The use of ultrasound-guided electrode positioning in near-nerve myography was investigated. This is a minimally invasive technique that allows repeated measurements to increase accuracy and hence decreases animal numbers. Ultrasound imaging of the sciatic nerve was performed in nine rats using a 55 MHz high-end transducer. Once visualised, a monopolar needle electrode was placed through the skin near this nerve. Upon stimulation, two surface electrodes, placed over the gastrocnemius muscle, recorded compound muscle action potentials (CMAPs). Reproducibility was tested having two teams of investigators perform the recordings consecutively. Reliability of the procedure was determined by comparing the ultrasound method to the conventional technique, which requires an incision through muscle and skin to expose the sciatic nerve. In all animals the sciatic nerve was visible on ultrasound images. Both methods showed CMAP latencies (duration was determined as the time interval between the onset latency and positive peak). The conventional method had a mean latency of 3.4 +/- 0.5 ms, our method had a mean latency of 3.3 +/- 0.5 ms. Reproducibility was excellent (observed latencies and amplitudes: 3.3 versus 3.3 ms and 25.6 +/- 5.1 mV versus 22.5 +/- 8.8 mV) resulting in a coefficient of variation for duration of 2.1% and for amplitude 6.7%. Interclass correlation coefficient was 0.828 for duration. Comparing the three different measurements no significant differences were found and our new method can therefore be considered reliable and comparable to the conventional method. Ultrasound-guided near-nerve needle positioning is a reproducible and reliable minimally invasive method for selectively eliciting CMAPs, which allows repeated CMAP measurements for studying nerve regeneration in rats. (C) 2010 Elsevier B.V. All rights reserved.