Where is the noise in SDT pain assessment?

Many applications of sensory decision theory (SDT) to pain research have used discrimination as a measure of pain or sensory sensitivity. This belief is based on the classical SDT assumption that discrimination and criterion represent separation of sensory and decision processes. This assumed separation stems from a model where all noise or variability is part of the sensory transduction mechanism. We present an alternative formulation that allows for decision variability as well as variability in sensory transduction. This formulation documented by computer simulation shows that decision variability and sensory variability are indistinguishable and that any measure of discriminability is degraded by both. Thus discriminability is influenced by both sensory and non-sensory factors. There is no way of knowing if a drug-induced change in discriminability represents an analgesic effect or a change of the observer's ability to make consistent judgments.