Table 2.

Changes in ARAT scores and movement phase durations

		ARAT	Reach Duration (msec)		Grasp Duration (msec)		Transport Duration (msec)		Release Duration (msec)
Healthy Controls mean ± 1 SE		57 ± 0	425 ± 17		236 ± 14		not available		not available
Participant			mean	n	mean	n	mean	n	mean	n
R005	Pre	38	442	10	314	10	964	10	186	10
	Post	57	548 ‡	10	263	10	935	10	−122 ‡	10
R007	Pre	20.3	1050	8	365	8	2196	7	590	7
	Post	28	1060	10	281 ‡	10	2762	10	1153	10
R008	Pre	43	558	10	327	10	2630	10	−1647	10
	Post	56	379 †	10	157 †	10	740 †	10	−15 †	10
R009	Pre	9.3	3870	10	2127	9	2550	2		0
	Post	11	2646	7	1483	7		0		0
R010	Pre	40	593	10	325	10	2334	10	−1455	10
	Post	53	963 †	10	302	10	1412 ‡	10	−11 †	10
R011	Pre	26.7	1692	10	447	10	4011	7	−641	7
	Post	31	2466 †	10	666	10	2477	7	63	7
R012	Pre	20	2004	10	1359	10	2910	9	909	8
	Post	24	1794	10	1754	10	3764 ‡	9	400	9
R013	Pre	15	1491	10	497	10	3990	1	−220	1
	Post	15	1516	10	827 †	10		0		0
R014	Pre	22	2022	10	1435	10	3811	7		0
	Post	31	1367 †	10	681 ‡	10	1893 ‡	10	1485	10
R015	Pre	20	3664	10	2194	5	2114	5	−550	1
	Post	24	1803 ‡	9	356 ‡	9	1427 ‡	9	−312	9
Group	Pre	25	1739	10	939	10	2751	10	−354	8
	Post	33 *	1454	10	677	10	1926	8	330	8
	Effect size (Est. N)	1.61 (6)	0.37 (60)		0.39 (54)		0.65 (21)		0.66 (21)

^*Post > Pre, p = 0.003

Bold type indicates individual or group pre-post differences that were statistically significant at the p < 0.05 level (‡) or at the p < 0.0005 level (†). Note that some significant changes were in the unexpected direction (e.g. increased reach phase duration for R005, R010, R011). For R005 through R015, n represents the number of trials included in the individual’s mean. Where n < 10, the participant was unable to complete the movement phase during every trial. For Group results, n represents the number of participants for whom data were available. The pre-training ARAT score is the mean of three baseline tests each separated by one week. The post-training ARAT score is from one test administered at the end of the training program. The effect sizes and estimated sample sizes (Est. N) are from post-hoc power analyses for each parameter (see Methods).