Fig. 2

Post-stroke recovery of dexterous hand use. Box and block test (BBT) score as outcome measure across time (training) and groups (Dextrain group vs. CT, conventional therapy group). A Primary outcome, BBT-change at post-training (T1-T0), did not differ between Dextrain and CT groups. BBT-change at 3-months follow-up (T2-T0) also did not differ between groups. B Within-group BBT scores compared to reference values in healthy control subjects (N = 30) at T0 (baseline), T1 (post-training) and T2 (follow-up); Dextrain and CT groups significantly improved BBT score with training (T1); the CT group showed retention of BBT gains at T2, while the Dextrain group showed further improvement at T2. *P < 0.05. C Degree of initial motor impairment and therapy-mediated gains in dexterous hand use. Exploratory analysis after median-splitting the groups according to Fmax% on T0 (line = median, box = 25–75%, whisker = min/max). Dextrain training was more beneficial in patients with moderate motor impairment (Low subgroup) than those with mild motor impairment (High maximal force subgroup; Z = − 2.32, P = 0.02). There was no significant difference in BBT-change between Low and High subgroups within the CT group. The Low motor impairment group also improved more in BBT-change after Dextrain than conventional therapy (Mann–Whitney U Test, P < 0.05)