brain and vision

how does the brain interpret the images ?

 

Log in or Sign up to hide all adverts.

huh, if you are talking about humans then:

light gets detected by the eye, the eye then converts what it sees into a series of electrical impulses which it sends to the brain, the brain reinterperts the impulses much as a screen does, with each impulse benig a different pixel, but with millions af pixels not just the 1024 or 1280 that we have on screens

 

Log in or Sign up to hide all adverts.

The path of light.

First to the cones and rods of the eye which are located in the retina. Changes in light intensity trigger the receptors to fire. To aid this the eyes vibrate finely (not sure of the amplitude or frequency of this motion) this causes receptors at high contrast boundarys to fire more rapidly, highlighting edges. This data moves from these receptors through several layers of nerves. Alot of processing is done in these layers. Edge detection, motion detection, noise reduction and filling in the blanks from blind spots. Adult human eye can have as much as 50% of the retina damaged. At this point the signals leave the eye and move deeper into the brain. The eye has about 125 million receptors but the optic nerve leaving the eye only has about 1million. This is because the information has be generalized.

The paths continue into the brain crossing over each other. One half of each optic nerve bundle crosses over so that each half of the brain see one side of the view.

Then onto the optical lobes near the back of the brain. At this point the left and right images combine. There are many layers of nerves here and the path of the signals are not direct. Nerves also travel in and out of this point sending and receiving signals from various parts of the nervous system. Spacial information is generated here, assorted lines and points with motion become objects, we are also preprogrammed to recognize certain objects like the human face. This area also controls which things to concentrate on, inhibiting unwanted information.

As we travel deeper into the Optical lobes more and more generalized information is generated. This is where a face becomes a recognized person or not, where a group of letters become a word with related meaning. Actions are generated here. Dodge or catch the ball, steering to avoid an obstacle.

Then onto the back of the brain.. This is where the high level stuff happens.. This area has input/output from almost all areas of the brain. We put true refined meaning to objects. This allows us to send signal back down to the Optic lobes reducing the strength of the signals to dodge the ball because you know its a 3D projection. To recognize a person so that you can place a name to them and a host of other meanings.

 

Log in or Sign up to hide all adverts.

Gaussian’s first post is excellent thread opener – hope thread lives long b/c it is a great mystery how the brain interprets an image to form a perception. I have a published opinion – too complex to post.

You can send Email to helen.worth@jhuapl.edu and ask for copy of "Reality, Perception, and Simulation: A Plausible Theory" (APL Technical Journal, volume 15, number 2 (1994) pages 154 - 163.)

Although most of my paper is a direct answer to Gaussian’s question, the last section notes that my answer also provides an answer to the classic freewill-vs. -Determinism problem.

What vslayer responded is true, but not really responsive to the question. The problem with the idea that there is a larger number of “neural pixels” displayed in the brain, is: “Who is looking at these pixels?”

Blindman’s first two paragraphs are also true, but after that there are errors, especially the idea that “right and left images combine” Exactly the opposite happens. In V1 (also called the primary visual cortex) begins a process of splitting apart various characters of the isomorphic image map (neural pixel image, if you like, still exists in V1&V2). For example, the features of color and “speed of movement” are sent to other entirely distinct (physically separate) neural tissue for additional processing. Exactly how many “features” of the image are so split off and sent to different neural processing areas is not firmly known. Most researchers would agree that at least 5 to 8 are processed in separate areas but up to 20 have been suggested by some.

This is a fundamental puzzle b/c the features never seem to rejoin anywhere in the brain. Neural science has learned a lot about how the image comes apart into separately processed features but has no idea how or where the unified perception we all see is formed. (This is the puzzle I address in my paper.)

