It is now possible to measure the effect of drugs on CNS function in elderly patients by a number of methods. All are safe, non-invasive and cause minimal discomfort to volunteers. Each has disadvantages and no reliance should be made on one. Computerised Axial Tomography (CAT) scanning is insufficiently sensitive to be used in research but has a place in the selection of subjects. Cerebral blood flow studies with xenon-133 can also be used in this manner as well as to study the effects of drugs on cerebral activity. To date, this use has received insufficient attention. The principle investigative techniques are the Electroencephalogram (EEG) and psychomotor tests. The former is reliable, objective and can measure onset, duration of drug action, and the minimum effective dose. Psychomotor tests are given as a battery which should include tests of performance, memory and learning, visual analogue scales and physical tests such as the measurement of sway. Subject selection is perhaps the most important aspect, since it determines data interpretation. Experiments in drug therapy must be carefully designed and tests selected to answer specific questions before the first subject is enrolled. Bias must be reduced as far as possible by making the experiment double-blind with the code not broken until the results have been analysed. A placebo or a proven reference drug should be included.