Development of a classification rule for four clinical therapeutic psychotropic drug classes with EEG power-spectrum variables of human volunteers.

Abstract

An objective rule for the classification of psychotropic substances has been developed. Classification is based on data from five basic studies simultaneously designed and performed and involving 75 healthy volunteers who ingested 20 different psychotropic drugs and 5 placebos in single oral dosages. Each volunteer took one psychostimulant, one antidepressant, one neuroleptic, one minor tranquilizer and one placebo in a double-blind Latin square cross-over design. The variables were 6 frequency bands, based on power spectrum estimates and determined by factor analysis, plus total power in the 1.5-30.0 Hz range. An objective classification rule was established by multi-group (5 groups) linear discriminant analysis. Reclassification of the substances by the new rule yielded correct results for 17 out of 20 psychotropic drugs and 4 out of 5 placebos. Of placebos from various studies not used for the establishment of the classification rule, 7/9 were classified correctly. The validity of the rule for other classes of substances will have to be verified in independent studies.