[Evoked potentials--effect of filter type on the results].

BACKGROUND Occupational hearing loss and its diagnostic methods are one of the subjects of great interest in the field of occupational medicine and hygiene. The primary aim of the work was to determine the characteristics of ready-made filters used to process the brainstem auditory evoked potentials. No data on the properties of these filters have prompted the author to verify the algorithm of their operations with particular reference to their effects on the results and their interpretation. MATERIAL AND METHODS A Nicolet model Spirit 2000 Unit and a phantom of the author's own design were used in the experiment. The phantom simulated the patient's skin resistance and brainstem responses to the presented auditory stimulus. The filter characteristics was verified in two stages. First, the signals generated by electromagnetic induction between the measuring leads and earphone coils were employed. Second, a precise source of the composite signal simulating the brainstem response was used. The study protocol comprised a series of tests using ready-made filters: a neuronal, a linear, and an arithmetical averaging filter. RESULTS The arithmetical averaging resulted in the most real representation of the reference signal, the use of the linear filter resulted in the phase shift of the high-amplitude reference signal, while the use of the neuronal filter resulted in the phase shift of the low-amplitude reference signal. CONCLUSION The use of filters may lead to the distortion of the input signal. When analysing the results of evoked potential determinations with use of filters, the type of the filter and its effects on the input signal must be taken into account.