In the neuron autonomous processes happen. Here we attempt to describe them in an abstract way. It is challenging, however, to write without physical and physiological details, so we frequently refer there and back to the corresponding chapters.
As discussed, controling the operation of its synapses is a fundamental part of neuronal operation. It is a kind of gating and implements an ’autonomous cooperation’ with the upstream neurons. The neuron’s gating uses a ’downhill method’ for gating: while the membrane’s potential is above of that of the axonal arbor, the charges cannot enter the membrane. As soon as the membrane’s voltage exceeds the threshold voltage, the synaptic inputs stop, and restart only when the voltage drops below of that threshold. The synaptic gating makes interpreting neural information and entropy, as we discuss it in [16] and chapter 5, at least hard.
For the cooperation of neurons, it is of fundamental importance to synchronize their operation. The neurons have low accuracy, while their concerted actions need precisely synchronized pulses (about two orders of magnitude shorter time resolution than the inter-spike intervals!).
Given that the voltage gradient is the pace of temporal change, a faster rush-in current in the upstream neuron (seen as a steeper slope [63]) can evoke firing, independently from the membrane’s voltage. This observation, alone, underpins that exceeding a a voltage threshold leads to firing. Receiving a synchrony signal forces an instant firing. After firing, the first synaptic input sets the time base as we have discussed it above.
It is interesting to note that, according to Shannon, a single spike does not carry information, given that the shape of the spikes are identical, only its time can deliver information. And, yet, a single spike can carry the information that a new collective operation (of neurassemblies) begins and the participating neuron’s operation must synchronize their ”local time” to a remote basetime. In the sense of time-space, the signal resets the time base of all receiver neurons to zero. That is, all their synchronized upstream neurons will reset their timebase to that synchrony signal. Consequently, the neuron will receive its input spikes on a relatively well-defined scale, despite that the sender neurons send their spikes at different absolute times; by automatically ”calculating” and applying the needed offset time. The neuron’s frequency stability is low, so the synchrony signal (the base frequencies) must be repeated relatively frequently for the system’s stable operation. Of course, the neuron does not know the absolute time. The local time’s starting time is the time when the first synaptic input arrives and its range of interpretation ends when a new computation starts or when the membrane’s potential goes back to its resting value.
It might also happen that (also depending on the residual membrane voltage) the outgoing spike’s delivering begins immediately after last spike arrives. Given that the rising edge delivers the important timing information, and the voltage gradients contributions received before the last spike somewhat faded in the meantime, one can understand Hebb’s observation in terms of learning: the last spike (before firing) contribute more than the ones received earlier.