Nerve fibre morphometry: a comparison of techniques.

This study compared different approaches to measuring nerve axon and fibre diameters and areas from transverse sections. A mock photomicrograph and mock tissue section, each with 100 identical, circular 'fibres' was constructed. Three measurement protocols were investigated: (A) circular approximation from minimum diameter; (B) circular approximation from the mean of orthogonal diameters; and (C) calculation of diameter and area from a digitized circumference. For each protocol, all 100 fibres on the photomicrograph were repeatedly measured using a digitizing tablet. Similarly, the fibres on the mock tissue section were measured using a digitizing tablet and microscope with camera lucida. The variance for these data was calculated. Protocols were compared on the basis of variability and the amount of digitizing time required. For diameter measurements, protocol B showed significantly lower variability than A or C (P less than 0.05), with only a modest increase in digitizing time over A. For area measurements, protocols B and C showed significantly lower variability than A (P less than 0.05), again with a modest increase in digitizing time. Measurements made using the microscope and camera lucida showed significantly lower variability than those made from the photomicrograph, but took more time. These data suggest that for diameter measurements, a mean of orthogonal diameters approach is best, and that for area measurements, a traced circumference approach is best as it is more flexible than the orthogonal diameter approach. While the microscope and camera lucida setup is more time-consuming to use, it eliminates the need for photomicrograph production.