We don’t understand the human brain

I was very surprised by this article about a study, started in 2017, that reports more than 3300 types of neurons in the brain. It is an order of magnitude more than they expected to find. We don’t know what most of them do. Other studies looking at how brain cells are connected (the connectome) have also discovered complicated patterns. Apparently it may be the extensive synapses that distinguish some of us from other animals, not even the cell types.

Excerpt:

The brain atlas, a $375 million effort started in 2017, has identified more than 3,300 types of brain cells, an order of magnitude more than was previously reported. The researchers have only a dim notion of what the newly discovered cells do. The results were described in 21 papers published on Thursday in Science and several other journals.

Ed Lein, a neuroscientist at the Allen Institute for Brain Science in Seattle who led five of the studies, said that the findings were made possible by new technologies that allowed the researchers to probe millions of human brain cells collected from biopsied tissue or cadavers. “It really shows what can be done now,” Dr. Lein said. “It opens up a whole new era of human neuroscience.”

Still, Dr. Lein said that the atlas was just a first draft. He and his colleagues have only sampled a tiny fraction of the 170 billion cells estimated to make up the human brain, and future surveys will certainly uncover more cell types, he said.

Biologists first noticed in the 1800s that the brain was made up of different kinds of cells. In the 1830s, the Czech scientist Jan Purkinje discovered that some brain cells had remarkably dense explosions of branches. Purkinje cells, as they are now known, are essential for fine-tuning our muscle movements.

Later generations developed techniques to make other cell types visible under a microscope. In the retina, for instance, researchers found cylindrical “cone cells” that capture light. By the early 2000s, researchers had found more than 60 types of neurons in the retina alone. They were left to wonder just how many kinds of cells were lurking in the deeper recesses of the brain, which are far harder to study.

With funding from the National Institutes of Health, Dr. Lein and his colleagues set out to map the brain by inspecting how brain cells activated different genes. At least 16,000 genes are active in the brain, and they are turned on in different combinations in different types of cells…

Truly fascinating stuff. To me it affirms my belief that the human brain is too complex to be understood by the human brain. Those who claim we’re on the verge of being able to upload minds, memories, and consciousness into the Cloud make me shake my head. Which further rattles my brain.

Yup. It seems to me implausible that the human brain can fully understand the human brain. Wouldn’t it take something larger to do so?

I expect we’re capable of learning a lot more about it than we know so far, however; and some of the additional potential knowledge is likely to be useful.

I think the human brain will understand the human brain. Just not our current human brains.

Fortunately, our cognition hasn’t been limited to the human brain for thousands of years. Maybe a human brain can’t understand the brain, but many, many brains working in concert can. I mean, there’s no one human who completely understands computers, either, but we’re still able to work together to make them.

YOU ARE CORRECT!

According to Gödel’s Second Incompleteness Theorem, it is not possible for any system demonstrate its own consistency. To understand anything, you must look from a higher level of abstraction.

I recall a somewhat amusing (to me) line in Steven Pinker’s How the Mind Works (at least, that’s what I recall it being from). I don’t recall precisely how it went and what I remember of it wasn’t enough for me to find anyone quoting it using Google. The basic premise was that the reason that we don’t really understand how the brain works is a lack of volunteers to have their brains studied in detail while they’re still alive. Nor has technology developed since the late 90s when he wrote that to make it much easier to do such a study with live subjects. All we can do for now is to look at corpses, and that’s not nearly as enlightening as being able to do on live subjects the same sort of studies that we can do on cadavers.

We obviously already understand some things about the human brain. At any given point, we can understand slightly more. At what point do you expect there to be a barrier beyond which we cannot understand incrementally more, and why should that be?

To “fully” understand something would seem to require that we know absolutely everything about the laws that govern the universe. On that basis, there is nothing we fully understand.

I think to “fully” understand something is really not a coherently defined idea, so I don’t buy that there is any fundamental principle that something cannot understand itself. There is only ever partial understanding of anything.

…but I think I understand enough to know that this is not relevant.

Slight nitpick, but i don’t think anyone ever claimed to have any model for how the brain works that had generated any “expected” number of neuron types. Which really just underlines your point that we know very little.

This is not an argument against being able to upload minds and memories. Contemporary large-scale AI systems are too complex to be understood by the human brain, too, yet we can build them. The mind is ultimately all just information, and the prerequisite for being able to map and extract this information and upload it to an operational model of the human mind is not any sort of holistic “understanding” in the usual sense of the word, but rather the development of sufficiently capable analytical tools for doing so. Like building AI itself, this is an incremental process that ultimately produces synergies and breakthrough results in which the seemingly impossible suddenly becomes reality.

Or, to put it another way … :wink:

Maybe, but when this is said by the people who did the study so also have no reason to disbelieve it even if they did not formulate a specific number.

“An order of magnitude more than previously reported” is what they said, according to your link.

There a significant difference. More than previously expected implies that there was some prior model generating an expectation, and this research invalidates that model. More than previously reported just means “we found some cool new stuff”.

No one even has the basic idea what dreams are, and we’ve had them since time began.

I’m fascinated by memory. Sometimes I can’t remember a fact, and I probe and try things to trigger the path to get into the “memory file”. I mean, one time I was trying to remember an actor, and I couldn’t remember his name, or any movies, or “see” a movie poster and read his name off it, but my wife said “his name is an adverb” and there it was! How the hell does that even work?? Is there a “brain file” cross referenced by part of speech?

eta: no I don’t remember who it was!

We don’t have much on the function. Some areas of brain for certain memories etc. The neuron level has some breakthroughs as far as how learning is done.
See classic work of

Harvard medical school professor of molecular biology said in 2018, that if a complete understanding of how the human brain works is a path a mile long, then we’ve come about 6 inches. That’s about 1/10,000%, otherwise not much.

It’s ironic that it will take the power of a super computer (our creation) to decipher the brain of its creator. Perhaps we are in the process of creating our replacements rather than our tools. Let’s face it, we are straining against the bonds of our biological limitations. Becoming a machine race may be the answer.

Some of us do grok the whole computer, and AI software too. Starting from vacuum tube circuits, 6-transistor radios and hifi gave a good base from which to watch the whole electronics tech thing grow. A privilege.

That has been my second career, people just go ‘Wow’ when I say we made stuff used behind the scenes in Hollywood and on TV. It was actually just work, although I loved my job.

Glamour is weird stuff, it’s funny watching a sprinkle of showbiz pixie-dust make regular folks act all drunk and silly.

[ Earlier projects on traffic speed enforcement cameras were not something I bragged about - other manufacturers had death threats and actual firebombings. ]

Wow!

I’m no expert on this sort of thing but I find brain organoids intriguing. I’ve read in various places that they are not miniature brains capable of thought, which I obviously believe. But I do wonder what exactly is lacking that would, if present, make a brain organoid into a mini-brain capable of thought.