Abstract
An optimum filtering method devised by D. O. Walter on the basis of the Wiener filter theory is implemented by means of a general purpose digital computer and is applied to estimation of visual potentials evoked by a single photic stimulus over the scalp.
Two essential points are stressed; one is that the Wiener filter me thod is more effective for minimizing noise and artifacts than any other method (comparison with the conventional average or arithmetic mean is made), and the other is the fact that presentations of photic stimuli should be aperiodic rather than periodic so as to reduce interference between evoked potentials and background processes such as the alpha activity.
The Wiener-filtered result and the averaged one are compared. The former is rather smooth and simple in waveform compared to the latter. A typical visual evoked potential recorded from a normal male by Wiener filtering exhibits a bi-phasic structure, It is almost flat during the first period (0-40 msec), and then shows a pattern characteristic of photic stimulus only in the range of 40-300 msec after the stimulus. Two predominant peaks, one positive in the range of 70-90 msec with amplitude of 2-10μv, the other negative in the range of 140-180 msec with amplitude of 4-14μv, are the most essential responses to a photic stimulus except for minor quantitative differences among individual subjects and experiments.
In the range beginning from 300 msec after the stimulus, the waveform becomes almost flat and shows no tendency to verify existence of the so-called rhythmic after-discharge The appearance of the rhythmic after-discharge Is no more than a result due to the data processing technique. The reasoning for the above fact is fully discussed in the present paper.
