A SEQUENTIAL DESIGN FOR PSYCHOPHYSICAL EXPERIMENTS - AN APPLICATION TO ESTIMATING TIMING OF SENSORY EVENTS

An experimental subject sequentially receives different levels of a st imulus, and data are recorded on response or non-response to the stimu lus. To ensure that the subject cannot predict the next stimulus level based on previous stimulus levels, a randomized design, based on a ge neralized Polya urn model, is used to allocate the stimulus levels. Th e goal of the experiment is to elicit information efficiently about th e relationship between stimulus level and response (either for an indi vidual subject or a group of independent subjects), by estimating quan tiles of the stimulus-response curve. Our design allocates stimulus le vels unimodally and symmetrically around the unknown median of the sti mulus-response curve. We discuss estimation under a broad family of di stributions and also fully discuss design issues and options. This des ign was used for an experiment in neurophysiology in humans to estimat e the timing of onset of kinesthetic stimuli, Such psychophysical stud ies can increase our understanding of normal and pathological function . We present data from that experiment. (C) 1997 by John Wiley & Sons, Ltd.