Summary
Associative neural network models are a commonly used methodology when investigating the theory of associative memory in the brain. Comparisons between the mammalian hippocampus and associative memory models of neural networks have been investigated [12]. Biologically based networks are systems built of complex biologically realistic cells with a variety of properties. Here we compare and contrast associative memory function in a network of biologically-based spiking neurons [22] with previously published results for a simple artificial neural network model [11]. We shall focus primarily on the recall process from a memory where patterns have previously been stored by Hebbian learning. We investigate biologically plausible implementations of methods for improving recall under biologically realistic conditions, such as a sparsely connected network. Network dynamics under recall conditions are further tested using network configurations including complex multi-compartment inhibitory interneurons, known as basket cells.
See the full content of this document
Extract
Improving Associative Memory in a Network of Spiking Neurons
1. Introduction
The CA3 region of the mammalian hippocampus has, at least in subrogions, sufficient recurrent connectivity between the excitatory pyramidal neurons that the network they form could act as an auto-associative memory [1, 23]. This possibility is reinforced by evidence from tissue slice experiments of the Hebbian induction of long-lasting changes in synaptic strength at these recurrent connections [4]. Auto-associative memory function in the CA3 region cannot be directly tested experimentally. It is not possible to instantiate precise patterns of neural activity for either storage or subsequent recall. Thus, computational models are required to assess the possible memory performance of the CA3 neural subnetwork of the hippocampus.The computational models presented here are devised to address the specific problem of whether network inhibition provides suitable control of pyramidal cell activity to allow the successful recall of previously stored patterns and whether methods which address unit usage (number of stored patterns a particular neuron belongs to), level of input activity and dendritic signal enhancement can be used to increase the quality of recall. They build upon a variety of models that include different levels of biological realism in exploring associative memory function in the hippocampus [14, 17, 22, 24]. The basic premise of ...See the full content of this document
Sponsored links
