Synaptic Delay
The connection point between two or more neurons is a short gap known as a synapse. When the action potential, an electrical signal, reaches the end of a neuron, it cannot jump the gap to the next neuron alone. Instead, chemicals are released into the gap which stimulate the next cell to fire. A short synaptic delay is associated with a chemical synapse because the chemicals are slower to release than the fast-moving electrical signal.
Certain chemical transmitters are released faster than others. In many parts of the brain, synaptic delays of several hundred milliseconds are common. The synaptic delay between nerve and muscle cells is significantly shorter to allow for fast muscle action.
Spike-Timing Dependent Plasticity
In the brain, the timing and delay of neural networks has important implications for learning. Synapses generally transmit directional signals from the first neuron to the next down the line, with no backward signaling. However, the status of the receiving neuron can change its reaction. If the second neuron has just transmitted a signal when it receives a new one, it cannot quickly recover and respond to the new input. Over time, the connection between the neurons weakens if this happens repeatedly.
If, however, there is a delay between the first neuron and the second neuron firing, the connection between the two will get stronger. More chemicals are released at the synapse, and the neurons may physically grow closer. These changes are part of how the brain creates and strengthens connections during learning.
Delay in Coincidence Detection Networks
Some neural networks in the brain are specifically designed to measure delays between signals. The auditory system uses delays to help you localize the position of sounds. When a sound from your right side travels toward your head, it reaches your right ear more quickly than your left ear. Your ears transmit this difference to the neurons in your brain, even though it is only a few milliseconds.
In the brain, a grid of neurons called a coincidence detection network calculates the delay between the two signals. After processing, you perceive the sound in roughly the correct left-right position. Because human ears are both the same height on the head, your brain cannot calculate vertical location of sound based on delay.
Artificial Neural Networks and Delay
Neuroscientists and computer programmers often create synthetic neural networks using computers to help understand the brain or to solve complicated problems. These networks are composed of units that transmit digital signals in a way that mimics the action of neurons in the brain.
Computer models of neural networks can work at a lightning-fast pace, but programmers may introduce delays to mimic natural biological processes. In a 1994 paper, scientists at the California Institute of Technology discovered that introducing delays created more stable networks. The stable oscillations of networks with delays more closely resembles the activity observed in the brain's neurons.