To the extent that the cognitive sciences actually consider the brain, the focus is clearly on neurons. Even the name of the field "neuroscience" suggests that neurons take the center stage. However, neurons are vastly outnumbered by glia, a different type of cell that is now known to be involved in sleep, memory, the fMRI signal, Alzheimer's and Parkinson's, although many neuroscientists are still resistant to the idea that glia are involved in information processing per se. In a recent review article, Watkins et al. focus on the role of these cells in the experience of chronic pain.
Although acute pain is easily treated by opioids, this is not the case for
"enhanced pain" (in which warmth is experienced as painful heat, or coolness as painful cold). Astrocytes and microglia are two types of glial cells thought to be particularly involved in this kind of pain.
As reviewed by Watkins et al., microglia will release proinflammatory molecules in response to activation by trauma, infection or ischemia, and may afterwards return to baseline or remain activated for a long period of time. Microglia and astrocytes can activate each other, with microglia typically responding first to trauma, and astrocytes showing subsequent changes. These changes can persist for long periods of time. Tellingly, inhibition of glial cells reduces the pain associated with a variety of stimuli.
Neurons can activate microglia as well, through a variety of molecules including chemokines, ATP, dynorphin, PGE2, TNF, IL-1 and IL-6. Watkins et al. report that chemokine injections increase pain sensitivity, by activating microglia and thus releasing proinflammatory molecules. Other chemokines seem to act similarly, as do injections of ATP into microglia. Watkins et al suggest that this wide variety of signaling molecules may indicate that neuron-to-glia signals are detailed with respect to the site, nature, and severity of neural damage.
Proinflammatory molecules not only increase neuronal excitability, but seem to make more lasting changes to the baseline activity rate of neurons (by increasing AMPA and NMDA receptor expression, by decreasing GABA receptor expression, and by downregulating glutamate transporters). This gross increase in neuroexcitability seems to underlie pain facilitation.
Watkins et al. report that similar mechanisms underlie opioid tolerance. Chronic exposure to morphine activates spinal cord astrocytes, and inhibition of astrocytes or the proinflammatory molecules they release is sufficient to slow the process of morphine tolerance. Pain facilitation due to opioid withdrawal is blocked by molecules that block glial function.
Chronic pain is often associated with reduced efficacy of opioids, and repeated exposure to opioids is often associated with increased pain sensitivity. Watkins et al. report that opioid sensitivity can be increased (both on short and long time scales) by blocking the release of proinflammatory molecules from glial cells; by blocking this function of glial cells, tolerance may become less likely.
Many claim that neurons are the "information processing units" of the brain, and suggest that glia have only a "support" role. The research presented by Watkins et al., however, clearly demonstrates a role for glia in the information processing of pain - opening the possibility that similar glia-blocking molecules might be useful in the investigation of glia's role in more cognitive information processing. In the meantime, many of these molecules - in particular interleukin-10, show extreme promise as a new treatment for chronic pain.
Clearly cognition is more complex than most modelers of cognition suspect. The glia-activated and sustained inflammation probably leads to accelerated loss of neurons. Perhaps this is why NSAID use seems to retard the development of some dementias.
I read a paper from Nature, maybe a year ago, about the control of ca++ realase by glia. If I remember, they said that glia could have a link with synchronizing neural spike !