Course:COGS200/2017W1/Group8FinalProject
Emotionally / Non-Emotionally Responsive Haptic Gaming for Post-Stroke Upper Arm Rehabilitation in Conjunction with Traditional Rehabilitation.
Introduction
Our study will look into how emotionally responsive virtual reality games could be used in collaboration with conventional stroke therapy in order to increase upper-arm strength and motor functions in stroke patients. According to the Ontario Stroke Network, about 426,000 Canadians are currently living with the effects of a stroke. Every year, nearly 14,000 will die due to stroke. As the leading cause of adult disability in Canada, it is important to develop better stroke rehabilitation methods that ensure that stroke survivors are able to gain back as much mobility and motor function as possible. Factors such as slow access to rehabilitation facilities, healthcare costs and the tedious and non-interactive rehab methods are contributing factors to the sometimes lackluster results from these therapies (Wang et al., 2014). Annually, strokes cost the Canadian economy 3.6 billion dollars (www.ontariostrokenetwork.ca). Faster and more effective rehabilitation contributes to stroke survivors being able to rejoin the workforce faster and care for their families to the best of their abilities. Designing more interactive and effective treatment for stroke patients can gain back some of the loses from lost wages and decreased productivity, while simultaneously resulting in an overall better quality of life for stroke survivors (Laver et al., 2012).
When it comes to stroke recovery, upper-extremity recovery is generally less successful than lower-limb recovery. Upper body function deficiencies persist in 55 to 75% of post-stroke patients. In comparison, 75% of patients learn to walk again after stroke (Meriens et. al., 2002), therefore more research on upper-extremity rehabilitation is necessary. The successful rehabilitation of upper arm function in stroke patients depends on both relearning motor skills and regaining the necessary neural pathways for such motor functions (Jack et al. 2001). Valuable training sessions are essential to achieving these goals, with the intensity, length, and quantity of these sessions showing direct relation to amount of functional improvement. In traditional rehabilitation settings, professionals will work with hospitalized patients for 30 minute sessions once or twice a day, while outpatients attend sessions once or twice a week (Jack et al. 2001). The use of robotics and VR programs allows patients to receive valuable feedback on their progress, as well as ensuring they have proper form for their exercises. This feedback and form advice can give patients high quality training sessions that can supplement their working with a specialist. This will both allow patients to receive more training time and allow specialists to spend more time with stroke patients who require additional support (Jack et al. 2001) (Hou, Xiyuan and Sourina 2013).
A positive rehabilitation environment correlates with stroke patients being more motivated to be rehabilitated (Maclean et al., 2000) Motivation is a complex phenomenon that changes over time within a single participant (due to factors such as depression or fatigue) and differs among participants (due to outside support or individual characteristics) (Colombo et al., 2007). While specialists are unable to directly affect many of these motivation factors, different types of rehabilitation can directly address the engagement of participants, which is linked to higher motivation. VR technologies create immersive environments conducive to increased motivation, that in turn leads to increased improvements in motor function (Jack et al. 2001). In addition, robotic technologies can provide real time, quantitative data pertaining to patients' improvement. This direct feedback is reinforcing to patients as they can concretely observe their own progress, leading to higher motivation and adherence to rehabilitation programs (Colombo et al., 2007). Using a mechanism that provides real time biofeedback will allow us to monitor the frustration levels of stroke patients in order to ensure that stress levels do not impede with rehabilitation.
Hypothesis
Our proposed research project will address the use of emotionally and non-emotionally responsive haptic gaming for upper arm recovery after stroke. We hypothesize that the use of an emotionally responsive robotic glove used in conjunction with an immersive game will improve upper limb control in stroke patients, more than a non-emotionally responsive game, and in turn, the use of a robotic glove and non-emotionally responsive game will lead to increased improvement of upper limb control in stroke patients more than in stroke patients using traditional rehabilitation techniques.
Methods
General Study Design
We we will use the parallel random clinical trials (RCTs) study design. RCTs involve randomly assigning participants to the treatment levels in a study. It is thought to reduce potential bias or causality associated with other experimental designs, which may lead to influencing a study’s results. In addition, the parallel RTCs design means once participants are randomly allotted to our three treatment levels, participants will carry out their rehabilitation in a parallel fashion with the other treatment levels (Ahn and Ahn, 2010;2015). A summary of this design can be seen in Figure 1.
