Academic journal article Cognitive, Affective and Behavioral Neuroscience

Exploring the Impact of Plasticity-Related Recovery after Brain Damage in a Connectionist Model of Single-Word Reading

Academic journal article Cognitive, Affective and Behavioral Neuroscience

Exploring the Impact of Plasticity-Related Recovery after Brain Damage in a Connectionist Model of Single-Word Reading

Article excerpt

The effect of retraining a damaged connectionist model of single-word reading was investigated with the aim of establishing whether plasticity-related changes occurring during the recovery process can contribute to our understanding of the pattern of dissociations found in brain-damaged patients. In particular, we sought to reproduce the strong frequency × consistency interactions found in surface dyslexia. A replication of Plaut, McClelland, Seidenberg, and Patterson's (1996) model of word reading was damaged and then retrained, using a standard backpropagation algorithm. Immediately after damage, there was only a small frequency × consistency interaction. Retraining the damaged model crystallized out these small differences into a strong dissociation, very similar to the pattern found in surface dyslexic patients. What is more, the percentage of regularization errors, always high in surface dyslexies, increased greatly over the retraining period, moving from under 10% to over 80% in some simulations. These results suggest that the performance patterns of brain-damaged patients can owe as much to the substantial changes in the pattern of connectivity occurring during recovery as to the original premorbid structure. This finding is discussed in relation to the traditional cognitive neuropsychological assumptions of subtractivity and transparency.

Traditionally, the science of cognitive neuropsychology has relied on evidence supplied by brain-damaged patients to make inferences about the cognitive systems subserving normal brain function. From the early work of Broca (1861) and Wernicke (1874) to the present day, a key feature of the discipline has been its reliance on information from single cases. In order to interpret this type of data, we are forced to make a number of key assumptions. These are often documented in the introductory chapters of neuropsychology textbooks (e.g., Ellis & Young, 1988; Shallice, 1988). For the purposes of this article, we are concerned primarily with two such (related) assumptions: subtractivity and transparency.

Caramazza (1984) defines transparency as the assumption that "pathological performance observed will provide a basis for discerning which component or module of the system is disrupted" (p. 10). Were this assumption to be violated on a large scale, the whole enterprise of cognitive neuropsychology would become vastly more difficult. For example, if a deficit whereby patients regularize irregular words could not be interpreted as the result of damage to a processing component in the language system, it would be difficult to see how we could progress in our understanding of brain function. It is not necessary, however, to propose such a major dislocation between behavioral performance and cognitive structure for the transparency assumption to be threatened. A more serious threat arises from the possibility that adaptation by the damaged system may alter performance in a way that does not transparently reflect the original undamaged cognitive structure. This could occur through the growth of a new compensatory-processing module or through an existing unrelated module adapting to perform the job of the missing component. Considerations of this nature have led to the need for a further assumption known as the subtractivity assumption (Saffran, 1982). Ellis and Young (1988) defined it as the assumption that "the performance of a brain-injured patient is explained in terms of the normal, intact cognitive system minus those components which have been lost as a result of injury" (p. 18). As we have indicated, it is traditionally assumed that threats to the subtractivity assumption stem from the growth of new neural-tissue/ cognitive modules or from the adoption of existing portions of the brain/premorbidly unrelated modules to perform new tasks. In this article, we will outline a new potential challenge to subtractivity stemming from the substantial readjustment of synaptic weights within the damaged modules themselves. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.