On the accuracy of memory for pain

Abstract

T he principal point made by Salovey and his associates 13 is that apart from an assimilative effect of present pain intensity on memory for the severity of prior pain, retrospective self-reports of pain appear to be accurate and reliable. Given the long-standing and widely shared opinion that such reports are prone to distortion, this is a point worth making. Still, the fact is that significant distortions have been detected in several studies involving memory for either acute or chronic pain. 3,7,8 Though it is conceivable that the results of these studies paint an illusory picture of inaccuracy and that, as Salovey et al. suggest, looks like recall bias may in fact be error variance, I think that it is far too soon to tell, for the simple reason that memory for pain represents a young and largely unexplored area of research that is long on interesting conjectures but short on compelling data. I think that the key question we should ask is not whether memory for pain is veridical, but rather, what factors affect the accuracy and other attributes of memory for pain? One such factor is the duration of the retention interval. 3,5,9 It seems reasonable to suppose that memory for acute pain, like memory for most other kinds of discrete episodes or events, becomes poorer with the passage of time. How much poorer is an open issue, as is whether memory for a particular episode of pain decays at the same rate as does memory for other kinds of episodic information (such as the emotion one experienced at some specific moment in the past14), and whether different forms of acute pain follow different forgetting functions. Though issues of a similar sort could be profitably pursued in connection with chronic pain, they may prove more difficult to answer in a clear-cut manner. Part of the problem here is that chronic pain is just