Cognitive skill learning in amnesia

Abstract

Amnesic patients and control subjects were trained on a computer-based cognitive skill task. Initially, the amnesic patients improved at a normal rate. A separate group of normal subjects did not improve on the task when they received random feedback from the computer on each trial. One month later, when additional training on the task was given, the control subjects performed better than the amnesic patients. There was little if any transfer by either group to a second skill task that had a different surface appearance but which required the identical cognitive strategy. Questionnaire data showed that the control subjects, but not the amnesic patients, acquired declarative knowledge about the strategy needed to perform the task. The findings show that cognitive skill learning can initially proceed at a normal rate in amnesia, presumably because declarative knowledge makes no material contribution to performance early in training. However, declarative knowledge may make an important contribution later in training and may therefore give normal subjects an advantage over amnesic patients.