Ocular Dominance Columns in Humans and the Limits of fMRI

Functional MRI (fMRI) is a very useful technique, but it lacks in resolution making some systems difficult to study. Adams et al. show in a study of ocular dominance columns in humans why good old staining is still useful when we reach the limits of imaging.

Ocular Dominance Columns

Ocular dominance columns are areas in the primary visual cortex (V1) of mammals that show a preference in activity to one eye over the other. (I think this applies to reptiles and amphibians as well, but I don't know that for sure).

Here is a brief schematic of the visual system. The primary visual cortex is at the back of the brain in the occipital lobe.

In order to get information to the primary visual cortex from the eyes, it must travel through the retinas to an area in the middle of the brain called the lateral geniculate nucleus (LGN) and then to the back to the occipital lobe. On the way it does something rather interesting. The two optic nerves that connect the eyes to the brain cross in a place called the optic chiasm. At the optic chiasm, all the neurons from the retina that correspond to one half of the visual field -- the information from one half of the world -- go to the opposite side of the brain. This means that information from both eyes goes to each hemisphere of the brain.

This process is depicted in the figure below.

We can follow information to the visual cortex in the diagram. Look at the orange part. The orange part corresponds to all the visual information on the right half of the world. Because the eyes are lenses, it ends up on the left half of the retina in both eyes. From there it travels in the optic nerves to the optic chiasm where all the orange information from both eyes goes to the left side of the brain.

When the information gets to the visual cortex it, the neurons carrying it synapse on a layer of neurons called layer 4C. However, the neurons from both eyes do not equally innervate the cells of layer 4C. It turns out that if you look laterally across the visual cortex there is a change in the relative innervation of the neurons from one eye to another. If you were to put in an electrode into this area and record, you would find that one column of cells responds best to one eye. Then if you moved slightly to the side the next column of cells would respond better to the other eye, and so on alternating across the cortex.

This is actually precisely an experiment that Hubel and Weisel did in the visual cortex of cats to demonstrate the existence of ocular dominance columns. They put electrodes in the visual cortex of cats and then shined a light into one eye or the other. You could observe that the activity measured in the electrode was specific to one eye or the other.

There are two reasons for ocular dominance columns, a functional and the other developmental.

• The functional reason for ocular dominance columns is that by knowing where a particular image is one the retinas of both eyes, the brain can calculate the visual angles to get to that image -- called the binocular disparity -- and make some statement about the distance to the object creating that image. This is important to perception of objects in space.
• The developmental reason for ocular dominance columns is that during a critical period in mammalian development the eyes compete for innervation of the visual cortex. (For more information read this paper.) The activity from each eye during that period determines how much of the visual cortex the neurons from that eye activate. Because both eyes are equally active during this period in normal animals the amount of cortex from each eye is about the same. (Experimentally you can verify this by suturing one eye closed during the critical period. This will cause the other eye to activate nearly all of the visual cortex because it had all the activity.)

You can actually see ocular dominance columns by staining the visual cortex in specially prepared animals. This experiment is performed by again closing one eye of the animal for a period. The brain is then dissected out, and the visual cortex is flattened into a sheet. The cortex is stained for a marker of metabolic activity called cytochrome oxidase. Because in this experiment, one eye is closed the ocular dominance columns from that eye will not be active and will hence have less metabolic activity. The ocular dominance columns from the other eye will be active and will have greater activity leading to a darker stain in these areas for cytochrome oxidase.

The experiment when done in macaques produces a pattern that looks like this.

This is a reconstruction of a flattened piece of cortex stained for cytochrome oxidase. The black parts would be the parts activated by one eye, and the white parts are from the other eye.

You can see that ocular dominance columns for thin stripes and bands along the cortex.

Ocular Dominance Columns in Humans

Adams et al. address one of the difficult experimental issues in studying ocular dominance columns in humans.

If you wanted to measure ocular dominance columns in humans how would you do it? Well, ideally you could measure them with imaging, but it turns out that the columns are too narrow to be measured with existing fMRI. You are not really allowed to go around suturing people's eyes shut, so that experiment is way out. Also, most researchers -- rather than using cytochrome oxidase -- tend to want to use a radioactive tracer injected into the blood because the resolution of the staining is better. However, I am pretty certain there are laws against such things as well.

This is where the researchers get all tricky. The found patients who had monocular blindness -- blindness in one eye -- and secured their permission to get part of their brain after they died. They then performed the same experiment of flattening out the brain and staining for cytochrome oxidase to see the ocular dominance columns. This is in effect the same experiment as suturing an eye closed, but from natural causes.

What you get when you perform the experiment is the following. This is from Figure 5 of the paper.

The top portion (A) is what the staining of cytochrome oxidase looks like on the flattened human visual cortex. Cytochrome oxidase staining is brown. The middle portion (B) is what the reconstruction of this looks like showing the ocular dominance columns from the functional eye as black. The bottom (C) shows the comparison of a similar experiment in a macaque brain. Note that the scales are different.

What did the authors find when they performed this experiment on 9 patients? They found several interesting things:

• One, human beings do have a visual cortex like macaques. We knew this before, but it is always good to check directly.
• Two, the pattern of ocular dominance columns in humans is similar to macaques but more irregular in humans. Ocular dominance columns are also wider in humans but of about an equal number.
• Three, there had been some debate about variations in size of the primary visual cortex (V1) in humans. This study showed that variations in the size are common and as common as in macaques. (There does not appear to be much evidence that this difference in size affects function.)
• Four, the visual cortex devotes more space to the center of the eye -- called the macula -- because vision is sharper there. However, the amount of cortex devoted to the macula had also been debated. This study found that imaging had overestimated the amount of the human cortex devoted to the macula. This says something about the limits of imaging.
• Finally, they looked at one patient where they lost vision very early in life. Remember where I discussed that ocular dominance columns form during a critical period where activity determines the spread of innervation from each eye. If you suture one eye shut before the critical period, then that eye does not form ocular dominance columns and the other eye take over.

  The researchers looked at a patient who lost vision in one eye at the age of 4 months -- before the critical period in humans. A similar finding was shown in this patient. The still functional eye had took over the entire cortex. This shows that there is a similar mechanism of ocular dominance column development in humans.

All in all, I think this study shows how some very simple but clever experiments can still resolve issues in science. Imaging is very important in neuroscience, but it does have limits.

Hat-tip: Faculty of 1000