Computer-assisted thermal threshold testing is a psychophysical semi-quantitative method of testing the function of small nerve fibres. At present, there is not enough data available on the effect of physiological variables on the threshold value and on the intra-individual variability of the test. Methods: Thermal threshold for heat and cold was determined in a group of 50 healthy volunteers (25 males and 25 females) in two localizations (thenar of the left upper extremity and dorsum of the right lower extremity). Each individual was examined with three different algorithms: two reaction time methods (reaction time inclusive – non-randomised and randomised variant of the Limits method) and one constant stimulus method (reaction time exclusive – randomised variant of the Levels method). To determine intra-individual variability, all tests were repeated in 30 individuals within one week and the upper normal limit of physiological intra-individual change of each test was determined. Results: Values of thermal threshold for cold decreased significantly whereas values for heat increased in examination of lower extremities, in men, and when using reaction time. The effect of age on thermal threshold was not significant. Intra-individual variability of threshold values expressed as the median of coefficients of intra-individual variability fell between 15 and 30% and was independent of test type, examined location, age or gender. Conclusions: Normal limits have to be determined with respect to gender, algorithm type and localization of the thermal element. Intra-individual variability of threshold values is acceptable, considering the psychophysical character of the test, is comparable to other quantitative tests of sensory perception and is not significantly influenced by test type or by physiological variables.

        Key words: thermal threshold testing, small nerve fibres, reaction time methods, constant stimulus methods