I’ve goofed on this topic pretty hard in the past, and still continue to make clarity goofs, so this is partly an explainer for myself to check concept coherence. Will source this up later.
Central pattern generators are astrocytic circuits which create a baseline rhythmicity to brain function. This rhythmicity allows functional groups in nervous systems to coordinate potentially asynchronous metabolic execution into a sequential series of reactions which ultimately make up behavior.
As an example of this, the heart is able to beat without being connected to external nervous system signalling because the heart has a set of CPGs which cooordinate activity local to the heart, and our limbs are able to perform complex independent movements because of the pattern generators located in the spine (for vertebrates).
On top of that, we have a core central pattern generator (this I should be more clear about when talking about CPGs) in the medulla, which provides a baseline signal to all the other CPGs in the nervous system.
When we think of nervous systems, it’s important to understand that they are not a single homogeneous unit, but instead a core unit with a bunch of functionally distinct “modules” built on top of it. Every single distinct “feature” that occurs in a vertebrate is almost always accompanied by a CPG circuit built into the group that enables that feature.
This extends beyond hearts and limbs into any “nuclei” or gross anatomical structure inside of the brain. The amygdala group will have it’s own CPG circuits, the habenula will have it’s own circuits, etc.
CPG circuits have a couple of pretty cool features. First, they allow state comparison between firing “external” to the functional group and it’s own baseline rhythms. Using our heart for example, the mechanic that hearts use to “know” how to speed up is comparing incoming input to the CPG vs. internal rate. The CPG then attempts to adapt internal firing rate to match incoming signal demand. When incoming signalling strength drops below a threshold, the CPG is again responsible for primary rhythm rates.
A large part of “natural ability” for any skill is largely a description of how tightly coupled these independent CPG circuits are with each other. Someone who is very coordinated likely has very synchronized movement group CPGs, someone who is “clumsy” (that is makes bad movement predictions) likely has CPGs which require more effort to synchronize.
This includes how “talented” or “untalented” people are at “thinking” constructs like how proficiently someone can process a specific cognitive task. Even “stuttering” is behavioral, and thus an artifact of internal<->external CPG integrations.
Now, because CPG circuits are astrocytically controlled, they also have another cool feature… they are “trainable”. A large part of early childhood development is composed of exactly this, learning how to integrate all of the various clocks in the system together, and as the clocks come into sync more behavior becomes available. The more we practice a particular task, the more nuanced the synchronization between CPG circuits.
Think of this as just another layer of the cellular internal<->external conceit, where instead of jumping straight to our top level functions, we likely have a few layers of “internal/external” grouping in between (which are the basis of structures like “nuclei”).
For any behavior or function, the “dysfunction” to “super function” dichotomy is quite likely an artifact of a) how adaptive local<->external synchronization is and b) the granularity of “learning” in a function group’s CPG circuits.
Anecdotal/Aside: One of the weirdest artifacts of most of our imaging techniques is that “activation” or “effort” appears to have an inverse relationship with “performance”. Put another way, during the “learning” phase of tasks nervous systems use a tremendous amount of activity, but a skilled or practiced version of the same uses almost none.
This completely flies in the face of the intuitive expectation, that “higher performance equals higher activation”. This has spawned countless stupid dorm room conversations like “We only use x percent of our brain, what would it be like if we used all of it at the same time!” (answer: you’d be dead, that’s a massive grand mal).
What we see in practice are these astrocytic CPG circuits going through an initial learning/synchronization state, once synchronization for a particular set of inputs is learned and encoded into local group neurons activation continues to decrease as long as no novel information is projected to the local group.
So what does this mean in the context of behavior, particularly “mental health” type constructs? It means that ultimately our goal should be to focus on discrete areas of high activation during the performance of certain tasks, and figure out how to speed up or slow down RNA transcription in that particular region. Someone who has a phobia for instance, will show a particular set of activation patterns when exposed to that phobia. By either speeding up or slowing down metabolic performance in the affected functional group’s CPG circuits, we can greatly reduce the burden necessary for the group to “learn” the necessary synchronization.
Edit: It’s kind of interesting thinking about drugs, especially hallucinogens, as fucking around with the CPG interactions in the nervous system.
Extending on this conceit, maybe we can start talking about stuff like “addiction” or “addicts” in the context of a brain trying to suppress or synchronize misaligned clocks, e.g. an alcoholic is an alcoholic because their dorsal stream and ventral stream clocks are misaligned, GABA nuking balances them out.
However instead of building the necessary behavior to synchronize the clocks when they are in a closer state, they get dependent on the chemical itself to synchronize.
We phenomenologically experience bad synchronization as distress, the greater or more locally intense the clock misalignment, the greater the distress.
