Thermal discrimination thresholds: a comparison of different methods.

Thermal testing was carried out on 55 healthy subjects in order to establish normal results and reproducibility of warm and cold thresholds. Diurnal variations of thresholds were investigated in a further 30 normal subjects. Then the sensitivity of different testing procedures was investigated in 33 patients with diabetes mellitus, but without severe polyneuropathy. Forced choice testing takes 6 times longer than the method of limits, and the results are not considerably different. It is thought that the forced choice algorithm does not provide a method for clinical routine. Another new approach, the double random staircase method, may help to exclude bias without taking too much time.