Posted by finelinebob on September 3, 2006, at 20:58:29
In reply to :-) » finelinebob, posted by Estella on September 3, 2006, at 6:45:57
> > Mathemagically, it's prolly pretty simple. You have A as your lowest detectible count or just set it to 1, black being 0. You have Z has the highest detected count. Violet light has a short wavelength (low number), red a long wavelength (high number). Match your lowest violet to A, your highest red to Z, and let the computer figure out the rest in-between.
>
> Okay. Thanks. I didn't know that. Though I'm worried about the computer 'figuring out' stuff...Bah! It's z-normalization essentially, without a set of negative values. Say your mapping program has the capability of displaying 98 hues plus black and white. Say your density counts range from 65 to 1380, min/max across all subject in your study. You may want to artificially (or perhaps theoretcially) set the min of your count range to 0, then again you might accept "65" as the minimal blood flow for least activation of the region studied. For a linear relationship, converting a scale of 1315 steps to 100 step is simple math -- we do it all the time to convert Celsius to degrees Fahrenheit. It's arithmatic, pure and simple.
> > Of course, the visible spectrum expressed that way is a linear progression. If your theory behind what you're studying suggests a non-linear response, then you have to get tricky.
>
> Woo hoo!!! Thats what I was thinking. Now hang on. I f*cked up my word for the week (abberant) so maybe I f*cked up my math for the week too.
>
> Linear (a):
> Non linear: (the rest
> ...
> Woo hoo. Also... Ah... Thats what log is for ;-)Logarthimic relationships are only one type of non-linear relationships. but any of them can be mapped to a linear scale with the proper transformation function.
> > Do they use linear or non linear for colour distributions on the scan do you know?No idea. I'd have to say the relationship is determined by the underlying theory. For instance, inscreased brain activity for a particular region on a particular task might show an exponential instead of linear increase in bloodflow to that region.
> High activity at tumor sites?
> Are lesions different?1) I'd imagine so. Biology is too squishy for me, but from what I understand tumors are sites of abnormally rapid cell growth. Sounds like it would require lots of artery and capillary formation. Denser blood supply - more pings.
2) Now you're getting into what **I** understand the term "brain plasticity" to mean. My father's second stroke damaged part of his language center, causing some aphasia. Any lesion cause by the stroke I'd imagine would show up blue to black.
Now, in his OT sessions, they gave him lots to read and lots of crosswords to do. He still has very limited aphasia, but nowhere near as bad as it was. Had he been getting his brain scanned while doing that language-heavy mental work, I'm guessing that area would have relatively "lit up" as his brain rewired itself around the lesion (if that's the theory you believe) or different, unexpected areas of his brain would have lit up (if that's the theory you believe).
Anyway, long time since I read anything about brain plasticity and I was never really into pure neuropsych as it was, so just my $0.02.
poster:finelinebob
thread:680731
URL: http://www.dr-bob.org/babble/20060901/msgs/682851.html