A computer-generated virtual customer purchases goods. The patient has to pick items, count the 4.80 (0.56) 0.12 total amount of money that the virtual customer has selected, and return the change. Interview task: The patient plays the role of a job seeker. A computer-generated virtual interviewer asks questions about 4.73 (0.59) 0.13 the patient’s personal information (eg, name, gender, age, nationality, education level, birth date, height, weight, spe- cialty, address, and family members) and assesses their language skills by reading text and describing pictures. Bedroom task: The patient plays the role of a mother. By communicating with the computer-generated virtual daughter, 4.80 (0.56) 0.12 the patient selects clothes for the virtual daughter according to the weather conditions and discusses breakfast and what to choose for dinner and the kind of transportation to use. Ordering task: The patient acts like a customer to order in a virtual restaurant; communicates with a virtual server syn- 4.80 (0.41) 0.09 thesized by the computer; requests a certain number of dishes, desserts, cakes, and drinks; and completes the payment. Park task: By buying tickets, paying, and asking for directions in the virtual ticket office, together with a computer- 4.80 (0.41) 0.09 generated virtual friend, the patient enters a virtual park. The patient and the virtual friend communicate with each other about the scene. Likert scale: 1, not important; 2, somewhat important; 3, moderately important; 4 = important; and 5, very important. CV: coefficient of variation. CV>0.2. Deleted on the recommendations of the 2 panel members. suggested that patients should complete answering within the Second Round allotted time. Some respondents suggested additional executive The results of the second round are presented in Tables 7 and functioning, reading, and writing training. However, this app 8 and Figure 3. All mean scores for importance were above is meant to improve patients' oral expression abilities, so reading 4.00, and the CVs were less than 0.20. The Kendall coefficient and writing are beyond our consideration. The revised content of concordance (W) was 0.335 (P<.001). Finally, a consensus contained 4 modules and 17 submodules. was reached on 4 modules and 17 submodules. One expert https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 8 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Table 7. Modules in round 2. Module Task 1 Oral expression therapy 2 Auditory comprehension therapy 3 Cognition therapy 4 Comprehensive application Table 8. Scores in round 2. Tasks Mean (SD) CV Oral expression therapy Read and repeat task: some phrases, including categories of digital, fruit, animals, vegetables, transportation, kitchen 4.93 (0.26) 0.05 supplies, daily necessities, body parts, food, address, location, action, and sports Read and repeat task: sentences, including proverbs and daily expressions 5.00 (0) 0 Answering some questions related to everyday life 4.87 (0.35) 0.07 Naming task: picture naming 5.00 (0) 0 Naming task: naming list of items with the exact nature 4.93 (0.26) 0.05 Describing pictures and videos pertaining to daily life 4.67 (0.49) 0.10 Auditory comprehension therapy Listening and matching: included digital, color, pictures, and words 5.00 (0) 0 Yes-no questions 4.93 (0.26) 0.05 Listening to the passage and answering questions 4.80 (0.41) 0.09 Cognition therapy Attention training 4.27 (0.70) 0.16 Memory training 4.13 (0.74) 0.18 Reasoning and problem solving 4.20 (0.77) 0.18 Comprehensive application Supermarket task: A computer-generated virtual customer purchases goods. The patient has to pick items, count the 4.80 (0.56) 0.12 total amount of money that the virtual customer has selected, and return the change. Interview task: The patient plays the role of a job seeker. A computer-generated virtual interviewer asks questions 4.80 (0.56) 0.12 about the patient’s personal information (eg, name, gender, age, nationality, education level, birth date, height, weight, specialty, address, and family members) and assesses their language skills by reading text and describing pictures. Bedroom task: The patient plays the role of a mother. By communicating with the computer-generated virtual 4.87 (0.52) 0.11 daughter, the patient selects clothes for the virtual daughter according to the weather conditions and discusses breakfast and what to choose for dinner and the kind of transportation to use. Ordering task: The patient acts like a customer to order in a virtual restaurant; communicates with a virtual server 4.87 (0.52) 0.11 synthesized by the computer; requests a certain number of dishes, desserts, cakes, and drinks; and completes the payment. Park task: By buying tickets, paying, and asking for directions in the virtual ticket office, together with a computer- 4.53 (0.64) 0.14 generated virtual friend, the patient enters a virtual park. The patient and the virtual friend communicate with each other about the scene. CV: coefficient of variation. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 9 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Figure 3. Consensus on 4 modules. In response to question 1, “What groups do you think this Usability Testing of the App mobile-based VR speech rehabilitation app is suitable for?”, 12 All 17 (100%) participants could open the app. A user-centered (70.6%) of 17 participants suggested that the user group could design philosophy was implemented to create additional include patients who were discharged or currently hospitalized. interactive and customizable features that indicated a high Participants stated that with this app, they could correct and degree of usability for patients with aphasia. Participants' refine their language skills, and 5 (29.4%) of 17 participants comments were collected and descriptively analyzed based on recommended that the user group should include people, the interview outline. The 4 responses were classified based on especially those who are under significant financial pressure their feedback. Figure 4 shows the functionality of the VR-based and do not have easy access to medical treatment but would aphasia therapy app. prefer an opportunity to learn. Figure 4. Functionality of the VR-based aphasia therapy app. VR: virtual reality. In response to question 2, “How does the mobile-based VR difficulty of going to the hospital and is more convenient as speech rehabilitation app differ from face-to-face language they do not have to worry about weather and traffic. In addition, training?”, all 17 (100%) participants mentioned the advantages 15 (88.2%) of 17 participants reported a high level of immersion of in-home apps. Participants stated that this app reduces the and engagement and thought that patients could thoroughly https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 10 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al enjoy themselves when using this app without judgment from in patients with poststroke aphasia influences language others. Additionally, there is no need to worry about rehabilitation outcomes [33]. self-abasement because of the inability to speak fluently. Language rehabilitation combined with cognitive rehabilitation Moreover, 4 (23.5%) of 17 participants critiqued the can improve patients' communication skills [34]. The cognitive effectiveness of this app because patients could not receive domain includes processing, attention, working memory, social immediate feedback, and the efficiency of using this app could cognition, reasoning, and problem-solving [7]. Therefore, to not be guaranteed. improve the patients’ oral expression abilities, we designed In response to question 3, “After using the mobile-based VR cognitive training pertaining to reasoning, memory, attention, speech rehabilitation app, what would encourage or hinder your visuospatial perception, and logical thinking. A range of brain adoption of the app?”, 15 (88.2%) of 17 participants mentioned regions are related to cognition, and these nonlanguage domains that the design and arrangement met patients’ daily play a role in the ability of patients with chronic aphasia [29]. communication needs; they felt more secure because the training To emphasize the communication demands of an individual's content was designed by a professional rehabilitation team. In life, the contextual domain is placed at the top of the cognitive addition, 15 (88.2%) participants mentioned that the app was communication competence model [7]. A Cochrane review easy to use, and 3 (17.6%) reported low levels of motion concluded that communication activity in real-world settings, sickness. or functional communication, could improve communicative In response to question 4, “What is your overall rating of the ability [1]. Communication practices should occur in an mobile-based VR speech rehabilitation app?”, all 17 (100%) ecological context. Therefore, we designed 5 virtual scenario participants indicated that the app was novel, and 5 (29.4%) of tasks focusing on everyday communication activities and 17 participants gave some recommendations, such as considering pragmatic conversational skills. Patients may receive language different training difficulties and increased functionality. All practice opportunities in natural context communication settings patients expressed interest in language rehabilitation using the in VR. These tasks are a comprehensive practice for the patients, new technology. and different short communication dialogues of everyday activities and cognitive exercises were integrated, including