Participant Selection Criteria
Participants will be chosen based on age, type of stroke, time since stroke occurrence, and whether this is their first stroke. These are important considerations as, in order to get as accurate a set of results as possible from our experiment, we will need patients who are more or less at the same stage of recovery. Our age range will be between the ages of 50 and 70. While the age range of 55 to 65 is when most people are likely to suffer a stroke, we have widened that age range to ensure we can get enough adequate participants. While fourteen days is considered a critical period for stroke rehabilitation (Biernaske et al., 2004), we have decided to use patients that have had a stroke within the last three weeks, as we feel two weeks is too short of a time frame to organize participants in our study.
As described by Raghavan (2004), there are three main types of upper-arm impairment that occur after a major stroke. These are learned nonuse, learned bad-use and forgetting as determined by behavioural analysis of tasks. In our study, we will be focusing on learned nonuse. This arises from patients not using the affected limb after the stroke, and can be caused by factors such as limb weakness, paralysis, or sensory loss in the limb. Of course, initially this nonuse of the limb is not voluntary, but may become habitual and the limb may be excluded from functional activities. This learned behaviour is why it is described as learned nonuse (Raghavan, 2004). The effect on motor functions could mean that the patient is not able to do simple tasks such as grasping a pen, holding a cup or folding a napkin. By focusing on major strokes which led to learned nonuse, our experiment is more likely to encompass the majority of the stroke spectrum. We aim to develop a treatment that can assist survivors of severe strokes to gain back as much of their fine motor skills as possible. By extending this treatment to activities which can be done at home, patients will be more likely to continue with it and gradually gain back normal motor function.
With ideal funding and ideal participant availability, our target sample size will be 60 patients (Bath, 2008). Sample size mainly depends on appropriate level of significance, the power of the study, expected effect size, event rate in population, standard deviation in the population. Other factors such as expected drop out rate, ratio of patients in each treatment level, as well as specific objectives of a study may influence the sample size (Kadam and Bhalerao 2010). Sample size is critical for scientific research, as too small a sample size can lead to false-negative findings (also called a Type 2 error, where the null hypothesis is incorrectly rejected).This kind of error can be incredibly harmful for scientific research because a treatment could have significant positive effects, but due to incorrect sample size, this is not detected (Biau, Kernéis and Porcher, 2008). For our study, we will accept a p-value of 0.05 or less (5, or less, out of 100 times a difference will be detected when no difference actually exists), a power or Type 2 error of 80% (1 in 5 times we will miss a difference when a difference exists). Other variables in sample size calculation will be based on accepted standards found in stroke rehabilitation research.
Equipment
A prototype consisting of a haptic glove, EEG sensors and a computer game will be developed. The computer game will serve as the virtual environment in which patients carry out rehabilitative exercises through the use of the haptic device. EEG sensors will capture feedback on the patient’s emotional state throughout gameplay so that difficulty levels can be adjusted accordingly in real-time.
The haptic glove will be the main medium of interaction and serves two main functions. Firstly, the haptic device inputs coordinates from the physical space onto the virtual space with the use of accelorometers and gyroscopes; thus, a simple turning of the wrist in mid-air in real space can be translated into the (successful or unsuccessful) unscrewing of a jar lid or the turning of a doorknob in the virtual world. Secondly, the device will simulate properties of actual objects, such as shape, size, rigidity and weight by providing kinesthetic feedback. By adjusting this feedback, an object in the virtual environment becomes lighter or heavier and smaller or larger. In turn, the object’s manoeuvrability will require different levels of motor control capacity.
Haptic rendering of objects in a virtual environment is achieved through kinesthetic and cutaneous feedback (Dahiya & Valle, 2013). Kinesthetic feedback is analogous to proprioception and provides information on object weight and shape while cutaneous feedback provides information such as temperature and moistness. Some sensations, such as viscosity or stickiness, arise from the overlap of kinesthetic and cutaneous feedback. Haptic rendering with the highest fidelity would incorporate both types of feedback.
In line with our goal of developing a user-friendly and portable system that can one day be used by outpatients in the comfort of their homes, the haptic glove will be wearable and lightweight. Possible devices to use for the study include the Grabity and Wolverine gloves, both developed by the Stanford University’s SHAPE Laboratory or the Dexta Robotic’s Dexmo Exoskeleton. Subject to available resources and collaboration, it may be possible to develop a new prototype tailored to the needs of stroke patients as ungrounded haptics (i.e. portable haptics) is a nascent technology and current devices target different aspects of haptic rendering (Choi et al., 2017).
