| |
Abstract:
Abstract: Striatal contributions to category learning were
investigated by examining the ability of patients with Parkinson's
disease (PD) or Huntington's disease (HD) to learn linear and
nonlinear categorization rules. Accuracy rates and indices derived
from quantitative models were evaluated. In the linear condition,
subgroups of PD patients emerged with approximately half of the
patients learning the rule as well as controls and the other half
of the patients not learning the rule. Analyses of the model
indices revealed that the impaired accuracy performance of the PD
non-learners was due to deficits in both rule learning and rule
application consistency. The HD patients were also impaired in
learning the linear categorization rule; however, no subgroups
emerged and their deficit was due only to impairment in
categorization rule learning and not rule application consistency.
For the nonlinear rule, no subgroups emerged for either the PD or
HD groups, although both groups were impaired in learning the rule
relative to controls. The model-based analyses indicated that their
deficit was due to categorization rule learning and not rule
application consistency. Unlike previous studies of categorization
learning in striatal-damaged patients, early-training deficits did
not eventually diminish later in training. The results of this
study provide strong evidence that the striatum contributes
substantially to linear and nonlinear categorization
learning.
|
