Do response times add to self-reported craving? A secondary analysis of a neuromodulation trial.

Craving is a core symptom and treatment target in tobacco use disorder. However, self-reported craving ratings often show weak or inconsistent associations with clinical outcomes, possibly because they overlook automatic motivational processes central to cue-reactivity. Response time to craving-item ratings has been proposed as an implicit marker of such processes, but its sensitivity to intervention and clinical relevance remain untested. In this exploratory secondary analysis of a randomized trial, tobacco-dependent adults completed an image-based cue-reactivity paradigm immediately before and after a single 1 Hz repetitive transcranial magnetic stimulation (rTMS) session targeting either the left frontopolar cortex (lFPC; n=12) or the vertex (Cz; n=11). We examined whether response time, modeled both as a mean-level measure and as a function of self-reported craving, was modulated by stimulation site and whether these indices were associated with smoking outcomes at baseline and one-week follow-up. Mean response time showed a significant Group × Time × Stimulus effect (χ²(1) = 8.64, p = 0.003): lFPC stimulation attenuated the typical post-exposure speeding observed in the vertex group, selectively for smoking cues. Models incorporating self-reported craving yielded convergent smoking-specific effects (Wald χ²(1) = 4.47, p = 0.034), but the curvature patterns were inconsistent and statistically fragile, limiting interpretability. Exploratory clinical analyses did not reveal reliable associations with smoking-related outcomes, although repeated nominal associations with tonic craving under smoking cues suggested a possible relationship. These findings provide initial evidence that response time may capture implicit components of cue-reactivity and function as a modulable behavioral endpoint sensitive to lFPC neuromodulation.