One point Blindman makes is excellent and important. A primary function of the two distinct layers of processing cells in the retina is data compression almost 100 to 1. His noting that it is essential for the eyes to “quiver” even when staring at one object is also good. Many experiments have been done that defeat this natural quiver. In the early days, before modern eye movement sensors were adequate, a light stiff shaft was actually glued to the eyeball. It held a tiny lens and slightly more distant an image on a small piece of film negative that was projected into the eye (distant light in dark room was the light source). Thus even though the eye continues to quiver the projection system moves with it and the image does not move on the retina. For example, the image might have been: T . What happens perceptually is most interesting (and used as part of my theory). When the light is first turned on “T” is accurately perceived for a few seconds, then one of the two line segments that make up the “T” is not perceived, even thought sill on the retina. That is a horizontal or vertical line is perceived. Same is true of more complex images. Typically various components (recall these separated processed “features”) are perceived in different combinations. Retina cell fatigue (same thing that makes you perceive the complimentary color when looking at a white wall after fixed staring for a few minutes at a colored spot) is an important part of the reason why selective fade out of feature elements occurs, but not the whole story as never does one perceive only the bottom half of the vertical segment of the T etc.

I will not go into details but to further prove that processing in the brain is also involved, search about the “waterfall effect.” (One can perceive an object is both stationary and yet simultaneously estimate its rate of motion from your separately processing of motion feature by different brain cells that have become fatigued by staring several minutes at steady moving object, like a waterfall.)

 

optical lobes....what the hell?? occipital lobes?

and is the question about the higher function of understanding images or the lower function of processing images...
There are many branchings to the optical path..and not all are accounted for in the major one which branches into 2(dorsal/ventral what vs where) then 3 or 4. which go to many other paths...

 

The optical lobes opps, Lateral Geniculate Nucleus is the correct name, they are located infront of the occipital lobes.
Thanks Billy T. I am but an amateur. It is interesting to note that the information crossover at the optic chiasma is not even. I am reading up on the new research.

 

Neurocomp2003 you are correct, but might want to look at “Separate visual pathways for perception and action” by M.A. Goodale and A.D. Milner in TINS, Vol. 15, No. 1 p20-25. Or at book ISBN 0198521367 (“Visual Brain in Action” by same authors, Oxford Univ. Press, 1995) Sorry about the age of these references, probably better are now available, but I retired to Brazil almost a decade ago and have lost touch with the recent literature. (Part of the reason why I am responding and hope with your help to raise this thread to the level it deserves.)

They persuaded me that the classic “what and where” paths are better understood as “what and how.”

Their view is based in part on the fact that when the monkey only has to look at the symbol on the lids of the food wells, he does not need to shift his gaze away to the “landmark” reference. The original experiments ignored this task difference.

I also want to strengthen your statement about many paths in the visual system (not all even go through the LGN!) After the LGN, some tracks go to the SC (superior colliculus), not to V1 (primary visual cortex in the occipital lobes).

This SC path is probably responsible for the amazing things a monkey can do some weeks after both his entire occipital lobes have been removed (I should use “ablated,” the term which helps confuse the animal rights people, but I prefer to be forthright about what is done.) For example, if you throw a peanut on the cage floor of such a “cortically blind” animal, most will, without hesitation or confusion, pick it up with a well-guided hand motion and eat it!

Blind humans exhibit “blind sight” but never reach this level of performance.

You seem to be able to contribute significantly to this thread. Please get a copy of my paper (see prior post) and tell me what you think of my theory as an answer to gaussian. I am especially proud of the “viewer in movie problem” I invented in the paper, which IMHO, destroys the traditional cognitive science view that we seen by “successive computational transforms” of the retinal data. (P. Churchland et. al.)

In answer to your question about higher or lower level of gaussian`s question I want the thread to be about how we perceive with vision, not a bunch of details about where various computations take place in the brain, although this information, illusions, cerebral accidents and ablation studies etc can be very helpful in answering the difficult question.

I am sorry I can’t send it to you by Email. I will try to get a pdf version from Johns Hopkins so I will be able to do so directly instead directing you to the Editor of the APL technical digest.
If you want to know more about me and my current concerns, visit site under my name.