Thinking about PTSD as an example of what happens when “bad learning” occurs and breaks sync when exposed to particular stimuli. Depending on where the circuit adapted would determine behavioral output, e.g. if adapted CPGs were near core brainstem CPGs, the effect is global and severe, if it’s embedded as part of a second or third level mechanic in a functional group (e.g. embedded a few levels down in the hippocampus) you get a much more specific “trigger”.
Shiiiit… Adverse Childhood Events are more severe and system altering because brainstem functional modules come online first and the higher level modules don’t come on line until later. Because the initial CPGs “train” later developing modules, they inherit similar responses.
Damn, that makes them pretty much hell to clean up as well, can’t simply “relearn” that particular module, you have to find the core offending module and fix all the downstream ones too. And god knows what type of dependencies that initial core has built internally on top of the original CPG learning. Wow. That’s uglier than I thought. Maybe this is why ECT is so effective, it nukes the “learning” for a little while.
Going back to addiction, sometimes those metabolic interactions in CPG circuits can be pushed in a direction or to a level that the organism simply can’t replicate on it’s own, creating a dependency on the external chemical to use the functional module. I’m thinking of it like being dependent on a particular non-endogenous amine. That module is always going to be firing in a distressed state when it doesn’t have the dependent chemical to match it’s new learned state, feeding a vicious cycle.
Can imagine other modules would try to “learn around” input generated by that, but on a simple mechanical level the modules probably only have a bayesian response, and are going to steer behavior toward shutting down distress rather than either shutting down/reducing interaction strength of the module.
Hah, under this conceit, caffeine headaches/withdrawals are a great example of the physiological/phenomenological effect of poorly matched CPGs due to external chemical manipulation of clocks.
Ref Dump, will integrate later
Addiction:
Dopamine-Evoked Synaptic Regulation in the Nucleus Accumbens Requires Astrocyte Activity
CPGs:
Astrocytic modulation of central pattern generating motor circuits
Multiple intrinsic membrane properties are modulated in a switch from single- to dual-network activity – Heh, I’ve never seen the phrase “stomatogastric nervous system” before, cool! This might be better recognized as one of the core functional networks.
Intersegmental Interactions Give Rise to a Global Network – YES. The “global network” is an artifact of CPG functional group reconciliations.
A role for inherited metabolic deficits in persistent developmental stuttering
Evaluation of recurrent GNPTAB, GNPTG, and NAGPA variants associated with stuttering
Astrocyte function from information processing to cognition and cognitive impairment
Physiology of the Digestive Tract Correlates of Vomiting
Forward Stepping Evoked by Transvertebral Stimulation in the Decerebrate Cat – Creepy as fuck, but useful for paralysis research.
Firing behavior of single motor units of the tibialis anterior in human walking as non-invasively revealed by HDsEMG decomposition – Which is the same thing as the cats, just way less fucked up.
Behavioral evidence for nested central pattern generator control of Drosophila grooming
Breathing Rhythm and Pattern and Their Influence on Emotion
Peptidergic modulation of a multi-functional central pattern generator in the pulmonate snail – Heh, ultimately, this is what it all comes down to. This is how “brains” work.
Serotonin as a volume transmission signal in the “simple nervous system” of mollusks: From axonal guidance to behavioral orchestration – Yep, although all “neurotransmitters” are functionally bayesian “volume” signals for different functional groups, it’s the “co-expressed” peptides/proteins that contain the actual “data”. This is actually cool enough I wish I had broken it out into another post.
CO2 exposure enhances Fos expression in hypothalamic neurons in rats during the light and dark phases of the diurnal cycle – Brainstem CPG circuits coordinating complex behaviors like sleep/wake cycles.
Distress:
Cell-type-specific epigenetic effects of early life stress on the brain
Long-Term Impact of Early-Life Stress on Hippocampal Plasticity: Spotlight on Astrocytes
Models:
Heteroclinic cycling and extinction in May–Leonard models with demographic stochasticity – As a thought exercise
Computational modeling predicts regulation of central pattern generator oscillations by size and density of the underlying heterogenous network – Yep, this model needs an adaptive CPG and it should being able to make some pretty substantial predictions.
Central pattern generator based on self-sustained oscillator coupled to a chain of oscillatory circuits – For my robotics bros looking for better movement algos…
Development:
Fmrp regulates neuronal balance in embryonic motor circuit formation – This is here mostly as a follow up reminder – Are CPGs all generated in the same phase of development as divisions or do they develop as discrete units?
Misc:
Association Between Temporal Asymmetry and Muscle Synergy During Walking With Rhythmic Auditory Cueing in Survivors of Stroke Living With Impairments – AKA, “how do brains deal with asynchronicity?” Kind of a cool idea isn’t it? To modify the strength of other inputs to CPGs to make up for weakened/undesired levels of inputs.
Convergent effects of neuropeptides on the feeding central pattern generator of Aplysia californica – CPGs are Adaptive
Neuromodulation Enables Temperature Robustness and Coupling Between Fast and Slow Oscillator Circuits – “Temperature” is a measure of energy.