A double dissociation between accuracy and time of execution on attentional tasks in Alzheimer's disease and multi-infarct dementia.

Two cancellation/attentional tasks: (i) Lines Cancellation (LC) and Multiple Features Targets Cancellation (MFTC) and (ii) a standard battery of neuropsychological tests, the Mental Deterioration Battery (MDB), were administered to 68 patients with dementia of the Alzheimer's type (DAT) and 40 patients with multi-infarct dementia (MID), who were accurately matched for the overall severity of dementia, and to 40 normal controls. Both accuracy and time of execution were considered in evaluating performance on the two cancellation tasks, which involved visuospatial exploration and psychomotor speed, but were differently demanding in terms of selective attention. On the first cancellation task (LC), requiring a lower attentional load, the two demented patient groups performed at the same level of accuracy. On the second cancellation task (MFTC), which was more demanding in terms of selective and divided attention, DAT patients were significantly less accurate than MID patients, making a higher number of 'false-alarm' errors. Conversely, the time employed in the execution of both LC and MFTC took longer for MID than for DAT patients, suggesting a greater impairment of psychomotor speed in MID. In the MDB, DAT patients scored significantly worse than MID patients on several measures of episodic memory (the immediate recall, delayed recall and delayed recognition of Rey's Auditory Verbal Learning Test) and on a test of visual-spatial memory. These data suggest that, while psychomotor speed and the lower (sensorimotor) levels of attention are preferentially impaired in subcortical forms of dementia such as MID, the higher levels of selective and divided attention are more markedly disrupted in the Alzheimer type of dementia.