Despite progress in guideline-driven cancer pain management, crucial challenges persist. Automatic Pain Assessment (APA) investigates behavioral aspects and biosignals, such as electrodermal activity (EDA) and heart rate (HR) variability (HRV), for providing objective and context-aware pain evaluation and monitoring. In this prospective, single-center study on cancer pain, we recorded EDA and HRV during a 3-block cognitive Stroop task and stratified patients by 0-10 Numeric Rating Scale (NRS) pain (<6 vs ≥6).
Since chronic pain can induce cognitive and attentional alterations, the aim is to assess the performance of the autonomic nervous system during a cognitive stress test, investigating the links between nociception (the encoding of noxious stimuli through autonomic or behavioral responses) and the neurobiological consequences of pain. We extracted tonic and phasic electrodermal activity (EDA) features, including skin conductance level (SCL), skin conductance response (SCR), recovery times, and time/frequency-domain HR/HRV indices, and then we compared groups across Stroop phases. Patients with lower pain intensity showed consistently higher SCL and stronger phasic SCR components across test phases (SCL mean p ≤0.05; SCR mean and SCR integral p≤0.01), whereas rise/recovery times did not differ; peak counts normalized by phase duration diverged modestly (significant in Stroop phases 1 and 3). HRV differences were limited, with a notable increase in HF power during the incongruent (phase III) block only. These exploratory findings provide initial evidence that cognitive stress–evoked autonomic responses may contribute to distinguishing pain phenotypes in cancer patients. Pipeline and results can be used for developing a multimodal APA framework aimed at personalizing cancer pain treatment.