Hacker News Reader

Comment by openasocket | Hacker News Reader

Are mice considered an effective model for autism? My understanding is we still don't know the exact mechanism of action of autism. Heck, there could be many different mechanisms, and we're lumping people with different underlying conditions into a single category because the symptoms line up. So our model of autism is entirely based on behavior. But if the mechanism of action in these mice is different from humans, you can't make any real predictions about how to treat/manage it.

View on HN

"Although findings in mice do not always translate into human treatments, the study may help to guide the development of strategies that could help..."

Pulled right from the article. Studies in mice are simply a way to point in the, hopefully, "right" direction without having to shoot in the dark with human test subjects needlessly.

A mouse model for human autism is contrived at best.

The cynic in me feels that some researchers are running out of ideas. The only experimental apparatus are poor knockout mice.

Arguably an analogy is trying to fix a compiled binary with a hex editor and replacing all your 0x11s with 0x0a. Then claiming the segfault is a good model of how another binary crashes (due to a null pointer exception or whatever really).

Oh my shit, they don't claim to have an answer. They have a direction. Not an answer. There's a very large difference. The entire language in this is extremely aware and careful that they don't have a cure. Quit acting like this is a CNN article saying "Scientists cured all disease because of a 2 person study". They have some survey based evidence that it happens in humans too. With a potential replication in mice, they feel confident it's an avenue worth pursuing. No one is holding their breath. It's an interesting study, with interesting line of evidence thought that they think is worth seeing if completely true. That's it, don't make it out to more than that. They share this information so other people can be enlightened to the idea and with professionals who are actually apart of the field help poke actual holes in the issues or offer other ideas to help guide it further. Not just "Ugh, you used mice, like, it's totally not gonna work man because it's like this one programming bug I once had."

> They have a direction.

Sure, the question is whether that direction significantly better than a randomly selected direction

IMO this isn't really relevant to the discussion; none of us here, nor the article itself imply that the mouse model is directly translatable, or that this is a cure instead of an interesting finding.

>Heck, there could be many different mechanisms, and we're lumping people with different underlying conditions into a single category because the symptoms line up. So our model of autism is entirely based on behavior.

I thought that was already clearly established.

Reply on news.ycombinator.com