3.1.1 Information flow to muscles: neural firing rate, precision=bits per firing, information flow in bits per second

Neural firing Rate:

Muscles do not hold their tension indefinitely. Rather, the neural firing has to be repeated at approximately 1 msec intervals for muscles to retain the same state of tension.

Precision = bits per firing:

Let us assume that we can count the number of neural connections per muscle. That would give us an estimate of the maximum precision that could be reached in controlling the tension of the muscle. Even though the firings are not necessarily synchronized, we can see this count as the bits of information per firing.

Since there can be more than one muscle connection per neuron, it is better to think of the number of independent signals that go to the same muscle.

This measure indicates roughly how many distinct states of tension the muscle can provide, and thus indicates the maximum precision of control over muscle tension that the neurons can provide. In general this number has to be estimated.

• Since movements are smooth rather than step-wise, it is likely that it takes quite a few bits.
• Redundancy in signalling would increase the bit count.
• 100 to 300 bits is probably not an unreasonable number, and might be low.

Information flow in bits per second

Combining the neural firing rate with the precision, we can estimate the information flow along the neurons to a single muscle. This measure is similar to the bit rate for data flow used for computers.

• Given our very rough estimates, we estimate an information flow rate of 100-300 Kbps (kilo bits per second) for controlling a single muscle.
• For walking, we need a least 4 muscles for the hip and knee, but at least twice as many if we include ankle flex and hip rotation. For simple walking we need 0.4 to 1.2 mega bits per second.