The ability to name pictures or concepts is widely used as a clinical measure of cognitive status and a naming deficit is often one of the first symptoms of dementia. Since healthy older adults also show some naming deficits, it becomes important to distinguish between naming problems associated with normal aging and those that might presage dementia. The goal of the proposed research is to develop a model of naming that is sufficiently detailed to provide specific processing mechanisms that account for naming deficits associated with normal and abnormal aging. The present proposal advances a neural network model of naming in which the user's knowledge of language is represented by a complex, highly interconnected network of simple processing elements organized at visual, semantic, and phonological levels of representation. The model is realized in the form of a computer program that can simulate the processing of an experimental naming trial. Neural network models of this type, in which processing mechanisms are fully specified, fill in the """"""""black boxes"""""""" of traditional information processing theories, and do so in language close to that of the brain itself. The approach is especially suited to understanding the underlying mechanisms of the naming deficits that accompany normal aging. Retrieval in the model is accomplished through spreading activation in the network. Since it is widely hypothesized that age differences in retrieval are related to rate of spreading activation, the proposed research addresses how activation is propagated in the network over time. The focus on the first 150 milliseconds of processing, an interval which has been shown in the current research to be one rapid change in activation for both young and older adults. A primed picture-naming paradigm is used in which a target picture is preceded by a priming stimulus that is related or unrelated to the target. Relatedness is modelled as shared connections between prime and target in the network. The difference in the efficiency of target retrieval as a function of prime type is assumed to reflect the influence of activation of the prime. Varying the time available to process the prime allows the time course of its influence on target naming to be traced. This in turn permits the direction, sign, and temporal dynamics of the connections in the postulated network model to be specified in detail. Network simulations shown in the proposal produce activation functions that are in striking parallel to the observed data. This outcome confirms the feasibility to the goal of modelling age differences in naming at the level of retrieval mechanisms.
| Bowles, N L (1994) Age and rate of activation in semantic memory. Psychol Aging 9:414-29 |
| Bowles, N L (1989) Age and semantic inhibition in word retrieval. J Gerontol 44:P88-90 |
| Bowles, N L; Poon, L W (1988) Age and context effects in lexical decision: an age by context interaction. Exp Aging Res 14:201-5 |
| Bowles, N L; Obler, L K; Albert, M L (1987) Naming errors in healthy aging and dementia of the Alzheimer type. Cortex 23:519-24 |
