Towards obtaining unbiased estimates of the total number of synapses in a brain region: problems of primary and secondary importance