Integrate and Fire

• Perhaps the simplest model of a neuron as an information processing element is the integrate and fire model.
• Neurons have cell bodies and axons,which can be over a metre long.
• On average, a neuron spikes about once a second.
• That sends an electrical impulse down the axon.
• The axons connect to other neurons over a synapse.
• When a spike gets to a synapse, it causes a chemical transfer of activation over the synaptic cleft.
• When a neuron collects (integrates) enough activation, it fires.
• So, as a model, synapses have weights.
• When a neuron fires, it passes that much weight over each synapse.
• The post-synaptic neuron collects the activation.
• If it goes over a threshold, it fires.
• That's the Mccullouch Pitts model (1943).
• You'll note that when a neuron is driven, it fires more than once a second, perhaps as frequently as 100 times per second.
• They do fire without activation; this is called spontaneous activation and is not part of the Mcculoch Pitts model.