The structure of recall in amnesia

This thesis reports seven experiments on the nature of the functional deficit in amnesia. Experiments 1 to 3 investigate patterns of recall for amnesic subjects and matched controls to investigate a hypothesised specific deficit in recall in amnesia. No significant evidence of a recall deficit in amnesia was found. However, a difference emerged between the two groups in the analyses of the stochastic relationship between recall and recognition. This revealed that in amnesic subjects recall is approximately independent of recognition, whereas in control subjects they are positively related. The second three experiments investigated a hypothesised selective deficit of spatial memory by comparing amnesic and control memory for the locations of objects or words placed on a grid. The hypothesis that intentional encoding of locations would improve amnesic spatial memory scores resulting in a trade-off of recall and recognition of the item's identities was also examined. No significant evidence of a selective spatial memory deficit in amnesia was found, nor did intentional instructions improve amnesic spatial memory scores. There was no significant evidence of a trade-off of item and location memory in the amnesic group. A further analysis comparing control and amnesic memory for the location of items scored by lenient criteria found no significant difference between the amnesic and control scores for number of items of this type, or for recall and recognition memory of these items. Fragment and schema models have been applied to normal memory for this type of contextual material. In a final experiment, the predictions of both types of model were contrasted with each other for data on singly and multiply cued recall provided by both normal and amnesic subjects. It was found that amnesics and normal controls formed fragments representing the unrelated triads and schemas representing the related triads. Both the schema and the fragment model parameters displayed uniform patterns of impairment. Thus amnesic memory may be argued to differ from normal memory quantitatively, rather than qualitatively. The implications of these findings for theories of memory and processing in amnesia are discussed