Hayward KS, Kramer SF, Dalton EJ, et al. - Despite a large and expanding body of research, the intervention dose and sample size of included studies were frequently insufficient to detect clinically significant effects. Furthermore, interventions have remained largely focused on subacute stroke recovery, with little change in recent decades. To advance stroke recovery research, a unified research agenda that establishes a clear biological understanding of timing, dose, and intervention type is required.

• Two hundred sixty-one studies with a total of 228 (n = 9,704 candidates) unique data sets were included.

• Between 1980 and 1984, the number of studies completed increased from one (n = 37 participants) to 91 (n = 4,417 candidates) between 2015 and 2019.

• The timing of the intervention has not changed, and the study sample size remains small.

• The majority of studies were rated as having a high risk of bias (62%).

• Participants in the study were enrolled at various stages of recovery: 1 hyperacute (24 hours), 13 acute (1–7 days), 176 early subacute (8–90 days), 34 late subacute (91–180 days), and 4 who were unable to be classified.

• The median dose for both the intervention and control groups was 45 (IQR, 600–1430) min/session, 1 (IQR, 1–1) session/d, 5 (IQR, 5–5) d/wk for 4 (IQR, 3–5) weeks.

• Electromechanical (n = 55 studies), electrical stimulation (n = 38 studies), and constraint-induced movement (n = 28 studies) therapies were the most commonly tested interventions.