Synergistic Brain Noodling

Check it out, check it out: this is the Lion Man, my favourite artefact. It’s from the last ice age, an incomprehensible 40,000 years ago, and is the earliest evidence we have of human thought drifting beyond the realms of reality — half man, half lionOr bear? Look at those lil ears.. It’s testament to what is, and always has been, core to the human condition: a dreamy, inevitable “what-if” response to any given reality.

Luckily, also core to the human condition is a cheery “huh, I wonder how...” which is what this little essay bravely endeavours to somewhat explain. So what gives rise to something like the Lion Man, and why were humans the first (and thus far only) animal to conjure up anything of its kind? The answer must exist in our brains, the only truly exceptional part of the human body.

Brains as Machines

We begin with the cheeky materialist assumption that the exceptional nature of our mind must be due to its physical form. And to understand the form, our line of attack will be the function that form evolved for. So let’s imagine, somewhat coldly and abstractly, the brain as an information processing machine; it continuously churns through information derived from sources both external (our beloved five senses) and internal (memories, moods, plans), to produce effects both external (physical motion) and internal (new memories, mood changes, updated plans). Don’t mind me as I abstract away the vast and fantastic mechanisms of this machine and let us just focus on the form and function of its information processing.

We’re going to categorise the information going into this machine (and passed between its component pieces) as either redundant or synergistic. Information is redundant when it is used by a system for the purposes of robustness. I understand this best when thinking about the senses, which through a rough and tumble evolutionary process have been made highly redundant. Our vision system for example processes a huge amount of redundant information in the overlap between the two eyes’ fields of visionSame for ears, and probably nostrils?. Information is synergistic when a system instead integrates it with other information to create new information that is not present in any single input aloneThesis, antithesis, synthesis, baby.. Our eyes’ redundant overlap also gives us depth perception, which each eye individually would be unable to doThough, full disclosure, I’m a bit confused because this implies that the information in the overlap is both redundant and synergistic? I suppose that’s true, just depends on which system is using it..

This ability to synergise information is not unique to humans. Indeed, the primary visual cortex, which handles “basic” functions like depth perception, is ancient, having evolved early in mammalian development. However, the human brain has prioritised synergistic processing to an extreme that no other species hasThat makes it seem very active. More accurate might be: “Evolution has sculpted the human brain to drastically value synergistic processing over redundant processing.”. How do we know this? Brain scans and some neat network and information theory that I’m still on kinda shaky ground with, so I’m gonna let Andrea LuppiMy supervisor, and proprietor of all these ideas. do some of the heavy lifting:

[Regard] the brain as a dynamical system, whose temporal evolution is influenced by the current state of its constituent parts (brain regions) and their interactions. This means that the brain’s current state intrinsically carries information about its own future — information that we can decompose into synergistic and redundant contributions. Specifically, we quantify redundancy as information about the system’s future that is equally available from the current state of each of the parts (brain regions). In contrast, synergy is quantified as the additional information arising from the interactions between regions.

Join me as I work through it for myself: if we take the visual cortex as our sandbox again, we can picture a hierarchy of systems built for understanding visual input, with the aforementioned primary visual cortex (V1) handling the basics: depth perception, edge detection, etc. The secondary visual cortex (V2) begins to look at textures and specific arrangements of shapes, while V3 captures motion, and so on up the hierarchyI don’t actually know anything about this. The systems are almost certainly not so neatly delineated, but that’s how it’s presented in basic literature.. At some layer in the hierarchy, other brain regions start chipping in. If we look at a building, what we perceive as simply a building in V1-4 or so suddenly explodes into contextualised thoughts and feelings as it reaches the higher end of the hierarchy and begins percolating through the rest of the brain: the architecture, what it must be like to live there, how difficult the outside of those upper windows must be to clean... These thoughts are not the product of the visual cortex, nor are they the product of any other single system: they are the miraculous integration of many different regions of the brain. That’s the essence of synergistic information processing, as I understand it.

