Objective: To evaluate the efficacy of a gravity-supported, computer-enhanced device (Armeo (R) Spring) for upper limb rehabilitation in chronic stroke patients. Material and methods: We included 23 chronic hemiparetic patients (chronicity: 328 +/- 90.8 days; distribution: 17 men and 6 women) aged 54.6 +/- 9.5 years, who had sustained ischaemic stroke (n =12) or haemorrhagic stroke (n = 11). All patients completed 36 one-hour sessions using the Armeo (R) Spring system. Arm movement was assessed at the beginning and end of the treatment programme and once more 4 months later. Main outcome measurements covered structure, activity, and function, as per the International Classification of Functioning, Disability and Health: Modified Ashworth Scale, Motricity Index (MI), Fugl-Meyer Assessment Scale (FM), Motor Assessment Scale (MAS), Manual Function Test (MFT), and Wolf Motor Function Test (WMFT). Results: Repeated measures ANOVA showed significant improvement (time effect) for all function scales (P < .01 for FM and MI) and activity scales (P < .01 for MAS, MFT and WMFT-ability, and P < .05 WMFT-time) without significant changes in muscle tone. The post-hoc analysis (Bonferroni) showed different evolutionary patterns for function and activity measurements, and clear benefits related to Armeo (R) Spring training, especially on activity scales. Conclusions: Armeo (R) Spring is an effective tool for rehabilitating the affected arm in patients with hemiparesis secondary to ictus, even in the chronic stage. (C) 2012 Sociedad Espanola de Neurologia. Published by Elsevier Espana, S.L. All rights reserved.