The need for a familiarization session and for a placebo session for each subject in pharmaco-EEG studies was examined. The data from 18 crossover studies of 22 substances at 33 doses in 112 subjects was used: 371 drug sessions, 195 placebo sessions and 53 familiarization sessions. The left occipital to vertex signal was quantified using power spectral density analyses. EEG changes under different conditions were compared using the spectral difference index. Less EEG change from baseline was found after placebo in the first session than after placebo in later sessions. Drug effects were defined by comparisons to placebo effects in the same subjects (related samples analyses) and to placebo effects in other subjects in the data base (independent samples analyses). Decisions using the two analyses were in accord for 27 substance doses: 20 were found active, 7 inactive. For one substance, the related samples analysis indicated drug activity, while the independent samples analysis did not (just below criterion). For five other substances, the related samples analyses did not distinguish drug and placebo sessions, while the independent samples analyses showed lesser EEG changes after drug than after placebo. First-session findings support the need for a familiarization session in pharmaco-EEG studies. Placebo findings indicate that past placebo session data are adequate for deciding whether a drug has CNS activity, without the necessity of a placebo session for each subject.