Abstract: Auditory tone burst signals are supplied to the ear of a human subject and the electroencephalographic responses are monitored, filtered, and provided to a sampler which samples the polarity of the EEG signal at predetermined times after the auditory signal has been applied. The existence of an expected response waveform of a definite shape and frequency in an individual who responds to the sounds causes a statistical shift of the ratio of sample polarities matching the expected polarity with respect to the total number of trials. The number of matched response polarities and the number of trials are counted and can be compared after the test is completed. A decision can be made as to whether a subject is responding by selecting a statistical confidence limit which allows the hypothesis that a subject is not responding to the auditory signal to be rejected if the ratio of responses to trials falls above the selected confidence limit.