EEG
In our prototype, the electroencephalogram (EEG) system will collect provide input the patient’s emotional state. This information will be used to adjust the difficulty level of the game in order to maintain the patient’s motivation throughout the treatment duration.
Our focus will be on the movement of the brain wave frequencies, especially alpha and beta waves, as these have particular implications for virtual reality games and video games (Berk, 2009). Alpha waves will occur when the patient is in a relaxed state of mind, and relaxed sense of awareness. Increased alpha waves are conducive for learning, as they relax the brain and allow for content to pass into long-term memory (Millbower, 2000). On the other hand, beta waves will occur when the participant is at a heightened sense of awareness and are highly alert. They are more likely to cause the participant to remember negative events in the past or think about them happening in the future increase. So, for example, if the EEG detects a higher level of alpha waves, the difficulty level of the game will be increased. When it detects more beta waves (denoting negativity or frustration), the difficulty of the game will be lowered. In our virtual game, these difficulty levels will mean different things.These may include increasing or decreasing the weight of an object, making a button easier or more difficult to push or making an object closer or farther to reach.
A single human emotion can be exhibited by a variety of physiological symptoms. Therefore, a high fidelity system of emotion capture should employ multimodal sensors, such as those that capture the heart rate and skin temperature (Mihelj et al., 2009). However, we propose using a simple EEG system in our prototype for two reasons. Firstly, we wish to create a user-friendly and cost-effective system. EEG devices are relatively ubiquitous and affordable and in fact, there already exist commercially available devices, such as the Emotiv system, that are pre-programmed to discriminate emotional states such as frustration, engagement, interest, etc. Given the age range of the participation group, there is also concern that excessive use of equipment would impact patient comfort. Secondly, multimodal information leads to complexity in mapping emotion based on physiological response as the number of input variables increase. Since EEG has been successfully utilized in numerous brain-computer interface (BCI) studies, including those on algorithm development (see Hou & Sourina, 2013; Reuderink & al., 2013; Sourina & Liu, 2011; Yoon & Chung, 2011), we consider this to be a well-established technology.
It can be difficult to identify specific emotions such as frustration or boredom based solely on physiological features, so we intend to utilize the two-dimensional arousal-valence model, which provides a descriptively maps brain activity with emotional states without explicitly labelling the emotions. Arousal describes the level of brain activity (e.g. concentration as opposed to being relaxed) while valence measures overall positivity or negativity of the mood (Mihelj et al., 2009; Sourina & Liu, 2011). We believe this model to be well-studied. Along with its more nuanced three-dimensional version, this model was prevalent across our literature review.
We will process EEG activity using the fractal-based algorithm developed by Sourina & Liu (2011). The algorithm yielded an accuracy rate of over 70% with only three electrodes. The minimalist equipment works towards our goal of a cost-effective and user-friendly prototype. Sourina & Liu mapped different arousal-valence values to fear (negative high arousal), sadness (negative low arousal), happiness (positive high arousal) and a general sense of pleasantness (positive low arousal).
In their 2013 study of a VR rehabilitative system, Hou and Sourina developed a model to determine difficulty levels based on emotional feedback and scores from the completion of rehabilitative tasks: if a low score is accompanied by a negative emotion, the difficulty is reduced; if a high score is accompanied by a positive emotion, the difficulty is increased; otherwise, difficulty remains unchanged. We will adapt this model for our study in the following manner as seen in Table1.
Arousal-valence states will fluctuate over the course of gameplay. States that persist for over sixty seconds will be used to determine if the difficulty level requires adjustment. This prevents over-adjustment for interim state. If the need to adjust a difficulty level arises when the patient is is performing a rehabilitative movement, difficulty will be adjusted by changing the kinesthetic feedback. If, however, the need to adjust a difficulty level arises when the patient is simply exploring the virtual environment looking for clue, the adjustment will arise as a form of hint on-screen.
Immersive game description
For this study, we have developed a computer game to use in conjunction with the haptic glove. The computer game will simulate a detective trying to solve a mystery. The game will have 10 levels, starting at level 1 and with level 10 requiring the most coordination and strength to reach. We anticipate that participants will reach level 7 by the end of the study period, but the extra three levels have been designed just in case participant’s exceed our expectations. These levels will test four aspects of upper arm motor function including range, speed, fractionation, and strength. Range involves finger extension and flexing, speed is the rate at which a participant can close their hand starting from an open palm position, fractionation entails bringing one fingertip to the palm while the other fingers stay extended, and strength will involve a participants grasping strength (Jack et al. 2001). Each level will include exercises relating to these four aspects, but increased levels will make these exercises more difficult.The participant will be required to repeat these exercises an x number of times, to practice the motion. We have provided examples of what these exercises will look like in the game for the first level, additional levels will have variations of these exercises at higher difficulties and in different situations to keep the game engaging.
