Rehab CARES Delivers Usable Robot Therapy for Chronic Stroke Recovery
A small but carefully followed cohort of adults with chronic stroke found that a community-based robot exercise platform could be tolerated well while still feeling engaging, usable, and clinically adaptable. Rehab CARES, built around the reconfigurable THEAbot system, paired upper-limb motor training with games designed to recruit both motor and cognitive processes. The experience matters because post-stroke disability rarely stays in one lane: weakness, slowed planning, attention deficits, and reduced confidence often coexist, and rehabilitation programs that ignore that overlap can feel mechanically repetitive or cognitively underpowered.
Six adults completed 12 therapy visits each. Their baseline profiles were heterogeneous, with Fugl-Meyer Upper Extremity scores ranging from 20 to 61 and MoCA scores from 22 to 28. That spread created a meaningful test of whether one platform could remain acceptable across different levels of motor impairment and cognitive reserve. It did.
The study tracked more than whether participants showed up. Pain, exertion, enjoyment, sense of control, excitement, frustration, effort, and usability were all measured across therapy visits. In practice, that produced a layered picture of how people experienced the robot rather than a simple yes-or-no verdict on feasibility.
Low Pain, Light-to-Moderate Effort, and Stable Enjoyment
Self-reported pain stayed low throughout treatment, an important finding for any rehabilitation tool that asks patients to repeat movements over multiple sessions. Exertion was generally light to moderate, which suggests the system challenged participants without pushing them into the kind of fatigue that can erode adherence. Enjoyment remained strong and stable across visits, a notable result in a setting where novelty often fades quickly after the first exposure to a device.
One sentence captures the bedside significance plainly: the robot did not become a burden during repeated use.
Affective ratings moved in a favorable direction as well. Sense of control increased over time, while excitement plateaued. Frustration and effort declined on workload measures, which is consistent with a user learning the interface and gaining confidence in how to engage with the games. In a rehabilitation system, that pattern matters because users who feel clumsy or overwhelmed often disengage before the therapeutic dose accumulates.
Game Selection Reflected Mixed Motor and Cognitive Profiles
Rehab CARES did more than offer a single repetitive exercise loop. The platform included eight games with different cognitive processes and varying levels of motor/cognitive demand, and participants selected among them during therapy. Game selection was only partially consistent with individual clinical profiles, suggesting that patients did not choose solely on the basis of impairment severity. Preferences and perceptual abilities likely shaped those choices, which is clinically plausible in a cohort where one participant may value challenge and another may prioritize clarity, speed, or visual simplicity.
That finding has direct implications for robot-mediated stroke rehabilitation. If a system can adapt difficulty but still leave room for patient choice, it may support adherence better than a rigid protocol. Cognitive engagement also becomes more than a secondary feature. In chronic stroke, the motor task is rarely isolated from planning, attention, or visual scanning, and a game structure that reflects that reality can make treatment feel more relevant to daily function.
The device architecture likely contributed to this flexibility. A reconfigurable one-degree-of-freedom robot can be simpler than a high-complexity exoskeleton, yet still support meaningful repetitive practice when paired with software that adjusts challenge dynamically. In this case, a haptic and biometric controller modified difficulty in real time, creating a feedback loop between performance and task demands. That kind of adaptive design is well suited to stroke survivors whose capacity may vary from day to day, or even within a single session.
Usability Rose Above Acceptability Thresholds Across Visits
Usability scores exceeded acceptability at every measured time point, and the final visit produced the highest score at 86 ± 4.5 out of 100. That trajectory suggests more than polite participation. It points to sustained acceptance of the platform over time, which is often the hardest hurdle for technology-based rehabilitation. Early enthusiasm can mask clunky controls, confusing prompts, or physical discomfort; later ratings are usually more honest.
Here, the later experience remained positive. The rising usability score fits with the reported decline in frustration and effort, and it suggests that repeated exposure may have allowed participants to learn the interface without being alienated by it. For community-based rehabilitation, where access, simplicity, and low burden can determine whether a program survives outside a laboratory, that is a practical outcome.
The strongest signal in the data is not a dramatic therapeutic reversal but a durable user response. Chronic stroke rehabilitation depends on repetition, and repetition depends on whether patients will return. A platform that is comfortable, understandable, and flexible has a better chance of generating enough exposure to matter.
Why Reconfigurable Robotics Matter in Stroke Care
Robot-assisted therapy has often focused on motor impairment alone, even though stroke recovery rarely respects that boundary. The combination of upper-limb weakness and cognitive inefficiency can make standard drills tedious or inaccessible, especially for patients whose attention, working memory, or perceptual processing has also changed. Rehab CARES addresses that gap by blending movement practice with cognitive demands and by adjusting difficulty as performance changes.
The heterogeneous response in this small cohort is not a weakness to dismiss; it is the exact reason adaptive systems are needed. A patient with low motor capacity but relatively intact cognition may engage differently from someone with better strength but reduced perceptual flexibility. By offering several games rather than one fixed task, the platform gives clinicians a broader menu for matching therapy to the person in front of them.
The study remains limited by size, and it does not establish efficacy for motor recovery or cognitive improvement. What it does establish is feasibility in a community-based setting, with a user experience profile that supports continued development. In a field where many technologies fail because patients do not like using them, a stable pattern of low pain, modest exertion, and rising usability is not trivial. It is the kind of early evidence that determines whether a platform deserves a larger trial with functional endpoints.