0
$\begingroup$

Does anyone know the number of neurons adjoining a cross sectional area of the optic nerve and the theoretical bit rate of the nerve?

I read that the ON cross sectional radius is 3-5 mm and the diameter of a neuron is .004 - .1 mm. Area of cross section: (5 mm)^2 * pi = 75 mm. 75 mm / .004 = 18,750 neurons/cross section. The average action potential of a neuron is 1 ms in duration, making maximum signal rate of 1,000 Hz. The action potential is both a frequency (on/off/unk) and amplitude (intensity).

I read that light at min intensity signals at 0-1 Hz and max at 30 Hz. At 30 Hz that’s 18,750 * 30 Hz = 562,500 bits/sec or 562 kB/sec or .5 mB/sec. at 15 frames per second that’s 37.5kB/frame. Now I think something is wrong because this is very slow considering all three cones and rods must signal over this slow bandwidth.

They are also orders of magnitude off from the 576 MB image calculated for the human eye. I read that the nerves are really signaling contrast changes, but it seems like a large amount of bits just for contrast.

$\endgroup$
4
  • 2
    $\begingroup$ Neurons are not encoding a digital signal in the sense of computers, so it may not make any sense to think about it that way. Neurons can't fire at 1000 Hz, even if an AP is 1 ms wide. Like I've suggested elsewhere, start by just learning about neuroscience. Read a whole undergraduate-level neuroscience textbook. Purves' "Neuroscience" or Kandel's "Principles of Neural Science" are good options. I know, it seems like a lot of work, but it will save you a ton of time wasted. $\endgroup$
    – Bryan Krause
    Commented May 31, 2022 at 16:04
  • 1
    $\begingroup$ Well, upon some simplification, we're talking digital vs. analog information. An 8-bit digital image can distinguish between 256 tonal values. The exact amplitude of a neuron action potential along an axon is continuous ("analog"), and perhaps more importantly, the summation/integration of electrical information along any point in the neuronal soma and dendritic regions, before the axon hillock, is entirely continuous, and as such is theoretically "infinite-bit" in "infinite locations", digitally speaking. So bandwidth is infinity. That's what you get when you compare digital and analog. $\endgroup$
    – S Pr
    Commented Jun 3, 2022 at 12:51
  • $\begingroup$ @SPr It seems you are implying the amplitude of the signal has infinite states. That makes sense. I think in OP I confused intensity and amplitude. Am I correct saying the signal amplitude is intensity and the frequency is on/off/unknown? Or is it the other way around? $\endgroup$
    – Nick
    Commented Nov 11, 2022 at 6:44
  • $\begingroup$ Start with basic neuroscientific principles such as action potentials as suggested above. An action potential is a signal propagated through an axon of a neuron, that's the most basic idea to grasp as to how neurons send information down the 'wire'. Here, the exact amplitude doesn't correspond to intensity, either there's a signal propagating or not, so it's binary in that sense; usually signal intensity is coded for in part by frequency of action potentials and is limited by its input (stimulus) and other things like their refractory period. $\endgroup$
    – S Pr
    Commented Nov 15, 2022 at 16:15

0

You must log in to answer this question.