Do we learn to see?
The Nobel-prize winning neuroscientist Torsten Wiesel delivered one of IISc’s centenary lectures this past Monday. The lecture was about the visual system, and the very broad question of the importance of nature versus nurture. Prof. Wiesel started with a general introduction to several notable features of the visual system: the columnar organisation of the visual cortex, binocular interactions, and orientation-selectivity in cells.
The cells in the retina are like pixels in a camera, except they have a centre-surround sensitivity: a bright spot at the centre is excitatory, so is a dark circle along the outer edge. The contrast between the centre and the surround is what matters, since these cells eventually add up the various potentials and pass the signal along if the net potential crosses a certain level.
Put a bunch of these cells in a line and you have a basic apparatus for orientation selectivity. Only when all three cells fire will you have a ‘line’, for example, because the cut-off potentials in the visual system are designed to be high. Here’s a caricature. In the figure, a line like ‘A’ will light up all three cells and therefore produce a spike in the resultant potential; a line like ‘B’, however, will excite only one of the cells and will not clear the threshold potential.
There are, of course, problems with this model. One that is immediately obvious is how the number of cells required scales with the five levels of the visual cortex. This, it turns out, is still an open problem.
In trying to address the nature-nurture question, Prof. Wiesel noted that in spite of the visual cortex being organised in five levels, all of which have the structure of self-organising maps (SOMs), it is still found that visual deprivation in infancy can lead to permanent visual impairment. He described experiments in which one eyelid of a newborn monkey was sutured shut.
It turns out that the eye that isn’t shut takes over almost all of the visual cortex and, left this way, remains dominant for life meaning that the other eye cannot ‘see’. It is also interesting that if the blinds are reversed – the sutured eye is opened and the eye that was allowed to see is sutured – the dominance in the visual cortex also reverses. The brain, as is well known, is quite plastic in early life.
This, however, only happens within a critical period from birth – about ten weeks. If therefore, the animal does not ‘learn to see’ in the first ten weeks after birth, it remains visually impaired.[hr]
There is a paper from 1977 titled ‘Forest before trees’ in the J. Cognitive Psychology by B. Navon. It talks about how it is impossible to miss the global picture and look only at the details. A review of that paper is up next.
[End. Fini. Kaputski. Wiesel!]