Principle components analysis of pain thresholds to thermal, electrical, and mechanical stimuli suggests a predominant common source of variance

Abstract 

We addressed the question whether pain thresholds to different stimuli measure independent aspects of pain or one common phenomenon. In the first case, different stimuli are required to completely characterize a subject’s pain sensitivity. In the second case, different stimuli are redundant and can be used to calculate composite scores across pain modalities. We measured pain thresholds to several stimuli (heat, heat/capsaicin, cold, cold/menthol, blunt pressure, 5-Hz sine-wave electric current (0–20mA), punctate pressure (von Frey hairs), and von Frey hairs plus capsaicin application) in 45 healthy men and 32 healthy women (aged 20–44 years). We observed that pain thresholds were significantly correlated with each other. Principal component analysis indicated that their variance was attributable more to the difference in subjects (variance estimate: 0.393) than to the difference in pain stimuli within a subject (variance estimate: −0.008). Among three principal components of the intercorrelation matrix with eigenvalues >1, the first, explaining 48% of the total variance, carried high loadings from all stimuli indicating that they shared a common source of half of their variance. Only minor variance components, each explaining <14% of the total variance, indicated a distinction of pain stimuli. There, a pattern of similarities and dissimilarities emerged agreeing with known distinct mechanisms of nociceptive responses to different stimuli. We conclude that characterizing a person as being generally stoical or complaining to any painful stimulus appears to be justified at least at pain threshold level.

Keywords: Pain threshold, Pain asessment, Experimental pain models