Discussion oral expression, naming, calculation, attention, memory, and reasoning tasks. VR technology is a promising rehabilitation Principal Findings tool and may be a useful alternative to conventional training [35], but the scene in VR is inaccurate. However, we could This study developed the first accessible and cost-effective create almost everything in VR, and patients experience it as mobile-based VR speech rehabilitation app in China. Through an actual situation, which makes it easy to try new therapeutic a 2-round Delphi study, panel members reached a consensus strategies. All learning in VR can be transferred to the real world on 4 modules and 17 submodules that benefit patients in gaining [12]. With clinician shortages and a higher prevalence of essential daily communication language skills and various skills aphasia, VR is not meant to replace skilled therapists but to ease in everyday life. Patients get opportunities to interact with the clinicians' burdens. virtual environment and can practice at any time. A review systematically identified and evaluated a series of mobile apps Limitations for speech-language therapy and found a lack of interactive and This study had some limitations. First, a range of brain regions engaging elements in the apps, which failed to make patients are related to attention and executive functioning, and these self-manage [11]. The content we designed based on VR nonlanguage domains play a role in the abilities of patients with emphasizes user experience, engagement, and visual appeal to chronic aphasia [29]. Owing to the limitations of intelligent VR, improve patient adherence. We developed real-life scenarios we cannot design some daily instructions (eg, washing clothes, and gaming factors for the VR environment. Patients get cutting vegetables, cooking). VR could be used to provide an opportunities to interact with a virtual environment to practice enriched environment where patients could master skills related language in natural context communication settings in VR. We to daily life that cannot be finished in the hospital. Learning in believe that the rehabilitation content will provide patients with VR can then be transferred to real-life situations. Second, a strong basis for returning to society and work. dialects vary across different areas of China. However, we did We developed the content by learning from a model of cognitive not involve dialects from different regions of China to provide communication competence [7]. Language processing is the options for those who have difficulty understanding and core of cognition and requires the participation of other expressing Mandarin. There are nonlanguage abilities that would nonlinguistic cognitive functions [25,26]. Left hemispheric be worthwhile to consider in aphasia rehabilitation, such as stroke leading to impairments in language processing often prosody and emotional tone in utterances. However, it is difficult affects other cognitive functions, such as executive function, to implement these in the app. In addition, the usability of the attention, visuospatial perception, logical thinking, and memory app was assessed by only 17 individuals, which may have biased [27,28], which play an essential role in aphasia recovery and the usability results. In the future, we will design a more rehabilitation [29-31]. Patients' language and nonlinguistic significant multicenter trial with a longer-term follow-up. cognitive functions cannot be separated owing to complex Additionally, this app is meant to improve patients' oral neurobiological networks [32]. Nonlinguistic cognitive function expression abilities, so reading and writing are beyond our https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 11 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al consideration. In particular, owing to the current limitations of significant potential to make our app an in-home app in the technology, we cannot plug writing training into VR. future, contributing to the automation of rehabilitation administration. However, further studies are needed to evaluate Conclusion the feasibility and efficacy of our app. We hope that our research We conducted a Delphi study and developed a mobile-based provides guidelines and references for others in the medical VR app. This study constitutes a step toward the development field. of a combination of health and VR. We believe there is Acknowledgments We gratefully thank the experts for completing the study surveys. This research was supported by the Fundamental Research Funds for the Central Universities of Central South University (project no. 2020zzts847) and the Hunan Provincial Health Commission (project nos. 2020SK51104 and 202114021494). Authors' Contributions All authors listed in this publication