Effect of testing procedure on Corsi's block-tapping task in normal subjects and Alzheimer-type dementia.

Corsi's block-tapping task was given to 30 normal subjects and 38 Alzheimer-type demented patients following two different procedures. The first is the most widely standardized (scoring criterion: 3 correct reproductions out of 5 sequences), while the second is more lenient since it does not require subjects to replicate a certain performance three times. Demented patients' scores were lower than those of controls in both conditions, and scores on the two tasks were significantly correlated for patients and controls. However, the requirement of replicating the visuospatial memory performances was more detrimental for demented patients than for controls so the two procedures cannot be considered equivalent.