Hacker News Reader

Comment by appleflaxen | Hacker News Reader

  The Internet has countless local area networks that then connect with larger, regional networks and ultimately with the backbone of the Internet. The brain operates in a similar way

These kinds of "news releases" from universities are really annoying to read. How could the brain be set up any other way than "local network areas connected by a bigger network"?

Is that really even notable?

These types of "stories" are simply fluff PR for academic research, and (IMO) get more traction on HN than they deserve given the lack of substance.

View on HN

Seriously. The term is scale-free network[1] and they could have just dropped it at that. It's not really even a new discovery, we've known that animal neural networks are scale free for at least ten years.

[1] https://en.wikipedia.org/wiki/Scale-free_network

Actually, yes - there are many topologies that a rat's brain could have but apparently doesn't.

For example, nature could have favoured more ring or star topologies: https://en.wikipedia.org/wiki/Network_topology.

Well then it would have been remarkable find indeed, as few would expect neural configurations forming a perfect loop or having a single point of failure.

Why not? Larger evolutionary structures do seem to have a single point of failure, e.g. the heart.

One neuron dying off is much more probable than a fatal heart condition. Heart is a multi-cellar, regenerative organ which often recovers from substantial damage.

Saying that the brain is organized like the web is functionally misleading but morphologically correct. Both show strong "small-world" network topologies (short mean path-length, high clustering coefficients) but do not function in very similar ways at all.

What is the difference between morphologically and topologically in this case?

There might still be some substance here, but it is all written in a very sensationalistic way. Like, we already knew the brain was modular, so the local networks were a given. We also knew the brain has hierarchical networks aggregating information from lower systems, higher up. We also know the brain handles local system-failures fairly well, reducing the space of possible network-topologies as far as I can understand it.

So, to me it seems like the only news-worthy stuff in there, is that they are actually mapping this stuff and wrote a paper on it.

Furthermore I see no link to the actual paper, or a title of it, making it harder to find...

There are many possible network topologies.

https://en.wikipedia.org/wiki/Network_topology

>These kinds of "news releases" from universities are really annoying to read. How could the brain be set up any other way than "local network areas connected by a bigger network"?

Obviously as one big network without any local network areas?

It's pretty evident that local here doesn't mean merely topically local, but of semi-autonomous functioning (separate utility, etc).

Didn't we already know that pretty solidly? Like the chunk at the back is for your eyes, etc [that's the only one I can remember consistently]

I guess we did.

But that's a different issue than asking "How could the brain be set up any other way" as if any other structure is impossible (as opposed to merely proven not to be the case in practice).

Reply on news.ycombinator.com