AIMS: Non-nociceptive somatosensory input, such as tactile or proprioceptive information, always precedes nociceptive input during a painful event. This relationship provides clear opportunities for predictive associative learning, which may shape future painful experiences. In this differential classical conditioning study we tested whether pain-associated tactile cues (conditioned stimuli; CS) could alter the perceived intensity of painful stimulation, and whether this depends on duration of the CS-seeing that CS duration might allow or prevent conscious expectation.
METHODS: Subjects underwent a classical differential conditioning task in which a tactile cue at location A (CS+) preceded painful electrical stimulation at location B (UShigh), whereas a tactile cue at location C (CS-) preceded non-painful electrical stimulation at location B (USlow). At test, we compared the pain evoked by a moderately painful stimulus (USmed) when preceded by either the CS+ or CS-. CS duration was manipulated between subjects. Participants were assigned to one of three groups: Long CS (4s, allowing conscious expectation), Short CS (110ms) and CS-US indistinguishable (20ms), preventing conscious expectation). We hypothesised that more pain would be evoked by the US when preceded by the CS+ relative to the CS-, and that the effect would be independent of CS duration.
RESULTS: Fifty-four healthy participants (31 females, age=26, SD=9) were included in the analysis. The hypotheses were supported in that more intense pain was evoked by the USmed when paired with the tactile CS+, than when paired with the tactile CS-; mean difference 3mm on a 150mm VAS (CI 0.4-4.8mm). CS duration did not moderate the effect. The effect was greater in those participants where calibration was optimal, as indicated by a relatively more painful UShigh.
CONCLUSION: We conclude that pain-associated tactile cues can influence pain, and that this effect is not dependent on stimulus duration. This suggests that explicit expectation is not a requirement for predictive cues to modulate pain. That the presence of the CS+ resulted in only a 5.3% higher intensity rating compared with the CS- may reflect a limitation of laboratory studies, where a limited number of trials, an artificial context and the use of experimental pain are likely to reveal only glimpses of what is clinically possible.
IMPLICATIONS: Pain-associated visual and auditory cues have been shown to enhance pain in laboratory and clinical scenarios, supposedly by influencing expectation of impending harm. We show that pain-associated somatosensory cues can also modulate pain and that this can occur independently of expectation. This points to a larger potential role for associative learning in the development and treatment of pain than has previously been considered. We suggest that research into associative mechanisms underpinning pain, as distinct from those that link pain to pain-related fear and avoidance, is worthwhile.