Computerized CFF: Reliability and Validity of Two Psychophysical Techniques

Abstract

Two computerized CFF measures have been implemented in our laboratory and used in several studies. One method is based on the traditional "method of limits" threshold determination technique (CFF-ML). Process-controlled optimization of individual stimulus parameters (starting points and center frequency of ascending and descending sweeps) were shown to give a significant increase in test-retest reliability as compared with a non-computerized version. In subsequent studies, test-retest reliabilities (> 24 h) for CFF-ML have regularly exceeded .90. Criticism against the method of limits for not separating sensory and response factors was the basis for the design of a forced choice, three-alternative, process-controlled CFF method (CFF-3). The reliability of CFF-3 appears to be high (.90-.95), there are virtually no order effects (even after 20 retest sessions of four minutes each), and the "exposure time" for a reliable threshold determination can be made short enough to study fast drug onset effects, for instance after smoking. Both CFF methods appear to have a satisfactory validity, allowing demonstration of significant effects in small samples of subjects, of e.g. variations in cerebral blood flow, effects of thyroxine substitution therapy in hypothyroid patients, and effects of sedative and CNS stimulant drugs.