Posted by finelinebob on September 2, 2006, at 18:41:59
In reply to Re: Biopsychosocial vs Biological Reductionism » SLS, posted by Estella on September 2, 2006, at 11:59:34
> > > The colours that are chosen to represent different levels of activation are completely arbitrary.
>
> > They are not arbitrary. Red is hot and blue is cold.
>
> So the scans show the temperature of different parts of your brain?
>Pardon me, but let me just interject some color theory into the mix. Lots of ways of describing color, but let's talk about HSV/HSB -- hue (color), saturation (intensity of the color compared to gray=0) and value or brightness (where it is on a scale from black to white) --the common colorspace for Photoshop, for instance. What I find interesting about most color mapping techniques is that while Hue is used to map different values of what is measured, Saturation is always at like 100%. To choose red to indicate high amounts of X being measured and violet to measure low amounts of X being measured, that choice might be completely arbitrary although the visual spectrum is "respected" in mapping colors inbetween deep violet and deep red.
But choosing red for "hot" meaning high energy levels and blue for "cool" or low energy levels is not completely arbitrary. In Western Culture, those two color-energy relationships are widely prevalent, so picking them to represent level of activity for whatever your machine measures, converting the numerical to the visual, is a natural "bias" to follow. People can relate to it readily.
What may NOT be arbitrary, tho, is the Brightness. I don't see this so much with, say, mapping non-visible wavelengths of radiation to the visible spectrum for x-ray spectroscopy of stellar objects. However, with mapping brain scanning technologies and what they measure to the visual spectrum, there is a distinct bias in lowering the Brightness of "cooler" areas. Although the colors remain vivd as much as they can, they do fade to black.
AFAIK, there is no scientific reason for this. Changing the brightness not only does not add to the interpretation of, say, a PET scan, it actually introduces a non-existent variable or, rather, a variable indicating the "goodness" or the "badness" implicated by the colors.
Any particular unstable radioisotope will decay, and the decay event may have a variety of outcomes. We know of isotopes whose decay events have one particular outcome that accounts for, say, 90% of all possible outcomes. That decay event will release a photon of a precise wavelength. When such an isotope in introduced into the bloodstream, areas of increased bloodflow will naturally display a greater number of events than areas with decreased bloodflow.
In other words, we are looking at a single dimension, a single degree of variation: the number of decay events. Any visual mapping of that information requires exactly one visual element to be altered. So, all any brainscan needs to do would be to map event counts to Hue OR Saturation OR Brightness. The use of Hue AND Brightness (with red having B near 100 and violet having B near 0) introduces a bias ... which I think depending on the situation can favor a favorable or unfavorable dx. An "unbiased" mapping would have no "fade to black" -- in those edges of the brain/skull/scalp where the bloodflow terminates, there is no topographical reason not to maintain an equal value for all hues until you hit a region of no recorded events; which could be white or black according to your preference as the mapper.
That being said, I have no problem with biases in scientific observations since I don't believe in objectivity -- but you need to acknowledge your biases for what they are if you're going to have any degree of intellectual integrity or honesty.
As an aside, I always find it a bit amusing that red tends to be assigned to high evergy levels and violet to low energy levels. The physicist in me just wants to giggle at how they got it backwards.
poster:finelinebob
thread:680731
URL: http://www.dr-bob.org/babble/20060901/msgs/682418.html