a) By detecting the direction of propagation of the signal using the magnetic flux characteristics surrounding axons and fasciculi b) Identifying the nerve terminal morphology c) Identifying the biomolecules in the nerve terminals and the adjoining synaptic clefts

Thanks Sunil

Antidromic responses follow high frequency short trains which is the most commonly used method (ability to faithfully follow of short trains at 300Hz) since synaptic transmission will fail. However, usually one also obtains additional information to prove that a response is antidromic: constant latency, discrete all or none threshold to stimulation, but most importantly demonstration of collision of the antidromic spike with an orthodromically conducted spike within the predicted collision interval. There are details of these methods in some of the older neurophysiological literature and most papers using the technique will reference one or more of these. I suggest you look at some of these reports.

Criteria listed by dr. Dostrovski are the more reliable when recording extracellularly.

A higher certainty level is only reached with intracellular recordings when, by decreasing the stimulus intensity to just below threshold, one can exclude the existence of an EPSP underlying the spike.

This depends on the technique as Prof. Baldissera pointed out. Using intracellular recording one can discern whether action potentials exhibit very short latencies and are not associated with a graded synaptic response, if using electrical stimulation. Also, antidromic responses to electrical stimulation lack the synaptic delay associated with orthodromic responses. Therefore, antidromic responses occur in close temporal association with the stimulus, whereas orthodromic responses will demonstrate 3-5ms delays...if they are monosynaptic....more if polysynaptic.