SIGNIFICANCE: Most stroke survivors experience residual upper-extremity (UE) motor impairment, which limits independence and burdens caregivers. Reacquisition of UE skills is a priority. A growing body of literature suggests that ample practice of slightly challenging tasks is necessary to drive reorganization in brain areas associated with improved skills. Task practice at the just-right challenge level—neither too easy nor too difficult—optimizes motor learning by offering opportunity for implicit error detection and motor strategizing. However, there is no standardized method to link task difficulty to client ability in stroke rehabilitation.

INNOVATION: In rehabilitation research and practice there is, at best, a tenuous link between objective client measurement and treatment decisions. However, it is critical to precisely understand the effect of treatment decisions for a given client to personalize treatment type and dose. The innovative standardized method tested here shifts current practice paradigms by clearly defining clients’ initial ability level, customizing treatment to the appropriate challenge, systematically progressing treatment to match evolving ability, and detailing how individuals within a sample respond/do not respond.

APPROACH: We hypothesized that repetitive task practice at the just-right challenge level would maximize UE motor skill reacquisition. The Rasch item response model was previously applied to the Fugl–Meyer Upper-Extremity Assessment (FMA–UE), a widely used poststroke assessment. The analysis calibrated item-difficulty and person-ability measures onto one metric and produced a keyform recovery map. The keyform illustrated a sequential progression of less difficult uni-joint to more difficult multijoint UE movements. As a measurement framework, the keyform pinpoints client-specific UE tasks having high, low, and 50% probability of success—that is, the just-right challenge level. Here, we tested the framework’s clinical feasibility.

METHOD: We used a proof-of-concept prospective study, and the setting took place in a stroke rehabilitation research laboratory. Inclusion criteria included the following: ischemic stroke > 3 mo, voluntarily shoulder flexion ≥ 30°, and concurrent elbow extension ≥ 20°. Subjects were recruited from an institutional review board (IRB)–approved registry. The keyform linked client ability measures to optimally difficult tasks. Feasibility was defined by subject pain/fatigue, motor function (Wolf Motor Function Test [WMFT]), UE motor ability (FMA–UE), and movement patterns (kinematics). Pre- and posttreatment scores were compared with paired t tests.

RESULTS: Ten subjects (59.70 ± 9.96 yr, 24.1 ± 30.54 mo poststroke) participated in nine progressively more difficult task-practice sessions without pain/fatigue. Subjects gained UE function (WMFT, p = .01), UE ability (FMA–UE, p = .00), and reduced compensatory movement (p = .02). Subjects improved shoulder–elbow coordination (p = .01), hand–eye coordination (p = .03), and shoulder–elbow active range of motion (shoulder flexion, p = .01; elbow extension, p = .06).

CONCLUSION: The FMA–UE keyform is a feasible standard method that personalized task difficulty throughout a rehabilitation program. Results from this small pilot study inform a large, well-powered, federally funded, ongoing randomized controlled trial (RCT).