eftychia: Perrine (fluffy silver tabby) yawning, animated (yawn2)
Add MemoryShare This Entry
posted by [personal profile] eftychia at 02:45pm on 2008-04-08

I was unable to fall asleep last night. So y'all get a post like this'n.

The more I learn about feline vision, the more I want to know about feline neurology and what learned and hardwired techniques are used to compensate for the limitations of their eyes (including the extent to which cats augment vision with sonolocation when leaping, pouncing, and swatting at moving targets).

According to what I've found on the web, feline visual acuity is in the 20/200 range, they can't focus as close as humans can, and the feline retina lacks the hi-def center region that the human retina has. (I'm not sure whether distortions caused by the vertical-slit pupil matter, because when I see a cat about to pounce, the pupils are usually open wide enough to be round -- but I welcome confirmation or contradiction of that from other observers.) But they're successful predators, and sometimes execute some impressive moves in pursuit of thrown or dangled toys (and, in some cases, fingers and toes). I wanna know how they do it, what kind of filters are implemented in the retina and optic nerve, what preprocessing is done in the brain, and what learned techniques cats employ to compensate for apparently having relatively low-res eyes (a price of their very useful low-light capabilities). Based on behaviour, I would have thought they see about as well as we do, other than that nifty bit about only needing one seventh as much light, anyhow, so they don't appear to be handicapped by the trade-offs in the eyes they wound up with.[*]

(Then again, there are still some things I'd like to understand better about how we see, as well. I could've sworn I'd posted recently about the business of human vision being able to resolve misalignments of line segments where the offset in the image is well below the 'pixel size' of the retina, and about my brain doing a 'software zoom' thing that feels like it's in my eyes even though I'm sure it's in my brain -- but at the moment I can't find those entries. Did I start writing them and never finish/post them? I suspect the answers there are related to my cat-vision questions. And maybe that "batters can't see the ball any more by the time they swing" business as well.)

Unfortunately, two of the things I want are a super spiffy remote functional-MRI setup -- so that it can map brain activity while the subject is leaping and pouncing and swatting at things -- and a boatload of neurology larnin' so I'd have a prayer of making sense of the fMRI data. And neither of those is going to be very easy to acquire. ( [info] realinterrobang suggested I invent remote-MRI ... I'm wondering whether the experiments could just be conducted inside the MRI machine with the cat un-restrained, and write a program to key in on a distinctive skull feature or something to re-algn the images afterward to compensate for the cat's movement around the chamber.)

So for now I'm trying to reverse-engineer my cat by taking note of exactly what kinds of mistakes she makes in the catching-small-missiles department, and hoping that there'll be patterns in the data obvious enough for even me to figure out. Not that the results will be likely to be unambiguous anyhow (did they ever figure out which of the competing hypotheses about how humans catch fly balls was correct?), but it's a start, and a lot cheaper than an MRI setup.[**]

In the meantime, of course, if there are any professional or amateur veterinary neurologists reading this ( [info] anusara, do you still check in here ever?), drop me some clues, eh? Useful search terms at least? Experiments to suggest (that don't involve surgery, of course)?

[*] This brings to mind another question that I didn't think of until I started typing this up: how far back did the feline eye features evolve, and what was the eye they evolved from like? (And yes, I'm aware that I'm probably horribly oversimplifying the very question, but you can translate to what I meant to ask, right?)

[**] I mean cheaper than renting time on one, of course; I don't think the machine would fit through my doorways, and I bet it draws more current than my house wiring can supply.

There are 16 comments on this entry. (Reply.)
geekosaur: orange tabby with head canted 90 degrees, giving impression of "maybe it'll make more sense if I look at it this way?" (Default)
posted by [personal profile] geekosaur at 06:59pm on 2008-04-08
I'm wondering whether the experiments could just be conducted inside the MRI machine with the cat un-restrained
If nothing else, you may successfully prove that cats can teleport.
 
posted by [identity profile] dglenn.livejournal.com at 07:29pm on 2008-04-08
I thought that was already established ... (though demonstrating it would hinder the experiment I'd planned, yes).
twistedchick: watercolor painting of coffee cup on wood table (Default)
posted by [personal profile] twistedchick at 07:19pm on 2008-04-08
I misread that as 'feline numerology' and started to wonder if it would be in base-4 or base-5...
 