conducted the research. XB was responsible for all stages of the project. XB, PHFN, QC, YT, RF, QT, QC, SL, ASKC, and XL participated in the design, analysis, and writing of the article. PHFN and QC provided expertise in the development of the app. PHFN, QC, and ASKC revised the article. All authors approved the final draft. Conflicts of Interest None declared. References 1. Brady M, Kelly H, Godwin J, Enderby P, Campbell P. Speech and language therapy for aphasia following stroke. Cochrane DB Syst Rev 2020 Jun 01:2016. [doi: 10.1002/14651858.cd000425.pub4] 2. Menon DK, Bryant C. Time for change in acquired brain injury. Lancet Neurol 2019 Jan;18(1):28. [doi: 10.1016/s1474-4422(18)30463-0] 3. Wu S, Wu B, Liu M, Chen Z, Wang W, Anderson CS, et al. Stroke in China: advances and challenges in epidemiology, prevention, and management. Lancet Neurol 2019 Apr;18(4):394-405. [doi: 10.1016/s1474-4422(18)30500-3] 4. Strong K, Mathers C, Bonita R. Preventing stroke: saving lives around the world. Lancet Neurol 2007 Feb;6(2):182-187. [doi: 10.1016/s1474-4422(07)70031-5] 5. Graham JR, Pereira S, Teasell R. Aphasia and return to work in younger stroke survivors. Aphasiology 2011 May 25;25(8):952-960. [doi: 10.1080/02687038.2011.563861] 6. Béjot Y, Bailly H, Durier J, Giroud M. Epidemiology of stroke in Europe and trends for the 21st century. Presse Med 2016 Dec;45(12 Pt 2):e391-e398. [doi: 10.1016/j.lpm.2016.10.003] [Medline: 27816343] 7. MacDonald S. Introducing the model of cognitive-communication competence: a model to guide evidence-based communication interventions after brain injury. Brain Inj 2017 Oct 24;31(13-14):1760-1780. [doi: 10.1080/02699052.2017.1379613] [Medline: 29064304] 8. Lavoie M, Macoir J, Bier N. Effectiveness of technologies in the treatment of post-stroke anomia: a systematic review. J Commun Disord 2017 Jan;65:43-53. [doi: 10.1016/j.jcomdis.2017.01.001] [Medline: 28171741] 9. Kurland J, Liu A, Stokes P. Effects of a tablet-based home practice program with telepractice on treatment outcomes in chronic aphasia. J Speech Lang Hear Res 2018 May 17;61(5):1140-1156. [doi: 10.1044/2018_jslhr-l-17-0277] 10. Lavoie M, Routhier S, Légaré A, Macoir J. Treatment of verb anomia in aphasia: efficacy of self-administered therapy using a smart tablet. Neurocase 2016 May 26;22(1):109-118. [doi: 10.1080/13554794.2015.1051055] [Medline: 26007615] 11. Vaezipour A, Campbell J, Theodoros D, Russell T. Mobile apps for speech-language therapy in adults with communication disorders: review of content and quality. JMIR Mhealth Uhealth 2020 Oct 29;8(10):e18858. [doi: 10.2196/18858] 12. Freeman D, Reeve S, Robinson A, Ehlers A, Clark D, Spanlang B, et al. Virtual reality in the assessment, understanding, and treatment of mental health disorders. Psychol Med 2017 Mar 22;47(14):2393-2400. [doi: 10.1017/s003329171700040x] 13. Marshall J, Booth T, Devane N, Galliers J, Greenwood H, Hilari K, et al. Evaluating the benefits of aphasia intervention delivered in virtual reality: results of a quasi-randomised study. PLoS One 2016 Aug 12;11(8):e0160381 [FREE Full text] [doi: 10.1371/journal.pone.0160381] [Medline: 27518188] 14. Zhang Y, Chen P, Li X. Clinical research on therapeutic effect of virtual reality technology on Broca aphasia patients. 2009 Presented at: Proceedings of the International Conference on Information Technology; 2017; Ioannina, Epirus, Greece p. 134. [doi: 10.1109/incit.2017.8257880] 15. Keeney S, Hasson F, McKenna HP. A critical review of the Delphi technique as a research methodology for nursing. Int J Nurs Stud 2001 Apr;38(2):195-200. [doi: 10.1016/s0020-7489(00)00044-4] 16. Powell C. The Delphi technique: myths and realities. J Adv Nurs 2003 Feb;41(4):376-382. [doi: 10.1046/j.1365-2648.2003.02537.x] [Medline: 12581103] https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 12 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al 17. Mullen PM. Delphi: myths and reality. J Health Organ Manag 2003 Feb;17(1):37-52. [doi: 10.1108/14777260310469319] [Medline: 12800279] 18. Brain Health, for Life. URL: https://constanttherapyhealth.com/ [accessed 2022-04-01] 19. Therapy Menu. URL: http://www.aphasiatherapyonline.com/main.html [accessed 2022-04-01] 20. Reed GF, Lynn F, Meade BD. Use of coefficient of variation in assessing variability of quantitative assays. Clin Diagn Lab Immunol 2002 Nov;9(6):1235-1239 [FREE Full text] [doi: 10.1128/cdli.9.6.1235-1239.2002] [Medline: 12414755] 21. Chen Y, Li Z, Peng J, Shen L, Shi J. Development of a care bundle for stroke survivors with psychological symptoms: evidence summary and Delphi study. Evid Based Complement Alternat Med 2020 Jun 30;2020:7836024 [FREE Full text] [doi: 10.1155/2020/7836024] [Medline: 32714421] 22. Xing Q, Zhang M, Zhao F, Zhou Y, Mo Y, Yuan L. The development