The first level of the game involves a detective (the participant) trying to find a missing person who has purportedly invented a time machine. A range of motion exercises will be required when the participant must catch fireflies (requiring a fist) to put in a jar to exchange with a lab technician. The lab technician will then allow the detective to have access to a box belonging to the inventor. The box will contain dusty papers that the participant will have to wipe off to read (requiring a wide open hand). Speed will be required when the detective begins to follow a person of interest, and a piece of paper will fly out of the person of interest’s book with the participant having to catch the paper. Once caught, the paper will reveal a clue. Fractionation will be needed when the detective must communicate in sign language with a deaf shop keeper who has valuable information. The game will have subtitles for the deaf person’s message and will show the participant what hand gestures they need to make. Grasping will be needed when the detective enters a room that has been ransacked. The detective will need to lift up and put down objects of various weight and sizes to reveal clues.
In addition to key actions listed above that must be performed in order to drive the plot forward, or move to the next chapter of the story, the patient will have to complete a fixed number of minigames involving the repetitive motion prescribed by their therapist. These minigames would be independent of the plot and would simply replicate the type of repetitive exercise that would otherwise be accomplished in the physical world. For example, a minigame may involve placing a water bottle in the palm and lifting the palm repeatedly towards the wrist ten times. Another example would involve pinching a pen between the thumb and forefinger and sliding it across a surface repeatedly. As with other game mechanics, the difficulty will be adjusted the in real time. The water bottle, for example, can be made to be heavier or lighter.
Study Design
For our study, we will have a control and two other treatment levels. All participants selected for our study will participate in 30% additional hours of stroke rehabilitation treatment, as compared to traditional upper arm stroke rehabilitation. This ensures that stroke rehabilitation will not be hindered by participation in our study. Those in the control/Treatment #1 will participate in 30% additional hours of conventional upper arm rehabilitation exercises. Those in Treatment #2 will have 30% additional hours participating in the immersive game with the robotic glove. Those in Treatment #3 will have 30% additional hours with the emotionally responsive immersive game and robotic glove. A summary of this can be found in Table 2.
According to the American Stroke Association, some stroke survivors continue their process of recovery well into their first or second year post-stroke. However, the critical period for stroke recovery occurs within the first 3 months of stroke onset (Bernhardt et al. 2017). As a means of monitoring the results of our patients as much as possible, whilst considering our available time and financial resources, we will conduct our research over a period of three months. Participants will spend three hours a week doing their exercises in a clinical setting. This will be in addition to standard stroke rehabilitation (not focused on upper arm rehabilitation) involving conventional occupational therapy related to psychological, speech and physical therapies. Participants will receive a check-up after every 3 weeks of treatment, as described under the “Motor Function Assessment”. After the three months of rehabilitation, there will be an additional follow up check up 3 months after the end of the study.
Control/Treatment Levels
Control/Treatment #1
In this control, patients will participate in traditional upper arm stroke rehabilitation. This will involve conventional occupational therapy. This can be split into four categories:
Physical therapy- We would employ the Bobath technique to stroke rehabilitation, which is one of the most widely used physical therapies for stroke patients as well as patients of cerebral palsy (Cathrin Butefisch, Horst Hummelsheim, Petra Denzler, Karl-Heinz Mauritz, 1994). The technique centres around posture and how correcting the posture first can lead to improved motor functioning. It emphasizes that patients need to be active while the therapist assists them to move by using key points of control and reflex-inhibiting patterns (Bobath, 1990).
Occupational therapy- Strueltjens et al., (2003) describes occupational therapy as treatment that “[facilitates] task performance by improving relevant performing skills or developing and teaching compensatory strategies to overcome lost performance skills”. This will be employed to ensure that patients can gain back self-care and leisure, as well as where their sense of independence will be boosted the most.
Speech therapy- Aphasia, which is a disorder of communication which impairs a person’s ability to use and comprehend language (www.stroke.org) affects about one-third of all stroke patients (Lincoln et al., 1984), which is why speech therapy is an important part of stroke rehabilitation. This therapy will help to improve language skills as well as, if necessary, develop alternative means of communication for the patients.
