Sensitivity was 20 μV/division, sweep speed was 2 ms/division, duration of pulse was 0.1–0.4 ms, and the machine used was a Tektronix electromyograph. Low-frequency filter, high-frequency filter, and rate of pulse were not specified.

This study was performed in the supine position.

Following the antidromic method [1], signals were recorded using surface electrodes from the median and ulnar nerves separately (Fig. 1). For the median nerve (R1), the active recording electrode (A) was placed around the proximal interphalangeal joint of digit II; the reference electrode (R) was placed over the distal interphalangeal joint of digit II. For the ulnar nerve (R2), the active recording electrode (A) was placed around the proximal interphalangeal joint of digit V; the reference electrode (R) was placed over the distal interphalangeal joint of digit V. The proximal interphalangeal joint rather than the base of the finger was selected as the site for the proximal ring electrode to reduce or eliminate the volume-conducted motor response. Adjacent digits were kept from inadvertently contacting each other. Ground (G) electrode was placed on the dorsum of the hand, between the recording and the stimulating electrodes (the figure shows the ground electrode placed on the palm).

Stimulations were applied over the median and ulnar nerves at the wrist. Median and ulnar nerves were stimulated separately (S1 for median nerve, S2 for ulnar nerve), both at 14 cm from the active recording electrode (A) placed around the proximal interphalangeal joint of digit II (R1) and digit V (R2). The cathodes were placed proximally. The author used only supramaximal stimulation with a duration of 0.1–0.4 ms, using the same sites of motor stimulations at the wrist (both median and ulnar nerves, S1 and S2, respectively) to compare median and ulnar sensory latencies in the same and opposite hands (see sensory conduction studies). Felsenthal chose the same stimulation points in an attempt to achieve some standardization.

Onset latency (ms) was measured from the stimulus onset to the initial deflection of the sensory nerve action potential (SNAP). Peak latency (ms) was measured from the stimulus onset to the peak of the negative deflection of the SNAP. All latencies were determined at the same session and at room temperature. Skin and room temperature data was not given. All latencies of the median and ulnar nerves were compared in all subjects in the same hand and in the opposite hand. Normal values (Table 1) were obtained from 100 normal hands of 50 healthy volunteers.