What is it about?
Designing a pulse-coupled neural network (spiking model) that reflects behavioral functions is a non-trivial problem especially if one were to maintain objective validity. A sensible approach is to design the pulse-coupled network on the basis of a level-coded neural network (firing rate model), i.e. from a higher level model to a finer grain model. This paper demonstrated the step-by-step construction of a pulse-coupled network capable of fear-like and relief-like responses.
Photo by Alina Grubnyak on Unsplash
Why is it important?
Since we assume that level-coded neural networks are closer to psychology while pulse-coupled networks are closer to biology, the pulse-coupled network must be developed step-by-step to maintain knowledge continuity. Any large leaps during the construction can result in the pulse-coupled network loosing objective validity with regards to behavioral function.
Read the Original
This page is a summary of: Construction of a pulse-coupled dipole network capable of fear-like and relief-like responses, Connection Science, May 2016, Taylor & Francis, DOI: 10.1080/09540091.2016.1185393.
You can read the full text:
The pulse-coupled neural network analogue of the level-coded network.
The PCNN analogue of G-DN. The network receives pulsed inputs. Solid lines indicate excitatory while dashed lines imply inhibitory connection. Connections for only one (highlighted) ENU per field/sub-field are shown for clarity. These are the highlighted ENUs.
The following have contributed to this page