In this prospective study, automated interference pattern analysis (IPA, ''Willison analysis'', modified by Stalberg et al. 1983) was compared to the quantitative evaluation of mean motor unit potential duration (QMUP) in 239 muscles from consecutive, unselected patients. The sensitivity and specificity of both methods were calculated with respect to the clinically derived final neurological diagnosis, with histology available for 120 examinations. Whereas specificities were not different for the methods, the sensitivity for detection of abnormal vs. normal was 49% for QMUP and 74% for IPA (P < 0.001). The sensitivity for detection of myopathy or neuropathy was 46% or 38% for QMUP and 75% (P < 0.001) or 53% (P < 0.05) for IPA. Thus, in all instances, IPA had superior sensitivity with unchanged specificity as compared to QMUP. The results of a rapid and purely qualitative visual MUP assessment were statistically not different from QMUP. Although widely used, neither of these methods has been evaluated for its reliability in unselected patients with various grades of disease. Our results indicate that in a routine setting, the best diagnostic strategy might be the automated IPA, which can be quickly obtained in several muscles, followed by muscle biopsy in unclear cases.