Objective: To examine in subjects with stroke using their nonparetic side how different levels of stimulus-response (S-R) compatibility, which require different levels of information processing, affect manual tracking control. Design: Descriptive study comparing finger movement tracking performance under S-R-compatible and S-R-incompatible conditions between subjects with stroke and healthy controls. Four two-factor analysis of variance tests with one independent factor (group, gender, laterality, or order) and one repeated measures factor (position) comprised the data analysis. Setting: University-based research setting. Patients: Forty subjects with chronic stroke: 20 right hemiplegia (average age, 65.2 ± 2.3yrs); 20 left hemiplegia (average age, 68.6 ± 2.3yrs). Fifty-one healthy controls: 24 using non-dominant hand (average age, 68.6 ± 2.1yrs); 27 using dominant hand (average age, 68.7 ± 2.0 yrs). All were right-handed. Main Outcome Measure: Tracking accuracy index (AI), based on root-mean-square error normalized to scale of each subject's tracking target. Results: In S- R-incompatible condition, AI of subjects with stroke was not significantly different from controls (F(1, 89) = 1.73, p = .19). In S-R-compatible condition, AI of control subjects was significantly better than subjects with stroke (F(1, 89) = 14.3, p = .0003). Conclusion: Manual tracking is impaired in nonparetic hand of subjects with stroke, suggesting that information processing, distinctly separate from motor weakness, may be an underestimated problem impairing controlled movements in individuals with stroke.