The mismatch negativity (MMN) recordings provide an objective measure of the preattentive, automatic auditory discrimination function. This article deals with issues central to the recording of the MMN and the interpretation of the results for clinical and electrodiagnostic purposes. The methods of acquiring as pure an MMN response as possible, i.e., one not contaminated by auditory cortical responses reflecting other functions, are discussed first. Second, other technical questions associated with the recording are reported on, e.g., what MMN parameters should be recorded and how, what is the smallest recordable MMN response, and what is the repeatability of the MMN recordings. Then, the effect of various physiological factors on the MMN (age, alertness, gender, topographic distribution of the MMN) is considered. The correlation between auditory discrimination performance and the MMN amplitude, observed in normal population, is dealt with. Finally, there is a short concluding overview on clinical findings of MMN recordings and discussion on their electrodiagnostic applications.