Hacker News Reader

Comment by bigiain | Hacker News Reader

I can't help but think that what they've discovered is some meatspace equivalent to the hash key or an index key to a memory. By "turning it on or off" you can lose or find a whole table row or hash value, but surely a "single neuron" can't "store" a generalised "memory".

View on HN

A neuron "could" store a generalized memory (the amount of information it stores is gigantic, w/o pendantically considered the DNA). However, it's more complex than that (there are many types of memories and many means of retrieval). Basically, memories are sets of synaptic connections, and a neuron has many synapses (average ~7,000 synaptic connections/neuron to multiply by our 10^11 neurons/brain). As you said, it may be a "key", but it should be seen more like the memory _is_ a complete list of hashes. Not {"key": memory} but {"key1": {"key2": {"key3": {...} } }, {...}, {...}, ...} and the set of {key1..keyN} is the memory. So if you removed whichever keyI in the middle, you lose the information. (That's not really true because there is high redundancy, but there are keys/synapses/nodes less redundant than others. The fact is, they don't fire on only "one" neuron, they fire at a very precise region but the light still goes through a population of neurons.)

This complexity make my head hurt. Who created this mess of complexity anyways without leaving so much as a code comment or design document to help us sort this all out? Intelligent design my ass.

Fascinating, thanks! Have you got some links or keyphrases I should feed Google or Wikipedia to learn more (at a "curious but not neuro-science trained dilettante" level of understanding)?

I'm in this lab http://www.lppa.college-de-france.fr and mostly learned through colleagues (I'm doing CS/AI/ML, not neuroscience). I've read some parts of "Neuroscience 3rd edition, Purves et al., Sinauer Eds." which has nice pictures but I think there are better books and better advices than mines on this. Perhaps start with Wikipedia, https://en.wikipedia.org/wiki/Hippocampus // https://en.wikipedia.org/wiki/Memory_consolidation

Thanks again! stuck in the reading queue for the weekend. (Phear my curiosity, people waiting for a table in my café!)

The laser can't be aimed at an individual neuron; it is aimed at a cluster of neurons. Besides, activating a single neuron will strongly activate the neurons it has strong synaptic connections to. The key here is that an individual cluster of neurons is directly correlated with an individual memory.

I don't know if it was used in this paper, but you can obtain single-neuron targeting using two-photon microscopy.

I'd see it as the keycode / directions / a path to connecting all of the pieces; At what level all of the pieces are stored at (biological/physical in brain, in subtle energies, other?), who knows?

Reply on news.ycombinator.com