This statistical point is true. However, there is a practical problem with experimental neuroscience - data is painful to get and the means are often comparable

This statistical point is true. However, there is a practical problem with experimental neuroscience - data is painful to get and the means are often comparable to the variances. So, with the data you have, it is often much harder to show A > B than to show A > C and B !> C.