Sunday, March 04, 2018

Harmonizing?

LSD is 'harmonizing' for the brain — and can change your personality for years, studies find".

???

The study found that taking LSD is “harmonizing” because it helps connect different parts of your brain in new ways while “reorganizing” it. The effects were temporary, but Newsweek reports that it’s a positive sign for people with some psychological conditions.

Wait, I thought they just said "for years." There's more...

And some in the tech industry have started to take LSD in small amounts to increase their productivity.

...

“I had an epic time," he told Rolling Stone after one day at work. “I was making a lot of sales, talking to a lot of people, finding solutions to their technical problems.”

Actually, that might explain some things happening in software recently...

But onward--the actual paper is Connectome-harmonic decomposition of human brain activity reveals dynamical repertoire re-organization under LSD The first thing that jumps to mind is "what in the world is connectome-harmonic decomposition?"

I'm glad you asked. Basically it is just a harmonic analysis of brain surface electrical patterns. Despite what they claim in the text, I don't see that the various harmonic patters bear necessary resemblance to conectivity--they appear to just be the spherical harmonics mapped onto a brain instead of a sphere. (And the text bears that out--this models the surface only. Other tools are supposed to address inner workings--I think.) The other examples plotted do not all fill me with confidence: e.g. the "mammal-skin"-shaped plates were supposed to show something significant? Hint: in a real mammal one boundary condition on the skin is that it wraps around.

This quote jumped out at me:

Taken together, these results suggest that the visual, somato-motor, and limbic, as well as the default mode network, are well predicted by individual connectome harmonics of a narrow frequency range, whereas the higher cognitive networks rely on a broader frequency range of connectome harmonics distributed over the spatial spectrum.

Or in English: the network activity for several brain functions can be predicted using only a few frequencies and harmonic patterns. That's a lot stronger than I think they can claim. Their measurement of network activity may involve only a few frequencies, but that's not the same as network activity as a whole.

One well known problem with using spherical harmonic decomposition of a pattern is that measurement uncertainties (and any sorts of gaps in the measurement coverage) make the higher frequency harmonics pretty meaningless. Been there, done that, got nothing.

So, maybe there's something to look at, but I'd be very careful about drawing conclusions.

Oh, yes, where did that "harmonizing" come in? The only use of the word in the paper is in the name of an ethics committee.

My guess is that it either comes from the vivid imagination of the science reporter, or a translation problem in talking with the Spanish researchers.

1 comment:

  1. Thanks for linking to spherical harmonics.

    The idea is a bit forced, don't you think? They wanted a headline, and made the paper fit.

    ReplyDelete