of a standardized framework for primary nurse specialists in diabetes care in China: a Delphi study. J Nurs Res 2019 Dec;27(6):e53. [doi: 10.1097/jnr.0000000000000330] [Medline: 31162295] 23. Ren H, Liu C, Wang R, Zhang M, Ma F, Li R, et al. Core competencies required for gastroenterology nursing specialists in China. Gastroenterol Nurs 2019;42(2):169-178. [doi: 10.1097/sga.0000000000000392] 24. Shen J, Xiang H, Luna J, Grishchenko A, Patterson J, Strouse RV, et al. Virtual reality-based executive function rehabilitation system for children with traumatic brain injury: design and usability study. JMIR Serious Games 2020 Aug 25;8(3):e16947 [FREE Full text] [doi: 10.2196/16947] [Medline: 32447275] 25. Ardila A, Rubio-Bruno S. Aphasia from the inside: the cognitive world of the aphasic patient. Appl Neuropsychol Adult 2018 May 23;25(5):434-440. [doi: 10.1080/23279095.2017.1323753] [Medline: 28535071] 26. Peach R. Cognitive approaches to aphasia treatment: application of the cognition of language to aphasia intervention. Semin Speech Lang 2017 Feb 15;38(1):3-4. [doi: 10.1055/s-0036-1597259] [Medline: 28201832] 27. Murray LL. Attention and other cognitive deficits in aphasia: presence and relation to language and communication measures. Am J Speech Lang Pathol 2012 May;21(2):S51-S64. [doi: 10.1044/1058-0360(2012/11-0067)] 28. Yao J, Liu X, Liu Q, Wang J, Ye N, Lu X, et al. Characteristics of non-linguistic cognitive impairment in post-stroke aphasia patients. Front Neurol 2020 Sep 30;11:1038 [FREE Full text] [doi: 10.3389/fneur.2020.01038] [Medline: 33117251] 29. Schumacher R, Halai AD, Lambon Ralph MA. Assessing and mapping language, attention and executive multidimensional deficits in stroke aphasia. Brain 2019 Oct 01;142(10):3202-3216 [FREE Full text] [doi: 10.1093/brain/awz258] [Medline: 31504247] 30. El Hachioui H, Visch-Brink EG, Lingsma HF, van de Sandt-Koenderman MWME, Dippel DWJ, Koudstaal PJ, et al. Nonlinguistic cognitive impairment in poststroke aphasia: a prospective study. Neurorehabil Neural Repair 2014 Nov 08;28(3):273-281. [doi: 10.1177/1545968313508467] [Medline: 24213963] 31. Brownsett SLE, Warren JE, Geranmayeh F, Woodhead Z, Leech R, Wise RJS. Cognitive control and its impact on recovery from aphasic stroke. Brain 2014 Jan;137(Pt 1):242-254 [FREE Full text] [doi: 10.1093/brain/awt289] [Medline: 24163248] 32. Hagoort P. The core and beyond in the language-ready brain. Neurosci Biobehav Rev 2017 Oct;81(Pt B):194-204. [doi: 10.1016/j.neubiorev.2017.01.048] [Medline: 28193452] 33. Gilmore N, Meier EL, Johnson JP, Kiran S. Nonlinguistic cognitive factors predict treatment-induced recovery in chronic poststroke aphasia. Arch Phys Med Rehabil 2019 Jul;100(7):1251-1258 [FREE Full text] [doi: 10.1016/j.apmr.2018.12.024] [Medline: 30639272] 34. Zakariás L, Keresztes A, Marton K, Wartenburger I. Positive effects of a computerised working memory and executive function training on sentence comprehension in aphasia. Neuropsychol Rehabil 2018 Apr 21;28(3):369-386. [doi: 10.1080/09602011.2016.1159579] [Medline: 26999324] 35. Dockx K, Bekkers E, Van DBV. Virtual reality for rehabilitation in Parkinson's disease. Cochrane DB Syst Rev 2016:CD010760. [doi: 10.1002/14651858.cd010760] Abbreviations ABI: acquired brain injury AI: artificial intelligence CV: coefficient of variation VR: virtual reality https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 13 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Bu et al Edited by N Zary; submitted 23.05.21; peer-reviewed by T Ong, G Cárdenas-López; comments to author 03.12.21; revised version received 28.02.22; accepted 05.03.22; published 07.04.22 Please cite as: Bu X, Ng PHF, Tong Y, Chen PQ, Fan R, Tang Q, Cheng Q, Li S, Cheng ASK, Liu X A Mobile-based Virtual Reality Speech Rehabilitation App for Patients With Aphasia After Stroke: Development and Pilot Usability Study JMIR Serious Games 2022;10(2):e30196 URL: https://games.jmir.org/2022/2/e30196 doi: 10.2196/30196 PMID: ©Xiaofan Bu, Peter HF Ng, Ying Tong, Peter Q Chen, Rongrong Fan, Qingping Tang, Qinqin Cheng, Shuangshuang Li, Andy SK Cheng, Xiangyu Liu. Originally published in JMIR Serious Games (https://games.jmir.org), 07.04.2022. 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 JMIR Serious Games, is properly cited. The complete bibliographic information, a link to the original publication on https://games.jmir.org, as well as this copyright and license information must be included. https://games.jmir.org/2022/2/e30196 JMIR Serious Games 2022 | vol. 10 | iss. 2 | e30196 | p. 14 (page number not for citation purposes) XSL FO RenderX

Journal

JMIR Serious GamesJMIR Publications

Published: Apr 7, 2022

Keywords: virtual reality; speech rehabilitation; stroke; app; Delphi

There are no references for this article.