posted by [identity profile] dglenn.livejournal.com at 09:11pm on 2008-04-08
Dayum! There's a combination of words I did not expect to find together! I wonder whether the Enoch Pratt Library gets that journal.
 
posted by [identity profile] weskeag.livejournal.com at 09:38pm on 2008-04-08
If they don't, JHU does
 
posted by [identity profile] weskeag.livejournal.com at 09:41pm on 2008-04-08
And I know UMCP has it (I've used it there)
 
posted by [identity profile] jmax315.livejournal.com at 10:08pm on 2008-04-08
IIRC, the old Amateur Scientist column had one on nuclear spin resonance detection in the late 50s or early 60s. Non-imaging, and with a very small sensitive area (a few cc), but it'd be a place to start. I have the CD at home; will look.
 
posted by [identity profile] weskeag.livejournal.com at 10:22pm on 2008-04-08
The one in the above reference apparently can image within a 180-cm sphere
 
posted by [identity profile] dglenn.livejournal.com at 08:58am on 2008-04-09
180cm is enough room for the experiment I have in mind. But the abstract mentions a 180mm homogeneous region -- was 'cm' a typo or is there a useable-though-not-distortion-free region ten times the diameter of the homogeneous region?
 
posted by [identity profile] weskeag.livejournal.com at 11:03am on 2008-04-09
All I know is what the abstract says, and cm was probably a typo :-(

Hm. Maybe it's time to get on the bike & go down to my local university for a looksee at the full paper....doing that today may be challenging...as $EMPLOYER actually has me working in the $OFFICE this week.
 
posted by [identity profile] en-ki.livejournal.com at 07:55pm on 2008-04-08
One of my idle schemes of the form "gee, if I had two friends who were experts in the relevant field and knew a venture capitalist or two, I could have a great startup here" has been to combine an fMRI setup and a motion-capture system, using the motion-capture data to realign the fMRI as the subject moves.

One possible problem is that, with enough movement on the part of the subject, the brain will move inside the skull, which the motion-capture system will not catch. (You might be able to align via feature detection instead: I gather MRIs are sequences of slices, not sequences of points.)

Another problem AIUI is that the MRI system has to physically move to capture different slices, so you would need to take a large number of extra slices to have a hope of capturing everything.
 
posted by [identity profile] jmax315.livejournal.com at 08:35pm on 2008-04-08
I think you're confusing MRIs, which are volumetric data (albeit over a restricted volume) with CAT scans, which are very slice-y.
 
posted by [identity profile] en-ki.livejournal.com at 08:45pm on 2008-04-08
Is the volume more or less slice-shaped, or what? I went to Wikipedia to refresh my memory (it having been a couple of years since my last fMRI) and it claimed that MRIs were slicey too.
 
posted by [identity profile] jmax315.livejournal.com at 10:04pm on 2008-04-08
Well, it is disc-shaped, but it's a _much_ thicker disc; more of a squat cylinder. And the actual volume that an MRI could read data from is even bigger; it's just a non-uniform field outside of that volume. The easiest way to deal with that for normal MRIs is to ignore data outside the uniform field volume, but you could probably correct for the distortion if you needed to (say, to track a moving cat...).

[edit: to/too blunder. Bah. Why do I even pretend to speak English?]
 
posted by [identity profile] jmax315.livejournal.com at 08:34pm on 2008-04-08
OK, first, while reverse-engineering cat, be very careful. I think you'd be fairly heartbroken if you didn't get her back together properly.

I'm wondering whether the experiments could just be conducted inside the MRI machine with the cat un-restrained, and write a program to key in on a distinctive skull feature or something to re-algn the images afterward to compensate for the cat's movement around the chamber.)

Yes; it's a fairly simple (said the guy who hasn't written code to do it) extension of the existing registration algorithms already used sometimes for MRI data. CPU-intensive as all hell, though.

Links

January

SunMonTueWedThuFriSat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24
 
25
 
26
 
27
 
28
 
29
 
30
 
31