Structure of the nerve cell
Thought and behaviour arise from the operations of the brain and nervous system. The building blocks of these are neurons or nerve cells. Sensory neurons are activated by sensory stimulation in the form of light, sound, touch, and so on, but most neurons receive stimulation from other neurons. Each neuron consists of a soma (or cell body), dendrites, a single axon, and axon terminals (or terminal buttons). The soma contains the nucleus of the cell, which is where protein synthesis occurs. The dendrites are branchlike structures that receive chemical stimulation, in the form of neurotransmitters, from neighbouring neurons. The axon is a fluid-filled tube that carries electrical signals down to the terminal buttons. The terminal buttons contain neurotransmitters, chemical messengers that are released into the synapse, which is the gap between the terminal buttons of one neuron and the dendrites of another.
The action potential
Within a neuron the signal that occurs is electrical in nature. The fluid inside and outside the cell contains electrically-charged molecules called ions. In a neuron that is at rest, the difference in charge between the extracellular fluid and intracellular fluid is about -70mV, and is known as the resting potential. Because of this difference in charges the neuron is said to be polarised. If stimulation from a neighbouring neuron exceeds a particular threshold then the membrane potential reverses, creating an exchange of ions. This is known as an action potential. The action potential is always of the same size; that is, it does not depend on the strength of stimulation. All that matters is that stimulation is sufficient to trigger the action potential. To view an animation of the action potential, click here.
Many neurons, especially longer ones, are covered in a myelin sheath that helps protect the integrity of the signal that passes down the axon. The Nodes of Ranvier help conduct the signal along the axon. These are tiny gaps between the myelinated segments of the axon where ion exchange across the membrane is possible. Essentially, in myelinated neurons the action potential proceeds in a series of jumps from one node to the next.
When the signal reaches the terminal branches at the end of the axon, there is some possibility for the strength of the signal to diminish, and so the release of neurotransmitters is not guaranteed. However, as long as sufficient signal does reach the terminal buttons then neurotransmitters are released into the synapse. The action potential is explained further in the following video:
Neurons and information-processing
How do computations actually arise through the action of nerve cells? In 1943 Warren McCulloch and Walter Pitts proposed an artificial neuron they called the Threshold Logic Unit (TLU), also called the McCulloch-Pitts neuron. McCulloch-Pitts neurons took binary inputs and produced binary outputs, enabling them to compute logical functions such as AND and OR. Later, in 1957, Frank Rosenblatt developed the Perceptron, a neural network that took matrix eigenvalues and produced binary output. This was the forerunner of modern connectionism.
Important biological evidence about the operation of neurons was obtained by David Hubel and Torsten Wiesel, who discovered that neurons in the visual system of cats were specialized for particular tasks. By inserting microelectrodes into cats' brains they were able to identify different types of cells. Simple cells only fired when presented with bars of light of a particular orientation; that is, some cells would only fire if shown a horizontal bar, other cells would only fire if shown a vertical bar, and so on. Complex cells would only fire if shown a bar of light of a particular orientation moving in a particular direction. This work won Hubel and Wiesel a Nobel prize (shared with Roger Sperry). In the following video David Hubel describes how this finding came about by accident:
Some further footage of the Hubel and Wiesel studies, using various dots and bars of light, is shown here: