Un nuovo metodo di filtraggio digitale per l'analisi dei potenziali evocati del tronco encefalico

Analysis and interpretation of auditory evoked brainstem potentials, as well as of any other kind of evoked potential, rely on two different steps: a) identification of each individual response wave or "wave labeling"; b) latency measurements. This two steps are closely linked since the identification of a given wave is basically the result of some previous initial proximation measurement of its latency. The major source of uncertainty is mainly due to the residual presence of background noise (low- and high-frequency noise). The aim of the present paper is to propose a new data analysis method that has proven o greatly increase the reliability of the entire procedure. The method is based on the use of a linear, non casual, zero-phase digital filter that can easily be implemented, even onto small computers. Its use is fast an simple and has not required any prior knowledge of the temporal structure of the signal being analyzed. This means that it can be successfully applied in the analysis of traces with large waveform alterations. Method performance is shown on the responses measured from a population of subjects suffering from multiple sclerosis, a pathology that often produces responses with large alterations.