The manipulation of living biological cells in microfluidic channels by a combination of negative dielectrophoretic barriers and pressure-driven flows is widely employed in lab-on-a-chip systems. However, electric fields in conducting media induce Joule heating. This study investigates if the local temperatures reached under typical experimental conditions in miniaturized systems cause a potential risk for hyperthermic stress or cell damage. Two methods of optical in situ temperature detection have been tested and compared: (i) the exposure of the thermo-dependent fluorescent dye Rhodamine B to heat sources situated in microfluidic channels, and (ii) the use of thermoprecipitating N-alkyl-substituted acrylamide polymers as temperature threshold probes. Two-dimensional images of temperature distributions in the vicinity of active negative dielectrophoresis (nDEP)-barriers have been obtained and local temperature variations of more than 20 degrees C have been observed at the electrode edges. Heat propagation via both buffer and channel walls lead to significant temperature increases within a perimeter of 100 mu m and more. These data indicate that power dissipation has to be taken into account when experiments at physiological temperatures are planned.