Motor function increased significantly in the treatment group. The treatment effect was significant when the initial FM-scores were used as covariates ( F(1.41)=8.513, p < 0.01). Motor function in treated patients with more severely affected arms (low FM-score: 0-29) improved but not significantly more so than controls with more severely affected arms. Motor function in treated patients with less severely affected arms (higher FM-scores: 30-50) was significantly improved compared to controls with less severely affected arms (p < 0.05). The Low-TENS did not decrease either pain or spasticity, nor did it cause a significant improvement in Barthel scores.