Psychological therapies- Lastly, the patients will also see a psychologist in order to ensure that their psychological and emotional wellbeing are monitored during the recovery process. Kneebone & Jeffries (2013) discuss how disorders such as depression and anxiety can begin after a stroke, and so access to a psychologist will be essential in ensuring that the mental wellbeing of the patients is sufficient to handle the physical demands of the other therapies
Treatment Level #2
Patients in this group will receive the same conventional treatment as described in Treatment Level #1. However, their additional hours of therapy will be spent participating in the VR game as described in the methods section. In this condition, the difficulty level is fixed at the beginning of each session (as easy, normal, or difficult), as per the advice of the patient’s therapist. No emotional responses will be collected from the patient, and thus progression to the next stage of the game or completion of the game will not be automatically adjusted by their emotional responses. The game will have set variables as to what patients need to complete to advance to the next levels, such as set object weights.
Treatment Level #3
Patients in this treatment level will receive conventional stroke therapy as detailed in the first treatment level. Additionally, they will be treated with an emotionally responsive VR game as detailed in the second treatment level. Information about the participant’s emotions will be tracked by an electroencephalographic (EEG) machine. We will use this EEG to track as well as record brain wave patterns whilst the patient is participating in the VR game, and adjust the difficulty level of the game according to the information we receive from the EEG.
Motor Function Assessment
To assess patient’s upper arm function we will use the Jebsen Test of Hand Function and the Fugl-Meyer Assessment to quantitatively assess patients’ motor skills. The Jebsen Test of Hand Function (Jebsen et al. 1969) is an established method of fine hand motor functions. This will be used in conjunction with the Fugl-Meyer Assessment, which is a more general method of assessing sensorimotor function recovery. This method has been widely utilized in the research of stroke recovery and has been proven to be a reliable method of assessment (Gladstone, Danells, and Black 2002).
Statistical Analysis
Our study fits the criteria for a superiority trial, as it determines if a new treatment is superior to a standard treatment (Ahn and Ahn 2010;2015). Therefore, we will need to test the significance of our results. We will need to do statistical testing for treatment #2 and treatment #3 in relation to our control (treatment #1), as well as test the significance between treatment #2 and treatment #3. We will determine the p-value for these relations and if the value is 0.05 or less, we will know our results are statistically significant. P-value is crucial in the analysis of research. P-value represents the probability of results being from a true null hypothesis, with the lower the P-value representing a higher likelihood that the alternative hypothesis is true. This value should be minimized as it represents a Type 1 error, where a null hypothesis is incorrectly rejected (Biau, Kernéis and Porcher, 2008).
Discussion
Equipment Choice
The use of a haptic glove will provide concrete data as to the progress of each patient. This is useful for progress tracking and gives the patient definitive evidence of their improvement. In addition, we chose the use of the robotic/haptic glove and immersive game, because this is relatively accessible technology that stroke patients could potentially use in their own homes (Hou, Xiyuan and Sourina 2013). There exist robotic gloves that provide physical support and facilitate the training of certain upper-extremity movement (see for example, Polygerinos et al., 2014). While providing this additional support can assist patient recovery, individual needs vary too much to be incorporated into the design of the prototype.
We will use an EEG set which, as detailed in Teplan (2002), typically comprise of the following components: electrodes with conductive media (electrode cap), amplifiers with filters, A/D converter, and a recording device.
We choose to develop our game with considerations to engagement as well as what actions a stroke patient would need to practice. The actions outlined in the description of the immersive game are actions that are routinely performed in everyday life. Using simulations akin to common, real world situations provides a natural pattern of recovery conducive to rehabilitation (Kwakkel, Kollen, & Lindeman 2004). We also chose the theme of a mystery game as we feel that this genre is quite popular in general.
Ethical Considerations
Our study will adhere to ethical guidelines as outlined by Emanuel, Wendler, and Grady (2000). This includes ensuring our proposed study has the potential to result in valuable scientific insights, and our research methods will adhere to scientific method criteria. Our participants chosen for this study will be based on fair selection with no selection due to privilege or other erroneous factors. In addition, our study has been designed with risk versus benefits in mind both in relation to the participants and society at large. This proposed research project will, once completed, be subjected to independent peer review for consideration. Participants will have informed consent of the study, with a clear outline of risks and benefits involved, and their privacy will be protected. Each participant will be able to quit the study at any time if they so choose to. Lastly, participant well being will be monitored with them being withdrawn from the study if their well-being is being negatively affected by the study.
