What a great article! For those of you interested in cognition, our brains have almost precisely the same problem of temporal delay and prediction. Each of ma

What a great article! For those of you interested in cognition, our brains have almost precisely the same problem of temporal delay and prediction. Each of many sensory-motor systems has about 50 to 150 milliseconds of jitter and offset. And the jitter and offset in timing depends on many factors—-intensity of the stimulus and your state of mind for example. How does the brain help “consciousness” create an apparently smooth pseudo-reality for us from a noisy temporal smear of sensory (sensory-motor) input spread out over 100 milliseconds or more? It is damn hard, but the CNS plays the same games (and more) as in this great Medium article—interpolation, smoothers, and dynamic forward prediction. Just consider input to your human visual system: color-encoding cone photoreceptors are relatively fast to respond—under 30 msec. In contrast, the rod photoreceptors are much slower integrators of photons—up to 100 msec latencies. So even at one spot of the retina we have a serious temporal smear between two visual subsystems (two mosaics). It gets much worse—the far periphery of your retina is over a centimeter from the optic nerve head. Activity from this periphery connects slowly over unmyelinated fibers. That add even more temporal smear relative to the center or your eye. And then we have the very long conduction delays of action potentials going from retina, to first base—the dorsal thalamus—, and then finally to second base—the primary visual cortex at the very back of your head. That is a long distance and the nerve fibers have conduction velocities ranging 100-fold: from 0.3 meters/sec to 30 meters/sec. The resulting input to layer four of your visual cortex should be a complete temporal mess ;-) But it isn’t. What mechanisms “reimpose” some semblance of temporal coherence to your perception of what is going on “out there”? Neuroscientist do not spend much time thinking about this problem because they cannot record from millions of neurons at different levels of the brain. But here is a good guess: the obvious locus to interpolate and smooth out noise is the feedback loop from visual cortex to dorsal thalamus. I mentioned the dorsal thalamus (formally the dorsal lateral geniculate nucleus) as “first bade” to visual system input. Actually it get 3X more descending input from the visual cortex itself—a massive feedback loop that puzzles neuroscientists. This huge descending recurrent projection is the perfect Bayesian arbiter of what makes temporal sense. In other worlds the “perceiver”, the visual cortex, provides a descending feedback to its own input that tweak synaptic latencies in dorsal thalamus to smooth out the visual world for your operational game playing efficacy. This feedback obviously cannot remove all of the latency differences, but it can clean up jitter and make the latencies highly predictable and dynamically tunable, via top-down control. Quite a few illusions expose this circuitry. for more @robwilliamsiii or labwilliams@gmail.com