of the pattern are below the AF threshold. Link weight updating then occurs, using one of several update rules, but in this case the update rule of ISV99I was used. Attention was spread using a diffusion approach by representing the weights of Hebbian links between pattern nodes within a left stochasti

s algorithm has been tested in sparse Hopfield nets, using both standard Hopfield net learning rules and Storkey's modified palimpsest learning rule ISV99I, which provides greater memory capacity in a continuous learning context. The use of key neurons turns out to slightly increase Hopfield net memory