Predictions
For our proposed study we have made the following predictions. We predict that the control group, those who would simply use traditional rehabilitation methods (as detailed in our methods) will experience the least amount of recovery overall, as compared to the other two treatment levels. This is because this kind of treatment is less engaging and not as emotionally catering as the other treatment levels (Jack et al. 2001). Second, we predict that the use of the robotic hand as well as traditional rehabilitation methods will result in an increase of recovery as compared to the control. This is because the addition of the robotic hand and associated game experience would ideally create a more interactive and fun environment that will result in an increase in effort on the patients part. Lastly, we predict that the use of the emotionally responsive technology will result in the highest recovery rate overall. This is because this treatment level will include the benefits of the robotic hand and traditional rehabilitation methods, in addition to creating a more comfortable environment for the patients. With difficulty settings that can decrease with destructive emotions and increase with constructive emotions, these patients will have higher commitments to rehabilitation practices and have faster rehabilitation rates. (Maclean et al. 2000).
Study Deficiencies
While our chosen sample size of 60 is sufficient to get statistically significant results, if resources (participants and monetary) are available, increasing this sample size could produce even more precise results.
We decided we would select participants that have had a stroke in the last three weeks. Ideally, we would get patients started on our rehabilitation study as soon as they were able to. We felt this was just not feasible as it takes time to survey the participants, and get enough participants who have all had their strokes around the same time.
We decided we would only select participants who have just had their first stroke. This was to minimize complications that may arise after having multiple strokes. Further research could look at the usefulness of this technology on participants who have had more than one stroke.
Our study will utilize the Fugl Meyer Assessment. This type of assessment can lead to a ceiling effect. This means that the assessment is not ideal for detecting changes once a patient reaches a high level of motor function (Gladstone, Danells, and Black 2002). We attempted to balance these limitations by also utilizing the Jebsen Test of Hand Function.
Conclusion
While previous research has been conducted on the usage of video games as a method of stroke treatment; the lack of research into emotionally responsive VR provides an opportunity to assess its effectiveness, and provides insights for physicians to determine if it is a cost-effective component of rehabilitation. The Heart & Stroke Foundation’s 2016 Report on the Health of Canadians states that nearly half of all Canadians will either have experienced a stroke, or have somebody they know experience a stroke at some point in their lifetime. This leaves a lot of people whose lives can be made better with more effective treatments for themselves, or their loved ones. Recent advancement in consumer grade electronics now allows for better and cheaper access to motion tracking technologies, as well as allowing access to previously expensive and science-based equipment like the EEG.
Insights as Researchers
From this project one of our major insights was that while the use of robotics and immersive gaming for clinical rehabilitation appears to show a lot of promise, and could have valuable applications for at home use. While rehabilitation work in-clinic is very important for stroke recovery, at home exercises can significantly improve motor function (von Koch et al. 2000). Often times stroke patients are less motivated to do these exercises on their own time. It is our thought that the system used in the our research project could increase commitment to the exercises, improving recovery. While the emotionally responsive system may be more difficult to use outside of a clinical setting, if the results from our study show that it leads to significant increases in motor function, then this technology could be adapted for home use.
Insights as Students
From this project, we learned that there has been quite a lot of research done on the use of robotics and virtual reality for stroke rehabilitation. This research has also been going on for quite some time. This surprised us as we initially thought this was a relatively new field. As a student, we were also surprised to know that some aspects of EEG is still somewhat contested: there appear exist different criteria on standards and the best types of measurements, particularly when we look beyond alpha and beta waves. It was also interesting to note that sample sizes for studies on technology-aided stroke rehabilitation appear to be quite small. This may be due to the cost of prototypes, meaning that fewer patients can participate in parallel. It may also be due to the fact that it is difficult recruit participants during the critical period post-stroke.
Further Research
The usage of commercially available parts used to construct the equipment we used along with the lower price points allow for a more accessible use of this form of rehabilitation. Further research could be done on the use of this system in a home setting. Particularly, there may be valuable insights in looking into how a haptic gaming system could increase patients motivation at doing home exercises. With some modification, namely in using a robotic leg instead of a robotic arm, this form of treatment could also be used in the rehabilitation of stroke patients with lower body mobility issues. With little modification, this treatment could be used for those with various other conditions that can cause impairment to upper arm function. This includes traumatic brain injury, cerebral palsy, or multiple sclerosis. For example, studies have shown that, even though there is no cure for cerebral palsy, using VR gaming has some benefits over not pursuing any treatment at all, or pursuing alternative treatments (Snider, Majnemer, Darsaklis, 2009). Patients of cerebral palsy were able to feed and dress themselves after the intervention, which before they were not able to do. Empirical data is still lacking in this area of study, but holds positive prospects for further research.
