Note that the synchronization means that the firing happens with timing that is not simply the reaction of the interconnection - hence the note that if it were just the interconnection it would have to be faster than light.
I’m still not quite clear about the idea. Is it that if given the model/configuration of the interconnections plus other neurons’ firing pattern, we can predicate one neuron’s immediate firing pattern mathematically (i.e. in no time), so the speed is infinitely high?
Consider the case of two cells that are both highly excited by some received sensory pattern. Both are switching from random very slow pulses to active higher firing rates.
These two cells are connected by reciprocal connections so that the information from these connections adds to excitation from the nominal receptive field for the respective cells.
The pair of neurons interact such a way that after the exchange of information for a few cycles the slower one speeds up and the faster one ignore the return pulse until they are synchronous.
This mechanism is more than just the reaction to the received pulse and the initiation of a pulse in response.
For the faster cell, it has already fired and the pulse arrives during the refractory period; the cell is unable to respond. For the slower cell the added excitation speeds up the oscillation bringing it into alignment with the faster cell.
This adjustment continues until neither cell is “the faster one” and they are synchronized. This timing adjustment is insensitive to the delay in propagation - the paths back and forth combined with the firing time of the synapses cancel out.
This action is greatly enhanced with the reinforcement in a larger network of cells that are participants in a distributed pattern - an engram. The length of the lateral reciprocal connections are important to enforce sparsity in the engram.