Neural network
AssigneeUniversity of Bath
InventorAlain Nogaret
Life of this patent
Estimating the term…
Abstract
A non-biological asynchronous neural network system comprising multiple neurons to receive respective input signals representing an input stimulus for the network, supply an output signal representing a spatio-temporal sequence of rhythmic electric pulses to an external system, wherein respective ones of the multiple neurons are connected using multiple mutually inhibitory links.
Claims27
An artificial asynchronous neural network system comprising a plurality of neurons each of which is operable to receive a respective input signal representing an input stimulus for the network, and to generate an output signal in response to receipt of such an input stimulus, the network system being operable to supply an output signal representing a spatio-temporal sequence of rhythmic electric pulses to an external system in response to receipt of such an input stimulus, wherein respective pairs of the multiple neurons are interconnected using mutually inhibitory links.
A central pattern generator comprising an artificial asynchronous neural network system comprising a plurality of neurons each of which is operable to receive a respective input signal representing an input stimulus for the network, and to generate an output signal in response to receipt of such an input stimulus, the network system being operable to supply an output signal representing a spatio-temporal sequence of rhythmic electric pulses to an external system in response to receipt of such an input stimulus, wherein respective pairs of the multiple neurons are interconnected using mutually inhibitory, wherein the multiple neurons are connected together in a predetermined configuration using such mutually inhibitory links, such that the neurons are operable to receive input signals representing an input stimulus for the generator and to supply such an output signal representing a sequence of periodic electric pulses for use in an external system.
A method of generating a coordinated spatio-temporal sequence of electrical pulses using an artificial neural network, the method comprising:
- interconnecting multiple CMOS-based neuron components using mutually inhibitory synaptic links by selecting a conductance value for artificial synaptic links which corresponds to an inhibitory synapse;
- providing an input stimulus for the network to cause multiple ones of the neurons to fire along a stable path in the network;
- and
- generating an output sequence of electrical pulses.
A pacemaker for use with a heart for coordinating contraction of the heart chambers by stimulating them with a coordinated sequence of electrical pulses, the pacemaker comprising:
- a central pattern generator including an artificial asynchronous neural network system comprising a plurality of neurons each of which is operable to receive a respective input signal representing an input stimulus for the network, and to generate an output signal in response to receipt of such an input stimulus, the network system being operable to supply an output signal representing a spatio-temporal sequence of rhythmic electric pulses to an external system in response to receipt of such an input stimulus, wherein respective pairs of the multiple neurons are interconnected using mutually inhibitory, wherein the multiple neurons are connected together in a predetermined configuration using such mutually inhibitory links, such that the neurons are operable to receive input signals representing an input stimulus for the generator and to supply such an output signal representing a sequence of periodic electric pulses for use in an external system, the pacemaker being operable to provide such a coordinated sequence of pulses responsive to an input stimulus.
Dependencies are inferred from the claim text.