A technique is described which verifies neural activity to a very faint continuous sine wave through subtraction of two different far-field whole nerve action potentials from one another. A brief transient is presented to an animal in order to elicit a supra-threshold action potential. The technique is then repeated, but on the second trial a near-threshold sine wave is mixed with the transient and another action potential is collected. The resultant evoked is then subtracted from the evoked potential generated by the transient alone and a small but persistent difference potential is acquired that presumably represents the unit activity occupied by the continuous sine wave. Four experiments are presented to show the validity of this technique, along with a surprising stability of the derived-response latency despite a 30 dB range of the probes. The technique may have promise in predicting behavioral responses to sinusoids acquired from individual animals.