ABSTRACT
The raw data for latency is the best value available for evaluating the hypothesis about temporal representation. Both speech perception and speech production involve the neural representation of temporal patterns which are highly structured at the scale of 50-200 msec, and also highly stable, within a defined range. A part from questions about the synaptic efficacy of signals relayed along slow conducting axons, the various biases leading to an underestimation of conduction delay from soma to synapse are rarely taken into consideration by those considering network dynamics in the cerebral cortex. Neural networks of the human cerebral cortex are likely to have the required temporal dispersion of signals to represent brief temporal patterns such as speech sounds, as well as sufficient concentration of influence at short delays to represent simultaneous patterns with adequate amplification. Excitatory synapses with Hebbian properties have been found in several parts of the central nervous system, but especially in the cerebral cortex.