Citations
Ahn, C., & Ahn, D. (2010;2015;). Randomized clinical trials in stroke research. Journal of Investigative Medicine, 58(2), 277-281. doi:10.2310/JIM.0b013e3181c9b2d4
Bath, P. (2008). Calculation of sample size for stroke trials assessing functional outcome: Comparison of binary and ordinal approaches: The optimising analysis of stroke trials (OAST) collaboration. International Journal of Stroke, 3(2), 78-84. doi:10.1111/j.1747-4949.2008.00184.x
Bernhardt, J., Hayward, K. S., Kwakkel, G., Ward, N. S., Wolf, S. L., Borschmann, K., . . . Cramer, S. C. (2017). Agreed definitions and a shared vision for new standards in stroke recovery research: The Stroke Recovery and Rehabilitation Roundtable taskforce. International Journal of Stroke, 12(5), 444-450. doi:10.1177/1747493017711816 (http://journals.sagepub.com/doi/pdf/10.1177/1747493017711816)r
Biau, D. J., Kernéis, S., & Porcher, R. (2008). Statistics in brief: The importance of sample size in the planning and interpretation of medical research. Clinical Orthopaedics and Related Research, 466(9), 2282-2288. doi:10.1007/s11999-008-0346-9
Biernaskie, J., Chernenko, G., & Corbett, D. (2004). Efficacy of rehabilitative experience declines with time after focal ischemic brain injury. Journal of Neuroscience, 24(5), 1245-1254. doi: 10.1523/JNEUROSCI.3834-03.2004
Colombo, R., Pisano, F., Mazzone, A., Delconte, C., Micera, S., Carrozza, M., . . . Minuco, G. (2007). Design strategies to improve patient motivation during robot-aided rehabilitation. Journal of Neuroengineering and Rehabilitation, 4(1), 3-3. doi:10.1186/1743-0003-4-3
Choi, I., Culbertson, H., Miller, M.R., Olwal, A., & Follmer, S. (2017). Grabity: A wearable haptic interface for simulating weight and grasping in virtual reality. Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology (UIST ‘17). New York: ACM. 119-139. doi: 10.1145/3126594.3126599
Dahiya R.S. & Valle M. (2013). Tactile Sensing: Definitions and Classification. In Robotic Tactile Sensing (pp 13-17). Springer: Dordrecht. doi: 10.1007/978-94-007-0579-1_2
Emanuel, E. J., Wendler, D., & Grady, C. (2000). What makes clinical research ethical? Jama, 283(20), 2701-2711. doi:10.1001/jama.283.20.2701
Gladstone, D. J., Danells, C. J., & Black, S. E. (2002). The fugl-meyer assessment of motor recovery after stroke: A critical review of its measurement properties. Thousand Oaks, CA: Sage Publications. doi:10.1177/154596802401105171
Hou, X. & Sourina, O. (2013). Emotion-enabled haptic-based serious game for post stroke rehabilitation. Proceedings of the ACM Symposium on Virtual Reality Software and Technology, VRST. 31-34. doi: 10.1145/2503713.2503738 Jack, D., Boian, R., Merians, A. S., Tremaine, M., Burdea, G. C., Adamovich, S. V., . . . Poizner, H. (2001). Virtual reality-enhanced stroke rehabilitation. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 9(3), 308-318. doi:10.1109/7333.948460
Jebsen, R. H., Taylor, N., Trieschmann, R. B., Trotter, M. J., & Howard, L. A. (1969). An objective and standardized test of hand function. Archives of Physical Medicine and Rehabilitation, 50(6), 311.
Kadam, P., & Bhalerao, S. (2010). Sample size calculation. International Journal of Ayurveda Research, 1(1), 55. doi:10.4103/0974-7788.59946
Kwakkel, G., Kollen, B., & Lindeman, E. (2004). Understanding the pattern of functional recovery after stroke: Facts and theories. Restorative Neurology and Neuroscience, 22(3-5), 281.
