T. Alvager, V. French, G. Putman, D. Herrmann, E. Anderson, and S. Schnitzer (USA)
Aging, cognition, artificial neural networks
The present research investigated whether an artificial neural network model could provide an adequate account of the decline in cognitive function with aging. The model proposed in this work, called here the Compression model, assumes that changes in learning with age involves the compression of data in a person's brain. Applied to aging, the model rests on the assumption that hidden units in the artificial neural network represent synaptic connections among neurons in the human brain. Furthermore, the model represents the observation that the density of synaptic connections among neurons in a brain changes with age by reducing the density of hidden units in the artificial neural network. Using backpropagation and recurrent neural networks, the model is shown to be capable of accounting for age effects on learning.
Important Links:
Go Back
