Hacker News Reader

Comment by vgoh1 | Hacker News Reader

So the paper describes a way to make new neurons. Any indication that adding new neurons will actually improve cognitive abilities? I know nothing about neuroscience, but I could see a scenario where it is analogous to adding a computer to a computer room, with nothing on the hard drive, and no network connection.

View on HN

If it's anything like an artificial neural network, adding new neurons or new neuron connections likely causes degradation in the overall abilities of the NN. Having neurons and connections is important, but the connections themselves have to be learned, selected, pruned to have overall value to the system.

I too would like to know more. Extending on the computer analogy, even if it is just adding more blank drives — does that mean your brain has additional _capacity_ for cognition?

Or is it more like those people who inject oil into their muscles in an attempt to look more buff without the work (and often end up looking grotesque instead).

i think most applications of neuronal cell therapy or regenerative medicine have focused on thinks like stroke. my understanding of the field is limited, but from what i understand the current tools are sort of blunt instruments. there are a huge amount of subtypes of neurons, and i dont think we are able to exert detailed control over which subtypes of neurons are created, and i think that after a certain number of divisions neural stem cells "forget" their subtype

i also dont think we understand enough about the biology of cognition to know how many of what kind of neurons to put where. not to mention any adverse effects, like synapses forming with non-target neurons, etc

It might be really useful to try to repair some non-cognitive areas of the brain. I.e. grow some new neurons to implant in the motor cortex that can relearn how to control limbs for people who had damage there.

What about just restoring substantia nigra in PD patients ?

Exactly, or what about strange new memories, i guess it would depend wholly on what layer of the neocortext the cell was converted on, but still, it could be very interesting.

Reply on news.ycombinator.com