Dynamic trajectory analysis to identify conditioned pain modulation profiles in pain-free volunteers: Capturing what is lost when focusing on mean effects

Pain is an unpleasant experience regulated by endogenous mechanisms. Conditioned pain modulation is a particularly well-studied inhibitory mechanism in which continuous painful stimuli (2 min) are used. It is usual practice to report the average between participants and between multiple measurement times. However, this practice limits our ability to highlight inter-individual variability and the presence of subgroups sharing similar responses. Nor does it allow us to appreciate the dynamic nature of this phenomenon. A cohort of 205 pain-free participants completed a CPM paradigm involving a 2-minute continuous heat pain stimulation (test stimulus) applied before and after a cold pressor test (conditioning stimulus). An extended mixed model was used to identify dynamic response trajectories for the test stimulus (PRE-Trajectories) and for the CPM response (DELTA-Trajectories). Four PRE-Trajectories were determined with three of them differing significantly from the mean-based approach. For the four DELTA-Trajectories identified, none were similar to that generated by the average. Moreover, an association was found between PRE and DELTA-Trajectories, namely that individuals with higher pain sensitivity patterns exhibited the highest inhibition patterns, and conversely, those with lower sensitivity exhibited lower inhibition. These results highlight the ability of extended mixed models to capture individual patterns of response to pain perception and modulation, most of whom are hidden sub-profiles missed by mean-based approaches. This analysis offers a promising tool for quantitative sensory testing that respects the dynamic nature of these mechanisms. Trajectory modeling opens future avenues for linking individual response profiles to chronic pain or pain outcomes. PERSPECTIVE: This study introduces a dynamic trajectory analysis for conditioned pain modulation, revealing profiles undetectable with traditional mean-based methods. This approach clarifies the dynamic nature of pain inhibition, improves characterization of individual responses, and offers substantial clinical value by linking these distinct profiles to relevant pain outcomes and potential treatment implications.