Statistical comparison of axon-scaled neurochemical production

Treatments designed to increase neurochemical levels may also result in increases in the numbers of axons that produce the neurochemicals of interest. A natural research question is how does one compare the average neurochemical production per axon between two (or more) experimental groups. Two statistical methods are proposed for this problem. The first method utilizes a delta-method approximation to the variance of a function of random variables while the second method is based on the bootstrap. These methods are illustrated with data obtained from perivascular norepinephrine following intracerebroventricular infusion of neurotrophin nerve growth factor in adult rats and are studied in a small simulation experiment. The delta-method confidence intervals exhibited better coverage properties than the bootstrap alternative.