What Does the Cerebellum Do?

Pianists take note- fine motor and rhythym control happens thanks to this part of the brain. But it isn't just for motor control anymore, either.

The cerebellum is the mini-brain appendage that has finer crenellations than the cortex, as much surface area, (when unfolded), more neurons, more regular structure, and has long been associated with fine motor control, judging from cases where it is defective. But in recent decades, its functions have ramified and now are understood to affect many core brain functions like cognition, pain, and affect, in a supplementary way. Just as we have supplemented computers with special processing units like GPUs, evolution seems to have devised a separate processing unit for our brains.

Removing the cerebellum does not generate paralysis, but severe deficits in movement control (fine-ness, rhythm, timing, balance) as commanded by the higher levels of the cortex. That means the cortical motor commands are not routed entirely through the cerebellum, but are copied to it and supplemented by its outputs on the way to the spinal cord. In evolution, it started as a small module to improve balance (becoming what is now the most primitive part of our cerebellum, the flocculonodular lobe). This gradually extended, in mammals, to refining all sorts of motor control, in the central areas of the human cerebellum. And finally, its lateral lobes are now interconnected with many areas of the neocortex, including executive, memory, and other non-motor locations, evidently to refine, based on feedback, many aspects of our cognition. In the evolution of humans, the cerebellum changed the most, of all brain regions, between Neanderthals and ourselves, suggesting that even so late in evolution, better fine control, whether of motor, social or other functions, became dramatically more important, perhaps through such activities as the creation and use of our many tools, of stone, wood, fibers, etc.

Outline of the typical circuitry of the cerebellum. Main inputs come from the left, vi mossy fibers (MF), which touch directly on the output DCN cells. Their major processes, however, go to granule cells (GC), whose axons form a vast parallel array innervating the dendrites of purkinje cells (PC), which in turn inhibit the deep cerebellar nuclei neurons (DCN), which provide outputs. Separately, by the major theory in the field, some error inputs come into the inferior olive (IO), which has extremely strong inputs to the purkinje cells and can change their long-term behavior, thus constituting training.

The remarkable thing about the cerebellum is its structure- a regimented, once-through architecture that can not have reverberating, recurrent connections like more complex parts of the brain, but instead is massively parallelized, featuring purkinje cells with large but flat pancake-like dendritic trees, shot through at right angles with the parallel fiber axons of the granule cells. The flow of information is input via mossy fibers to the granule cells, which activate the purkinje cells, which inhibit the dense central nuclei cells, which are the source of all outputs. The dense central nuclei cells also get some inputs directly from the input mossy fibers. The overall logic seems to be that the granule cell - purkinje cell circuit selectively dampens what would otherwise be a direct input-output from the mossy fibers through the dense central nuclei cells.

One functional map of the cerebellum, from a very interesting general review of its functions. It is clear that while motor functions are strongly represented, the cerebellum engages many other cognitive issues.

Additionally, center parts of the cerebellum that are most relevant for motion are topographically mapped to body regions, much as the sensory and motor cortexes of the cerebrum are. This supports the idea that the main cerebellar function is a very regimented, if hugely adjustable and sensitive, information transformation from input to output.

There is one more input, from the inferior olive, which gets inputs from the spinal chord, and higher levels of the brain. These neurons have activating processes going to the deep central nuclei and particularly strong connections (climbing fibers) to a purkinje cells, one climbing fiber per target cell. These connections are strong enough to overwhelm all the granule cell inputs, and are thought to be the key "training signal", which, in response to pain or error, adjusts the strength of the granule cell-purkinje cell network. This seems to be what is happening under the hood, piano playing-wise, when one has the jarring experience of hitting wrong notes, and gradually finds that the fingers unconsciously and spontaneously learn to avoid them. What was perhaps a stop-gap tuning mechanism for critical needs of accurate motion turned out, however, to have wider applications.

Detailed tracing of the connections between hippocampus (injection site) and the cerebellum (imaged above). A virus was injected, which travels slowly in retrograde fashion up the axons of neurons projecting from other regions of the brain, in this case, neurons projecting from the cerebellum into the hippocampus. The images show staining (brown) of regions of the cerebellum, with cell bodies in blue.

A recent paper looked at hippocampal connections of the cerebellum, which seem to mediate spatial orientation / navigation- another fine-tuning kind of process. Defects in these connections are seen in autism, for instance. Experiments in mice show that the cerebellum provides some inputs to, and affects and alters activities in the hippocampus, known for roles in short-term memory, and navigation / orientation / mapping. These researchers undertook to track the detailed connections between the two areas, and also established that some portions of each organ oscillate together, at the theta (6-12 Hz) frequency. This oscillation is very strong in the hippocampus, characteristic of being in motion or needing short-term memory, and known to function in spatial navigation. Indeed, they sampled individual purkinje neurons in the cerebellum (of mice) that were phase-locked with this hippocampal rhythm. And they found that for some of these areas, the coherence of the rhythms increased detectably as the mice learned a new navigation task. The cerebellum, as all brain areas, has various rhythms of its own, and to find that some of those may entrain, or at least functionally correlate with, those of other interesting regions of the brain, is very interesting.

"... oscillations within the theta range are thought to support inter-region communication across a wide variety of brain regions. Our finding that cerebello-hippocampal coherence is limited to the 6–12 Hz bandwidth is in keeping with previous studies on cerebro-cerebellar communication in which neuronal synchronization has been observed between the cerebellum and prefrontal cortex."

• Stiglitz on Facebook.
• Should the government ensure full employment?
• Over the Rainbow, quiet arrangement.
• Lies, lies, and more lies.
• Heaven, retracted.
• Sleazy moronic rapist, still president.