ConceptNeuronLogicalityEnergyEfficiency

Related: ConceptNeuronLogicality

Rate-based signals at the inputs and outputs of the brain have led many to presume that all the internal processing of the brain is likewise rate-based. But when you consider the types of things your brain needs to do, though, you can convince yourself that interior signals of the brain represent meaning in individual neural spikes or clusters of redundant spikes.

• Source: Brain Simulator II _ The Guide - Charles Simon
• Chapter: Chapter 4: Applications of Neurons, Frequency/Rate Detection

As an alternative to the simple ConceptNeuronLogicality: Using the slightly more complex ConceptNeuronModelLeakyIntegrateAndFire, identical logic functions can be implemented with a different scheme: Instead of continuous firing representing a logic 1, let any single spike represent a 1.

All the complexity of the biological neuron may add some efficiency but as more complex models are described, recall there is no theoretical need for them—you could do everything with the IF model and fixed-weight synapses.

The great advantage is that this network requires almost no energy when logic is inactive. We know that the brain is tremendously efficient. My brain uses only 1/10th the energy of the CPU in my desktop computer. One of the ways it does this is by not having any neuron spike unnecessarily because spiking requires energy.

• Source: Brain Simulator II _ The Guide - Charles Simon
• Chapter 4: Applications of Neurons, Saving Energy