Explaining temporal resolution with water-works of the visual system
Most people are confused about temporal resolution. That includes my students, and even BBC science programmes. So I created this diagram to communicate the basic concept, with the example of human visual processing, using a water-works metaphor.
Why water-works? I’m trying to explain an unfamiliar concept in terms that everyone can understand intuitively. By using a hydraulic metaphor for the nervous system, I’m following in the footsteps of Descartes, who in the 1600s knew almost nothing about the brain or even nerves but nevertheless had a pretty good notion of what was going on:
And truly one can well compare the nerves of the machine that I am describing to the tubes of the mechanisms of these fountains, its muscles and tendons to diverse other engines and springs which serve to move these mechanisms, its animal spirits to the water which drives them, of which the heart is the source and the brain’s cavities the water main.
Here’s my more modern, yet still hydraulic, take on temporal resolution of the visual brain:
At top is the stimulus, which consists of two patches. The top patch alternates between green and red, and the one below alternates between leftward-tilted and rightward-tilted. This image is projected onto the retina. The retina processes the stimulus somewhat before passing it on to the cortex (via the thalamus’ lateral geniculate nucleus), where one population of neurons determines the stimulus color, and another set of neurons determines the stimulus orientation. These processes have high temporal resolution, meaning that they determine color based on a relatively short interval of signals. This is why we can perceive colors correctly even when they are presented at a rapid rate, say 9 colors per second.
The resulting color signals ‘pour’ into the binding process, which has poor temporal resolution. A long interval of signals must accumulate before the process can compute the feature pairing. For the presentation rate depicted, the consequence of the long integration time is that multiple colors and orientations fall within an interval that is essentially simultaneous from the perspective of the binding process. The binding process cannot determine which color and orientation were presented at the same time. At this rate, we can perceive which colors were presented and which orientations were presented, but not the pairing between them (Holcombe & Cavanagh 2001). Click here to see this for yourself.
Large bucket = long interval of signals mixed together before the output is determined= poor temporal resolution = long integration interval = “slow” process.
But the last term in this equation can get us into trouble, because in everyday language the word ‘slow’ conflates temporal resolution with latency. My next post will add to the illustration in an attempt to make the distinction clear.
Rene Descartes (1664). Traite de l’Homme (Treatise of Man) Harvard University Press (1972)
Holcombe AO, & Cavanagh P (2001). Early binding of feature pairs for visual perception. Nature Neuroscience, 4 (2), 127-8 PMID: 11175871
[UPDATED 9 August 2011 to remove reference to non-existent movie link]