| |
Abstract:
U. Pennsylvania and U. Michigan Patients with Alzheimer's
disease (AD) and frontotemporal degeneration (FTD) chose one of two
categories based on a description (e.g. "A round object 3 inches in
diameter"). One category had fixed dimensions (e.g. "quarter"); the
other, variable (e.g. "pizza"). The described value fell between
the 2 categories on the relevant dimension. The descriptions were
"brief" (as above) or "rich" (e.g. "A round object 3 inches in
diameter found in an arcade"), biasing toward the fixed-dimension
category. This task was performed at separate times with 2
different instructions. "Definitional" instructions indicated that
one category must be correct, requiring subjects to identify and
reject the fixed-dimensions category (a quarter cannot have 3-inch
diameter). "Similarity" instructions asked for the category most
similar to the description. With brief descriptions, controls chose
the (correct) variable-dimension category significantly more under
definitional (80%) than similarity instructions (56%). Patients
showed no such difference: Variable-dimension categories were
chosen 60% by 11 AD patients and 55% by 25 FTD patients with
definitional instructions (significantly less than controls), and
53% by AD and 50% by FTD under similarity instructions. Rich
descriptions biased all subjects toward fixed-dimension categories,
indicating sensitivity to feature content in descriptions. Under
definitional instructions, controls chose the variable-dimension
category 59%, AD 44%, and FTD 41%; under similarity instructions,
controls chose the variable-dimension category 38%, AD 36%, FTD
34%. AD and FTD patients cannot process the critical features
necessary to define and categorize an object.
|
