It’s almost certainly accurate to say that all “thought” has the same active “physical” component (salience) as expressed movement does.
A big miss we’ve had with regard to the cerebellum prior to the last decade has been the idea that “movement” is a product of separate systems than “thought”. It’s becoming pretty clear that “movement” is a type of “thought” and “thought” is a type of “movement”, and that our brainstem processes these things largely the same way.
I feel like this needs to be repeated in as many ways as possible, that “movement” and “thought” aren’t “linked”, they are the same thing as far as nervous systems are concerned.
Why is this important?
We can describe conditions like “mania” as an extremely high salience state.
We can describe conditions like “severe depression” as an extremely low salience state.
We can describe conditions like “bipolar” as salience cycling.
This far, far more accurately describes both expressed behavior and the mechanics of behavior in a general way, rather than attempting to sort expressed behavior into largely arbitrary categorizations.
It also transcends the “disordered/ordered” motif of psychiatry and allows us to preempt behavioral patterns that may not be “harmful” now but could possible develop into “harmful” behavior in the future.
It gives us a clearer lens to understand the difference between what makes the behavior “harmful” or “beneficial” and apply better training/therapy to redirect that behavior.
Importantly, it gives us the framework to understand that our attitudes about the behavior largely drive our definitions rather than a quantifiable physiological component.
It frees behavior from the prison of the “mind” and “free will” and brings it into the realm of the quantifiable and physiological.
This is a path toward greater understanding and empathy, toward truly understanding the mechanics of our behavior and creating tools which benefit both the individual and society.
Again, why is this important?
Understanding that expressed behavior and thought are the same elucidates the connection between really discussions at the core of psychology, sociology, etc.
On an individual level, we can look at concepts like “motivation” as a chain of behavior between stimuli incept and expression.
Motivation requires salience bound to stimuli in order to initiate behavior (including thought). If no salience is bound, no thought and no resulting behavior occurs. Low salience states result in lower levels of “thought”. High salience states result in higher levels of “thought”.
Each step in the chain requires a certain amount of salience to proceed to the next “expression step”. Each expression step is valence evaluated and either “inhibited” or “excited” (more like “sustained”). This is the literal physiological mechanic happening in the cerebellar-cerebral looping which takes place in the cerebellum.
Most of our behavioral training is “learning” (building behavioral responses) how to preempt these behavioral chains before they complete. Individuals “inner voice” or hidden speech is still decodeable to a pretty high degree with EMG for example, but we “learn” to preempt those chains before the rest of the chain is completed, this “inhibition” results in suppressed speech.
Vertebrate nervous systems have two synthetic general states, one that is feedforward/Glu/sensory based, which is essentially all gas no brakes. We “learn” or create behavior in this system/state by carving inhibitory exceptions into it. Effectively one half of learning is what not to do.
The second synthetic general state is the feedback/GABA/valence based system, which carves exceptions into a general inhibition. It stops all behavior from proceeding through the next loop and passes through only if matched conditions are met.
All expressed behavior (including thought and action) is the result of hierarchical construction, allowing stop/go evaluations to take place at each step. It is the metabolic differential of these two systems which determines whether behavior proceeds to the next loop.
It is this second state that most of current psychiatry is focused on.
That system’s mechanics are *entirely*/causally/whatever driven by how ponto-olivary climbing fibers are “synapsed” with purkinje cells in the cerebellum.
During early neonatal development, astroytes in the human brain for instance are initiating the connections between these brainstem circuits and cerebellar circuits. As these connections begin to establish, humans “learn” how to “move” and “think”. All developmental milestones are an artifact of how “efficient/effective” astrocytic binding of climbing fiber/purkinje connections are.
If we look at speech for instance, we must first establish the “thought” of a word, before it can be heirarchically bound to appropriate motor behavior to produce the word. This happens as the appropriate filters on the global state are applied in a specific enough way to create the behavior. These filters are general (baby talk/cooing/babbling) and eventually specify into increasingly more specific behavioral pathways to produce clearer speech.
One of the big mysteries for me has always been regarding why “schizophrenia” is so relatively late onset developmentally. Under this model, astrocytes are not binding behavioral roots 1 to 1 (or close to it) between climbing fibers and purkinje cells. In early development this is less of an issue, as we have a large pool of excess unbound climbing fibers. As this unbound fiber pool begins to deplete, astrocytes can link “unrelated” concepts, a type of “overfitting” of these connections.
There are no core “schizophrenia”/mania presentations (not the “spectrum” shit) which do not have some type of detectable movement uh.. “peculiarities” associated with it. This is an artifact of these fibers activating multiple behavioral chains due to the overfitting, and the movement is the valence system attempting to normalize out that overfitting.
For other types of “autism” (I’m describing all climbing fiber/purkinje class astrocyte interactions as “autism”), like “profound autism” we will likely be able to start finding in the autopsy data that these individuals have too few climbing fibers in the first place and are unable to bind more complex data because of it.
We should also be able to start doing more advanced types of imaging (hopefully ultrasound will work here) and we will see lower climbing fiber/purkinje activity for things like chronic, severe depression.
Even conditions like PTSD fit nearly into that model, as a similar type of effect to “schizophrenia”, where the intensity of the stimulus resulted in overfitting of behavior along these ponto-olivary fibers.
Broader questions like “why are some people more resilient” have less to do with any particular gene or “signaling” (outside of genes “guide” construction), and more to do with ponto-olivary climbing fiber/purkinje cell metabolic serves.
