Treatments designed to increase neurochemical levels may also result in inc
reases in the numbers of axons that produce the neurochemicals of interest.
A natural research question is how does one compare the average neurochemi
cal production per axon between two (or more) experimental groups. Two stat
istical methods are proposed for this problem. The first method utilizes a
delta-method approximation to the variance of a function of random variable
s while the second method is based on the bootstrap. These methods are illu
strated with data obtained from perivascular norepinephrine following intra
cerebroventricular infusion of neurotrophin nerve growth factor in adult ra
ts and are studied in a small simulation experiment. The delta-method confi
dence intervals exhibited better coverage properties than the bootstrap alt
ernative.