The Works

So the crux of the matter is this: the more a brain region’s activity is linked to the activity of another region, the more synergistic the information flow between the two regions is deemed to be. The activity of the lower levels of the visual cortex are independentI think. from other regions, it’s staying in its lane and just noticing edges and colours, therefore connections between it and other regions are computed to be mostlyRarely working with absolutes in brain wiring. redundant. Compare this to a highly interconnected region like the default mode network, a highway interchange-esque region of the brain which is active when thinking about personhood and narrative and is deeply dependent on information flowing into it from other regions, information we therefore compute to be mostly synergisticGlobalism for the brain. Though I think this is a matter of resolution and arbitrary-ish region delineation as well. If we zoomed into the primary visual cortex, we’d see lots of internal networks with their own synergistic and redundant connections between each other. It’s just when we group them into one region and call it the primary visual cortex that we can say that most of the information flowing in and out is redundant. Fractals, innit?.

So, how to measure all of this? First, we carve the brain up into 232 regionsMethods uncertain, ask me in a few weeks. and make all possible pairings of them. Then for each pairing calculate how much more accurately we can predict each individual region’s future by looking at the pair of regions rather than just the individual regionSee previous note.. The extent to which the paired regions have more predictive power than the individual region is the level of flow of synergistic information between the two brain regions. And if the paired regions have no extra predictive power than the individual region we conclude most of the information flow is redundant.

Andrea has applied this analysis to an aggregation of 100 fMRI brain scans and the results are wicked cool. We can see hotspots for synergistic and redundant processing in the brain, and then we actually have a mapYarkoni, ‘Large-scale automated synthesis of human functional neuroimaging data’ of which brain regions are linked to different fundamental functions of the brain (from auditory processing to autobiographical memory), so we can see which activities emphasise redundant processing and which emphasise synergistic processing by laying the hotspots over the map. And we see that activities like visual perception, motor activity, pain, and reading all fall on the redundant processing side of things, while social cognition, rewards-based decision making, and emotion are all highly synergistic activities. This intuitively feels right; the more intelligent/dare-I-say-human a task, the more important synergising different streams of information is. Isn’t that great? Like, we have the numbers, and we can back up our intuitions with data, neat as anything!

And then we get to feel superior about our species by running the same analysis on macaqueMonkeys. Making a note of it because I initially understood it as a funny pronunciation of “macaw”. brains, which showed that for them, redundancy is far more prevalent than synergy in the prefrontal cortex, despite the prefrontal cortex being among the most synergy-dominated cortices in humansAn exciting, morally impelling application of this research is that a macaque in a chemically induced coma will wake up when certain synergistic centres are invasively stimulated, but only if the exact centre is hit, little bit to the left or to the right and nothing happens.. And pleasingly, it’s the prefrontal cortex that underwent substantial expansion over the course of human evolution, emphasising again that our minds are honed like nothing else for taking in the world and forging out of it something wild and new and exciting — like a lion man.

Xx

P.S.

I haven’t coded a comment/compliment the author modal yet, so please write your lavish praise on a post-it and check back later. Meanwhile:

Who made the world?
Who made the swan, and the black bear?
Who made the grasshopper?
This grasshopper, I mean —
the one who has flung herself out of the grass,
the one who is eating sugar out of my hand,
who is moving her jaws back and forth instead of up and down —
who is gazing around with her enormous and complicated eyes.
Now she lifts her pale forearms and thoroughly washes her face.
Now she snaps her wings open, and floats away.
I don’t know exactly what a prayer is.
I do know how to pay attention, how to fall down
into the grass, how to kneel down in the grass,
how to be idle and blessed, how to stroll through the fields,
which is what I have been doing all day.
Tell me, what else should I have done?
Doesn’t everything die at last, and too soon?
Tell me, what is it you plan to do
with your one wild and precious life?