Humans learn skilled acts in order to effectively interact with their environment. A loss of the ability to perform skilled acts is termed apraxia. Apraxia has been thought to be of theoretical interest, but the ecological implications of apraxia are controversial and have not been fully studied. We examined ten patients with unilateral left hemisphere cerebral infarctions (eight of whom were apraxic) and compared their mealtime eating behavior to a group of neurologically normal, age-matched controls. The stroke patients were less efficient in completing the meal. They made more action errors and were less organized in the sequencing of mealtime activities. Because the patients made more errors while using tools than when performing nontool actions, their deficit could not be accounted for by an elemental motor deficit. A positive relationship was found between mealtime action errors and the severity of apraxia. These findings suggest that limb apraxia may adversely influence activities of daily living. (JINS, 1995, I, 62–66.)

We assessed the prediction that object structural cues could benefit the learning of object–action relationships in ideomotor apraxia (IMA). A total of 15 patients with left-hemisphere stroke, 11 of whom exhibited IMA, and 10 healthy subjects were trained to match novel gestures to novel tool pictures that were either High- or Low-Afforded by their associated tools. Learning was assessed with recognition and production tests. Only IMA patients demonstrated better recognition of High- than Low-Afforded gestures, and their recognition of High-Afforded gestures was statistically comparable to the other groups. This finding suggests that apraxics may rely abnormally on object structure when learning to associate novel gestures and tools. Finally, the “affordance benefit” was associated with relative sparing of structures in the dorsal visual processing stream. These data are consistent with the proposal that two routes may mediate skilled action, one specialized for stored information, and the other responsive to object structure, and that deficient gesture learning may be compensated by “bootstrapping” intact dorsal stream coding of action. (JINS, 2007, 13, 997–1008.)