Upper extremity (UE) paresis is well known to be a debilitating side effect of stroke, causing many adults yearly to rely on the assistance of care partners for their basic activities of daily living (ADL). Targeted UE therapy can help improve capacity, but the optimal dosing and intensity remains unclear.1 Thus, alternative therapy options that can be used as an adjunct to conventional therapy in a variety of therapeutic settings (hospital, outpatient clinic, and home) should be easily accessible to stroke survivors and their care partners. Electromyographic (EMG) biofeedback is one method that has been shown beneficial for arm strength after stroke.2 Therapeutic games increase motivation and engagement with therapies, which may benefit adherence and even efficacy.3-6 Studies of game-based home stroke telerehabilitation have found high rates of patient acceptance and compliance and have demonstrated significant motor gains.7

In this issue, Seo and colleagues8 randomized 20 people with chronic stroke and moderate hand impairments to receive 9 training sessions of either unilateral or bilateral gamified EMG biofeedback therapy. This therapy used a set of previously described games played by controlling the horizontal and vertical positions of a cursor on a screen.9 Four key muscles for wrist and hand function were used to control the games. EMG activity from the 4 muscles was converted into 2-dimensional signals using principal component analysis applied to EMG activity from the unimpaired arm during a calibration procedure. The unilateral group controlled the games solely with the impaired hand, and in the bilateral group control was gradually shifted from the unimpaired hand to the impaired hand over the training sessions. The authors found that both groups improved their skill playing the games, as measured by the participants taking shorter, more direct trajectories and completing the games in less time after training, with no difference between the groups. Both groups also demonstrated a small, but statistically significant, improvement in function assessed with the Wolf Motor Function Test (WMFT). Importantly, as the authors note, this change exceeded the minimal detectable change but not the minimal clinically important difference of the WMFT.9

Compared with another study of gamified EMG biofeedback for arm rehabilitation after stroke,10 Seo and colleagues8 used 4 (v 2) EMG channels on the paretic limb, placed them more distally, and compared the use of bilateral versus unilateral training. Given the results were similar for both bilateral and unilateral groups, we focus on the other differences. The use of principal component analysis on EMG activity from the intact limb during calibration is a natural way to reduce the 4-channel activity into a 2-dimensional control signal, but also abstracts away what is happening at the individual muscle level. Mugler et al10 explicitly show a reduction in co-contraction of agonists and antagonists through biofeedback training. As such, subsequent analysis of the data from Seo et al8 may similarly show that their protocol helps participants improve activation of their muscles outside their initial synergy patterns. An alternative hypothesis is that this training protocol does not reduce impairment at the level of independent muscle activation (ie, neural restitution) and that the improved game performance reflects skill activating and controlling muscles within those synergy patterns. This is important to distinguish, as skill training may reflect a different site and type of neuroplasticity.11

The fact that studies by Seo et al8 and Mugler et al10 demonstrate improvements in arm function indicates that EMG biofeedback therapy is useful for both proximal and distal muscles after stroke and is consistent with prior work.2 One limitation of this study, also noted by the authors, is the lack of a control group with either a placebo or active but non-biofeedback-based intervention for comparison. This issue is particularly relevant as the same rater performed the pre- and post-WMFT assessments, who presumably could distinguish a new versus return evaluation. This limits the robustness of the single-blind design, as the blinding is only between unilateral and bilateral groups who both showed similar benefits, and the demonstrated improvement in WMFT should be considered in light of this.

One key area for translation to usual care for this readership to engage with is how this type of technology can be deployed as a home exercise program. Studies of home-based, gamified therapies for arm recovery after stroke show they are efficacious with high adherence.7 EMG biofeedback in this study used laboratory equipment, but advances in wearable sensors now allow multichannel EMG games to be deployed using only a smartphone.13 Stroke survivors who have limited access to outpatient therapy services, who have adequate psychosocial support, and who are motivated by therapeutic games will benefit from these technologies readily available to the average consumer.

Most stroke survivors will attend therapy services to function better at home, work, and in their community, and Waddell et al12 found that a large majority of self-identified goals for chronic stroke survivors target improved performance of ADL. While Seo et al8 demonstrated stroke survivors could learn to control their muscle activation patterns, they did not include functional task practice as part of the protocol. Combining EMG-based therapy with functional task practice may produce a greater functional benefit and better support people's goals. Additionally, using EMG-controlled games as a home-based therapy combined with salient and environmentally specific functional task practice could help to examine the carryover effects of improved muscle activation to functional tasks the authors mentioned. These updates to future study designs will further entice clinicians to consider this type of therapy intervention routinely.

This study is impressive for its demonstrated adherence, with all 20 participants completing the study and only 4 out of a total of 180 sessions missed. Although enjoyment was not formally assessed, the authors anecdotally noted that participants liked the study. This is reassuring, as a key benefit of gamified therapies (ie, serious games) is to increase the engagement and intrinsic motivations of participants and ultimately to increase the effectiveness of treatments.5 Greater engagement with games is also correlated with greater improvements in Fugl-Meyer assessments.14 Thus, the combination of EMG from wearable sensors, gamification, and mobile computing provides a natural path to translate this treatment into an engaging, effective, home-based therapy.

REFERENCES