Abstract Background: Innocuous cooling of the skin activates cold-specific Aδ fibres, and hence, the recording of cold-evoked potentials (CEPs) may improve the objective assessment of human thermo-nociceptive function. While the feasibility of CEP recordings in healthy humans has been reported, their reliability and diagnostic use in clinical conditions have not been documented. Methods: Here, we report the results of CEP recordings in 60 consecutive pa- tients with suspected neuropathic pain, compared with laser-evoked potentials (LEPs) which are the gold standard for thermo-algesic instrumental assessment. Results: CEP recording was a well-tolerated procedure, with only ~15 min of sur- plus in exam duration. The reproducibility and signal-to-noise ratio of CEPs were lower than those of LEPs, in particular for distal lower limbs (LLs). While laser responses were interpretable in all patients, CEPs interpretation was inconclu- sive in 5/60 because of artefacts or lack of response on the unaffected side. Both techniques yielded concordant results in 73% of the patients. In 12 patients, CEPs yielded abnormal values while LEPs remained within normal limits; 3 of these patients had clinical symptoms limited to cold sensations, including cold-heat transformation. Conclusions: CEPs appear as a useful technique for exploring pain/temperature systems. Advantages are low cost of equipment and innocuity. Disadvantages are low signal-to-noise ratio for LL stimulation, and sensitivity to fatigue/habitu- ation. Joint recording of CEPs and LEPs can increase the sensitivity of neuro- physiological techniques to thin fibre- spinothalamic lesions, in particular, when abnormalities of cold perception predominate. Significance: Recording of cold-evoked potentials is a well-tolerated, inexpen- sive and easy-to-use procedure that can be helpful in the diagnosis of abnormali- ties in the thin fibre- spinothalamic pathways. Supplementing LEPs with CEPs allows consolidating the diagnosis and, for some patients suffering from symp- toms limited only to cold, CEPs but not LEPs may allow the diagnosis of thin fibre pathology. Optimal CEP recording conditions are important to overcome the low signal-to-noise ratio and habituation phenomena, which are less favourable than with LEPs.