Fragile X syndrome (fra(X)) is the most common inherited form of mental retardation. It is characterized by constellation of impairments in linguistic, cognitive, social, and sensory/motor skills. Theories differ regarding the underlying psychological cause of the cognitive phenotype; however, many researchers agree that those with Fra(X) experience problems with working memory. A working memory deficient would express itself most profoundly in sequencing tasks. Those with fra(X) perform poor than those behavioral model and simulates task performance with an with an artificial neural network. The network model is based on me urological substrates discovered to be significantly different in those with fra(X); in particular, the hippocampus (HC) appears to be larger. The model will use data from a modified """"""""n-back"""""""" task, the sequential card playing tasks. Simultaneous processing matched typically developing participants and those with fra(X) will play this experimenter-designed card game. In this task they will be required to recall either one, two, three, or four cards in a row. The neural network will simulate how these two different populations perform. Those with fra(X) may have an abundance of neurons in the HD; paradoxically, having too many neurons leads to poor memory and slower learning. These neurons are represented by hidden units in the network model, and an overabundance of hidden units means that the model does not abstract important elements from the environment. The network too much idiosyncratic, moist, and extraneous information and this is infelicitous for sequencing. Proper sequencing relies on the activation and storage of meaningful, well-abstracted neuronal patterns.
The specific aims of this grant are 1) to complete the development of a game-like, sequential card playing task that can be used to assess visual working memory in a motivating manner, 2) to compare fra(X) and typical populations' performance on the card task, and 3) to implement a neural network that will provide a neurologically-motivated, more process- oriented account of the working memory differences between the two populations. In addition, by adjusting the parameters of the neural network model, individual differences within the fra(X) population can be stimulated. This project represents a step towards research in which network models can help to make recommendations for most optimal remediation strategy based on an individual's performance profile.
| Johnson-Glenberg, Mina C (2008) Fragile X syndrome: Neural network models of sequencing and memory. Cogn Syst Res 9:274-292 |
