Monday, August 27, 2012

A man's reach should exceed his grasp, though this makes for awkward masturbation

A more informative blog calls our attention to an article in the British Medical Journal:

The gist of the story --
---------- SPOILER ALERT ----------
-- is that reporters for the print media cannot be relied upon to investigate a SRS science topic themselves, preferring to skim quickly through medical press releases looking for sensational highlights that they can further dramatise, however much it distorts the researchers' actual intentions. Shock! The blogger is flabbermacked by this finding -- if not gobsghasted -- and feels that it deserves to be known more widely. "It’s an article I can’t believe I missed when it came out earlier this year".
---------- END SPOILER ALERT ----------
Obviously it would have been better if Schwartz or the BMJ had accompanied the article’s publication with a more dramatic press release.

On the other hand, without journalists inflating someone's observations out of all recognition, future generations of researchers would lack for high expectations. When someone taught volunteers the useful knack of controlling the output of a single neuron so that they could turn pictures of Marilyn Munroe on and off, reporters somehow managed to twist that into "a machine that could record dreams", but was that really a bad thing? It taught successors on the academic scene to think of a Dream Machine as a practical idea -- it had NEVER OCCURRED TO ANYONE BEFORE -- and that's how neuroscientists at Berkeley came to develop a machine to analyse activity in the visual cortex and reconstruct movie trailers. Once again science is fulfilling people's needs.


Without fictitious accounts to fool scientists into thinking that anti-gravity was a realistic accomplishment, the principles of anti-gravity would never have been discovered.

13 comments:

ifthethunderdontgetya™³²®© said...

I dreamed I was playing cards in a bar with people from high school last night, he related.
~

fish said...

In all seriousness, a question for you, Dr. Perception. The article and the scientists themselves seem to draw a parallel between visual stimulation and imagery in the brain that is not visually generated. Is that a fair comparison? My initial reaction is that they are unlikely to be related.

mikey said...

An interesting [digital] analog (sorry - couldn't resist) is the RFB Protocol for operating a remote computer with a GUI. All the actual computing is done remotely, and only the frame buffer is captured and transmitted to the remote client. The client than merely captures the I/O events and sends them back up to the remote system(s).

You get all the benefits of using as many and as powerful systems as you want, while only requiring the bandwidth for a static video image.

Your doods up there are looking in peoples frame buffers when they access stored information. That's cool...

mikey said...

Also, too, I can't wait to meet the people who find Riddled using the search term "awkward masturbation".

My kind of folks, fer sure...

Substance McGravitas said...

I dreamed I was playing cards in a bar with people from high school last night, he related.

Look out! That bulldog is giving his partner an ace under the table!

Smut Clyde said...

The article and the scientists themselves seem to draw a parallel between visual stimulation and imagery in the brain that is not visually generated. Is that a fair comparison?

I did notice the suggestion in the Berkeley paper that their "framework could potentially be used to decode involuntary subjective mental states e.g., dreaming or hallucination)".

My own theory (and it is mine) is that "visual thinking" and dreams and such as do *not* involve primary visual cortex, i.e. there's not much back-projection. I haven't seen any strong evidence that it does, anyway.

But isn't it a cool study!
(1) They have a system that can take a moving image and decompose it into a list of edges and movements and directions. Gabor filters blah blah blah motion energy blah blah blah.

(2) The researchers spend hours in the fMRI machine watching movie trailers. Eventually they have enough data for a statistical model that matches the components of a moving image from (1) -- edges & movements & directions -- to the point-by-point brain activity within visual cortex. This is a *predictive* model, specific to each individual, that can take a NEW movie and predict how the brain responds.

What they *can't* do is work backwards and convert a description of brain activity back into the unknown movie that produced it -- the model is way too complex. So they use a clever reverse-correlation trick:

(3) For a humungous database of movie clips, calculate the brain activity for each one...
Find 100 or so clips where the predicted activity comes closest to the activity from an unknown movie...
Superimpose them.

wiley said...

My favorite was the journalists writing about the antioxidants in martinis. It was a joke issue of The New England Journal of Medicine in which the martinis that had antioxidants were "shaken, not stirred".

There was also an article about girls named "Debbie" or "Betty" being more likely to get VD in that issue. The reporters didn't read enough to see that the whole thing was a parody of scientific studies--- one they published every year.

Scientists need a good laugh now and then. Bet they stopped laughing when they saw the benefits of martinis being trumpeted throughout the media as if it were the result of an actual scientific study. Fucking spoil-sports those "journalists" were.

Smut Clyde said...

Focussing the Riddled Dream Machine on ITTDGY...

Substance McGravitas said...

Scientists need a good laugh now and then.

You can't animate bodies made out of corpse parts all day long.

Smut Clyde said...

Scientists need a good laugh now and then.

People used to complain that they couldn't tell between the serious papers in The Worm-Runners Digest and the spoof ones, until McConnell eventually started printing the spoof ones at the back (upside-down, so the journal had two front covers).

Big Bad Bald Bastard said...

Is there a Jane Russel neuron? AFAF

Hamish Mack said...

The difficulty with the Jane Russell neuron is that it usually triggers the Twin Peaks neuron and the mash-up ain't pretty. Though the coffee is good.

fish said...

My own theory (and it is mine) is that "visual thinking" and dreams and such as do *not* involve primary visual cortex, i.e. there's not much back-projection.

This is what I was thinking too. I imagine (do not try to capture these images) that taking readings during dream states could show whether the visual cortex is activated or not. Not a perfect test, but presumably related.