2-D spatial frequency dependence of VEPs: A neural network analysis

An approach to black-box identification-function problems is introduced. Visual evoked potentials (VEPs) were recorded from normal subjects using stimulus patterns that evolved from a totally random distribution of intensities to an ordered distribution representing a bar. The contribution of all spatial frequencies to the amplitudes of the evoked potentials was studied by performing a 2-D FFT on the stimulus images. The FFT spectra were used as inputs to a neural network as a black box approach in order to study the VEP spatial frequency tuning curves of the subjects.<>