Premium
Methods for planning experimental designs that allow measurement of several concurrent drug effects
Author(s) -
Overton Donald A.,
Shen C. Frank
Publication year - 1989
Publication title -
drug development research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.582
H-Index - 60
eISSN - 1098-2299
pISSN - 0272-4391
DOI - 10.1002/ddr.430160215
Subject(s) - design of experiments , computer science , linear regression , drug , regression , variety (cybernetics) , regression analysis , reliability engineering , machine learning , statistics , artificial intelligence , mathematics , pharmacology , engineering , medicine
Psychoactive drugs concurrently produce a variety of effects on the formation and retrieval of memories and on performance of learned responses. Experimental designs used in psychopharmacology are usually arranged so that some drug effects do not occur or do not influence the dependent variables of interest and so that the strength and statistical significance of the remaining effects can be concurrently determined. Moderately complex designs are often required to allow the simultaneous evaluation of the several drug effects expected to occur in a preparation. In general, the inclusion of more experimental conditions in a design will allow concurrent measurement of more drug effects, but it is often less than obvious exactly which effects can be measured by any given design. Multiple linear regression provides a method for determining which effects can be measured with certain types of designs and hence provides a method for selecting designs adequate to deal with the drug effects anticipated in any particular experiment. This paper presents illustrative examples of the use of multiple linear regression to evaluate the utility of selected experimental designs. It also describes some new experimental designs more adequate than those previously employed for the study of state‐dependent learning and concurrently occurring drug effects.