next up previous contents
Next: Brain imaging techniques Up: Neuroscience Lecture 2 Previous: Neuroscience Lecture 2   Contents

Single neuron electrophysiology

Around 1950 - microelectrodes record the activity of individual nerve cells.

Ratliffe and Hartline (1959) Studied retina of horseshoe crab by stimulating photoreceptor cells with bars of light and measuring activity in the ganglion nerve cells to which they were connected. Found that ganglions were more responsive to the edge between light and dark rather than the centre. Explained by positing excitatory inputs along with inhibitory ones from neighbouring detectors (``lateral inhibition''). Cell's response depends on the difference between them. Ganglion cells connected to receptors near the edge would receive as much excitation as those connected to the centre but less inhibition, so giving more response, so functioning as edge detectors. (See fig. [*].)

Figure: Edge detection - Stillings et. al., 1987, p. 281, fig. 7.6
13#13

Lettvin, Maturana, McCulloch and Pitts (MIT, 1959) found neurons in a frog's eye that responded to specific patterns. There were separate detectors for edges, moving contrast, overall dimming, and interestingly, for convex edges. Together they could detect small, dark roundish shapes in motion, thus seeming to serve as ``bug detectors''.

See Lindsay and Norman, p. 76 (crab-cat)

David Hubel and Torsten Wiesel (1962) worked with cats. Recorded activity further along the neural path ... in the visual cortex. Shared (with Roger Sperry) the 1981 Nobel Prize for Medicine and Physiology.

See: http://www.cim.mcgill.ca/ siddiqi/308-558-2001/james_biology.ps.gz

Recording elctrodes were placed in the cat's ``striate'' or primary visual cortex (pvc) while other stimuli and finally bars of light were presented to the eye.

These results thus suggested a hierarchy of processing cells. But one problem with this model is it does not say where the hierarchy will end - surely there aren't specific cells for each perception? Besides empirical evidence (eg. response latencies being shorter for complex than for simple cells) does not support a simple hierarchical model.

Hubel and Weisel also found that the visual system of the cat contains some abilities that are ``wired in'' at birth and others that develop with age and experience. The visual system will not develop, in fact parts of it will atrophy, if the cat is not exposed to appropriately patterned light after birth (eg. only horizontal bars). The cat must be exposed to a visually varied environment, and be allowed to move about and explore it using both eyes. The timing of these early experiences can be very specific.

Critical periods have been studied for vision in cats and monkeys. In the context of language, where too critical periods are known to exist, the vision results have been used as analogy.


next up previous contents
Next: Brain imaging techniques Up: Neuroscience Lecture 2 Previous: Neuroscience Lecture 2   Contents