Christodoulou, Chris and Bugmann, G. and Clarkson, T.G. and Taylor, J.G. (1993) The temporal noisy-leaky integrator neuron with additional inhibitory inputs. In: Mira, J. and Cabestany, J. and Prieto, A. (eds.) International Workshop on Artificial Neural Networks: New Trends in Neural Computation. Lecture Notes in Computer Science 686. Springer, pp. 465-470. ISBN 9783540567981.
Abstract
The Temporal Noisy-Leaky Integrator (TNLI) neuron model with additional inhibitory inputs is presented together with its theoretical mathematical basis. The TNLI is a biologically inspired hardware neuron which models temporal features of real neurons like the temporal summation of the dendritic postsynaptic response currents of controlled delay and duration and the decay of the somatic potential due to its membrane leak. In addition, it models the stochastic neurotransmiller release by the synapses of real neurons, as pRAMs are used at each input. Using the TNLI, we investigated the effect of synoptic integration between excitatory and inhibitory inputs on the transfer function of the neuron. We observed that inhibitory inputs increase the fluctuations of the input current and reduce the slope of the sigmoidal transfer function of the neuron, which highlights one of the differences between biological neurons and formal neurons.
Metadata
| Item Type: | Book Section |
|---|---|
| School: | Birkbeck Faculties and Schools > Faculty of Science > School of Computing and Mathematical Sciences |
| Depositing User: | Sarah Hall |
| Date Deposited: | 16 Mar 2021 19:04 |
| Last Modified: | 09 Aug 2023 12:50 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/43531 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