Maclean, N. (2000). Qualitative analysis of stroke patients motivation for rehabilitation. Bmj, 321(7268), 1051-1054. doi:10.1136/bmj.321.7268.1051 (http://www.bmj.com/content/bmj/321/7268/1051.full.pdf)
Merians, A.S., Jack, D., Boian, R., Tremaine, M., Burdea, G.C., Adamovich, S.V., Recce, M., & Poizner, H. (2002). Virtual reality—Augmented rehabilitation for patients following stroke. Physical Therapy 82(9), 898-915. doi: 10.1093/ptj/82.9.898
Mihelj, M., Novak, D., & Munih, M. Emotion-aware system for upper extremity rehabilitation. (2009). Proceedings of IEEE Virtual Rehabilitation International Conference. 160-165. doi: 10.1109/ICVR.2009.5174225
Polygerinos, P., Wang, Z., Galloway, K.C., Wood, R. J., & Walsh, C . (2014). Soft robotic glove for combined assistance an at-home rehabilitation. Robotics and Autonomous Systems, 77, 135-143. doi: 10.1016/j.robot.2014.08.014
Reuderink, B., Mühler, C., & Poel, M. (2013). Valence, arousal and dominance in the EEG during gameplay. International Journal of Autonomous and Adaptive Communications Systems, 6(1). 45-62.
Schulz, K. F., Altman, D. G., Moher, D., CONSORT Group, & for the CONSORT Group. (2010). CONSORT 2010 statement: Updated guidelines for reporting parallel group randomised trials. Bmj, 340(7748), 698-702. doi:10.1136/bmj.c332
Sourina, O. & Liu, Yi. (2011). A Fractal-based algorithm of emotion recognition from EEG using arousal-valence model. BIOSIGNALS 2011 - Proceedings of the International Conference on Bio-Inspired Systems and Signal Processing. 209-214.
von Koch, L., Holmqvist, L. W., Wottrich, A. W., Tham, K., & de Pedro-Cuesta, J. (2000). Rehabilitation at home after stroke: A descriptive study of an individualized intervention. Clinical Rehabilitation, 14(6), 574-583. doi:10.1191/0269215500cr364oa
Yoon, H. J. & Chung, Y. C. (2011). EEG spectral analysis in valence and arousal dimensions of emotion. Proceedings of the 11th International Conference on Control, Automation and Systems. 1319-1322. doi: 978-89-93215-03-8 98560/11/
Laver, K., George, S., Thomas, S., Deutsch, J., & Crotty, M. (2011). Virtual reality for stroke rehabilitation. Cochrane Database of Systematic Reviews, (9), CD008349. doi:10.1002/14651858.CD008349.pub2
Wang, G., Zhang, Z., Ayala, C., Dunet, D. O., Fang, J., & George, M. G. (2014). Costs of hospitalization for stroke patients aged 18-64 years in the United States. Journal of Stroke and Cerebrovascular Diseases, 23(5), 861-868.
Raghavan, P. (2015). Upper Limb Motor Impairment Post Stroke. Physical medicine and rehabilitation clinics of North America, 26(4), 599.
www.stroke.org
Bütefisch, C., Hummelsheim, H., Denzler, P., & Mauritz, K. H. (1995). Repetitive training of isolated movements improves the outcome of motor rehabilitation of the centrally paretic hand. Journal of the neurological sciences, 130(1), 59-68.
Bobath, B. (1990). Adult hemiplegia: evaluation and treatment. Elsevier Health Sciences.
Kneebone, I. I., & Jeffries, F. W. (2013). Treating anxiety after stroke using cognitive-behaviour therapy: Two cases. Neuropsychological rehabilitation, 23(6), 798-810.
www.ontariostrokenetwork.ca
Berk, R. A. (2009). Multimedia teaching with video clips: TV, movies, YouTube, and mtvU in the college classroom. International Journal of Technology in Teaching & Learning, 5(1).
Millbower, L. (2000). Training with a beat: The teaching power of music. Stylus Publishing, LLC..
Snider, L., Majnemer, A., & Darsaklis, V. (2010). Virtual reality as a therapeutic modality for children with cerebral palsy. Developmental Neurorehabilitation, 13(2), 120-128.
Teplan, M. (2002). Fundamentals of EEG measurement. Measurement science review, 2(2), 1-11.
Lincoln, N. B., Mulley, G. P., Jones, A. C., McGuirk, E., Lendrem, W., & Mitchell, J. R. A. (1984). Effectiveness of speech therapy for aphasic stroke patients: a randomised controlled trial. The Lancet, 323(8388), 1197-1200.
Steultjens, E. M., Dekker, J., Bouter, L. M., van de Nes, J. C., Cup, E. H., & van den Ende, C. H. (2003). Occupational therapy for stroke patients. Stroke, 34(3), 676-687.