I struggled to find the meat in this article and a quick google of the terms involved only really brings back the same article, which is almost 3 years old now.<p>On top of that the story is a cookie cutter rewrite of every upcoming ML/math technique which are always poised to provide some amazing insight into the human mind.<p>A more honest title would be "algebraic topologists think algebraic topology is great, academics who have employed algebraic topologists want it to be great but don't really understand it".<p>Edit: for anyone looking for the actual meat this appears to be it:<p>Objective Morphological Classification of Neocortical Pyramidal Cells (2019)<p><a href="https://academic.oup.com/cercor/article/29/4/1719/5304727" rel="nofollow">https://academic.oup.com/cercor/article/29/4/1719/5304727</a>
"They might even unravel the greatest mystery of them all: consciousness."<p>Don't get me wrong. Brains are interesting. Consciousness is a real puzzler. But... the mass of the known universe is estimated to be about 10^53 kg. In all that mass, there might be a few other mysteries here and there. Believing that the "greatest mystery" is between our two ears feels a tad bit egoistical.
Too late to the thread but I will give it a shot.<p>First of all, the Blue Brain Project (equivalent of the US Human Brain Project) has been widely criticised within the neuroscience community for its failures to invest the funding appropriately [0,1] as well as failing to produce their exaggerated claims of "brain on a computer", even 3 years after this article was published.<p>Secondly, it is insanely hard to even map in all the connections between neurons with their thousands of synapses. Imagine one neuron connects to another 1000 neurons and a cubic millimeter of mouse cortex contains about 9×10^4 neurons. Even 1mm3 of cortex hasn't been mapped yet for a rodent (mouse/rat) brain. Mapping the whole rodent brain will easily take another decade [2].<p>Thirdly, even when you have mapped the complete brain, inferring states of consiousness from it is a far fetched hypothesis and most neuroscients don't even dare to go there because it is so absurd to talk about it from incomplete simulation models.<p>Given the complexity of wiring of neurons and the sheer amount of specific nature of this wiring, building simulation models assumes that the connections between neurons are totally random which is an age old view which computational neuroscientists are now starting to accept (except the BBP, for now). Contrary to the BBP simulation models, this [3] model actually pays attention to the observed wiring from past experiments and is the closest to anything "realistic" but of course with several caveats.<p>[0] <a href="https://www.scientificamerican.com/article/why-the-human-brain-project-went-wrong-and-how-to-fix-it/" rel="nofollow">https://www.scientificamerican.com/article/why-the-human-bra...</a><p>[1] <a href="https://www.theatlantic.com/science/archive/2019/07/ten-years-human-brain-project-simulation-markram-ted-talk/594493/" rel="nofollow">https://www.theatlantic.com/science/archive/2019/07/ten-year...</a><p>[2] <a href="https://pubmed.ncbi.nlm.nih.gov/31055238/" rel="nofollow">https://pubmed.ncbi.nlm.nih.gov/31055238/</a><p>[3] <a href="https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1006535#abstract1" rel="nofollow">https://journals.plos.org/ploscompbiol/article?id=10.1371/jo...</a>
This doesn't have that much substance in the article. I hope that you haven't heard of the entropic theory of the brain, its one of the most fascinating theories and its very exciting! <a href="https://www.frontiersin.org/articles/10.3389/fnhum.2014.00020/full" rel="nofollow">https://www.frontiersin.org/articles/10.3389/fnhum.2014.0002...</a>
<a href="https://mind-foundation.org/entropy-as-more-than-chaos/" rel="nofollow">https://mind-foundation.org/entropy-as-more-than-chaos/</a>
I didn't understand the article at all, but it did make think about about how many dimensions are required for humans to understand how objects in life relate to each other.<p>I counted: a point, a line, a plane, a solid, a solid at a point in time, a solid that might exist in some possible future, and a solid at a point in time that might have existed if our past had been different.<p>That's all I could come up with (and I'm not even sure if the last one is a dimension).<p>It's probably a coincidence (there's nothing crankier than numerology), but it did make me happy that it adds up to seven.
Perhaps I am missing something, but the use of dimensionality in this article seems a bit overblown. The article says that a digraph meant to represent a rat brain formed 8-cliques in response to some input data. You could represent this 8-clique as a 7D structure ... but not necessarily. I am a bit worried that a layperson will read the article and think that the human brain is literally 7D in some actual way
> these transient structures, which appear and disappear like sandcastles on a beach<p>That's a good metaphor for building up elaborate programming concepts in your brain then someone bothers you and it slips away. So you have to sit quietly and build it back up (quicker this time).
>[what is..]consciousness<p>Isn't this an answered question? Consciousness is the sensation we get when we point our ability to predict others actions back on itself. The inverse of putting our-self in someone else's shoes. This was scientifically confirmed by both the <i>decision delay</i> experiments, as well as brain imagery that showed the same circuitry activate whether you ask someone why <i>they</i> did something; or why <i>someone else</i> did something.<p>That is not to say all the secrets of the brain have been understood or uncovered, but this seems to be searching for a more feel-good answer to a question that has largely been answered.
tldr: researchers build computational models of mice's brains then study the models through methods of algebraic topology e.g. looking at artificial neural cliques (<a href="https://en.wikipedia.org/wiki/Neural_clique" rel="nofollow">https://en.wikipedia.org/wiki/Neural_clique</a>) up-closed i.e. at a level of detail unobtainable with an actual slice of brain tissue, and measuring their dimensionalities<p>Here are the stuff the project published so far: <a href="https://www.epfl.ch/research/domains/bluebrain/blue-brain/publications/" rel="nofollow">https://www.epfl.ch/research/domains/bluebrain/blue-brain/pu...</a>