Why are some people good at a certain tasks (“talent”) is an artifact of these ponto-olivary fibers astrocyte bridges being particularly metabolically robust at those junctions. Carrying this to an extreme, we could call things like “savant syndrome” as possibly metabolically overweight astrocytes on particular connections.
It’s even likely that our whole construct of “cognitive flexibility” or “intelligence” are largely driven by three things, a) the number and density of ponto-olivary climbing fibers, b) the metabolic “robustness” of the astrocytes binding these fibers to purkinje cells, and c) the synaptic morphology between the bridge astrocytes and climbing fiber/purkinje cells.
Expanding on this a bit (jeez what a rant), nearly all imaging we currently do turns out be be heterogeneous garbage because all of it is measuring downstream artifacts of these brainstem connections. Much like your muscles, the more “information” (or “memory”) astrocytes acquire, the larger they get. Astrocytes that aren’t used as much are not just physically smaller, but they induce less “synaptic” modeling, resulting in reduced morphology in their local group.
When we say for instance that “schizophrenia is associated with reduced hippocampal ventral CA1 volume”, that reduced volume is not causal in any way to the “schizophrenia” symptoms, instead it is a downstream artifact of processing in the brainstem. When we say that when normalized, larger cortical volumes are associated with higher “intelligence”, those larger volumes are a downstream effect of increased brainstem processing. When we say that volumes in certain areas are reduced in say the ventral habenula complex in “depression”, again that’s downstream of brainstem metabolic activity.
All this to say that when we are looking at all this imaging and electrophys work and making assumptions about what they tell us about behavior, it’s important to understand that those regions are not the salience points of behavior, but rather a reflection of differences in valence processing. And it’s not the valence itself that is the root of behavior, but rather a “shaper” of expressed behavior.
References:
A gene expression signature in developing Purkinje cells predicts autism and intellectual disability co-morbidity status – Genetic work like this is always a gross mush, but the results demonstrating low expression of the particular targets they were looking for in Purkinje cells being associated with ID is interesting, with increased expression resulting in “not ID”. Just as interesting is that we’ve been talking about how cerebral and cerebellar cortical function are essentially inverses of each other, and here we get a tiny bit of confirmation bias demonstrating just that.
Transcriptomic analysis of isolated and pooled human postmortem cerebellar Purkinje cells in autism spectrum disorders – Whoa, some real red meat statements in this one – “PC dysfunction is the most consistent neuropathological finding to date, with as many as 75% of cases showing reductions in number.” – “Components of a condensed form of the ECM, perineuronal nets (PNNs), are highly enriched in the deep cerebellar nuclei (DCN), surrounding synapses from PCs onto DCN neurons and have recently been shown to be important for cerebellar plasticity and the formation of associative memories.” YES. – “In the present study, we isolated PCs from the surrounding tissue and found down-regulated genes in a variety of immune related pathways.” Remember the immune/cognition link? This paper feels like I lifted a lot of the discussion directly from it. Still, it’s an RNAseq paper so, you know, salt it.
Modified climbing fiber/Purkinje cell synaptic connectivity in the cerebellum of the neonatal phencyclidine model of schizophrenia – “We found that, while the global cerebellar cytoarchitecture and Purkinje cell spontaneous spiking properties are unchanged, climbing fiber/Purkinje cell synaptic connectivity is increased in juvenile mice.” (Overfitting)
Quantification of Behavioral Deficits in Developing Mice With Dystonic Behaviors – “Converging evidence from structural imaging studies in patients, the function of dystonia-causing genes, and the comorbidity of neuronal and behavioral defects all suggest that pediatric-onset dystonia is a neurodevelopmental disorder.” The opening sentence is the thesis, lol.
Glutamatergic cerebellar neurons differentially contribute to the acquisition of motor and social behaviors – Interesting in that it suggests the direct chain between “movement” and “thought” related activity.
Roles for cerebellum and subsumption architecture in central pattern generation – Yes. CPGs are vertebrate stimuli-salience binding points. The cerebellum binds on top of that salience. We see this in humans by following the olivary climbing fibers through to purkinje cells and out.
Reward-Based Learning and Emotional Habit Formation in the Cerebellum – A tiny bit tangential, but a really interesting discussion about the cerebellums role in creating unconscious hierarchical behavior stacks, e.g. binding the “word” to the “speech” in a more efficient manner than cortical/BG valence binding.
Functional Outcomes of Cerebellar Malformations – Really thorough look at the architecture of the cerebellum, from the climbing fibers to the DCN mapping functions to output. Again, demonstrates that “thought and action” function in a hierarchical fashion, composed along purkinje cells and mapped into the DCN.
Dynamic organization of cerebellar climbing fiber response and synchrony in multiple functional components reduces dimensions for reinforcement learning – Fairly certain I’ve posted this somewhere before, but it’s an interesting model demonstrating the chain between “thought” related activity and “movement” activity.
Climbing fiber multi-innervation of mouse Purkinje dendrites with arborization common to human – So, here’s a caveat, particularly with regard to “schizophrenia” and overfitting – while in early development there’s a one to one mapping between climbing fibers and purkinje cells, as we age, we can bind multiple climbing fibers into a single purkinje and this usually works out okay. However when the revese happens, single climbing fiber into multiple purkinje – this may result in an overfit.