Jonah Lehrer interviewed Andrew Koob on the Scientific American site (here) about the importance of astrocytes, one type of glial cells. It seems that we will not have even a rough idea of how thought works until we understand these cells.

Interestingly, as you go up the evolutionary ladder, astrocytes in the cortex increase in size and number, with humans having the most astrocytes and also the biggest. … cell counts in the brain revealed glial cells to be nearly 90% of the brain (this is where the neuron based idea that we only use 10% of our brain comes from). … many researchers have completed experiments on the communicatory ability of glial cells with neurons … it was discovered glial cells respond to and release ‘neuro’ transmitters. … calcium waves are how astrocytes communicate to themselves. Astrocytes have hundreds of ‘endfeet’ spreading out from their body. They look like mini octopi, and they link these endfeet with blood vessels, other astrocytes and neuronal synapses. Calcium is released from internal stores in astrocytes as they are stimulated, then calcium travels through their endfeet to other astrocytes. The term ‘calcium waves’ describes the calcium release and exchange between astrocytes and between astrocytes and neurons. Scientists at Yale, most notably Ann H. Cornell-Bell and Steven Finkbeiner, have shown that calcium waves can spread from the point of stimulation of one astrocyte to all other astrocytes in an area hundreds of times the size of the original astrocyte. Furthermore, calcium waves can also cause neurons to fire. And calcium waves in the cortex are leading scientists to infer that this style of communication may be conducive to the processing of certain thoughts. … In this theory, neurons are tied to our muscular action and external senses. We know astrocytes monitor neurons for this information. Similarly, they can induce neurons to fire. Therefore, astrocytes modulate neuron behavior. This could mean that calcium waves in astrocytes are our thinking mind. Neuronal activity without astrocyte processing is a simple reflex; anything more complicated might require astrocyte processing.

This just shows how far we are from understanding the brain.