Purchase this article with an account.
Erich E. Sutter; Visually modulated endogenous activity: A component of active visual processing. Journal of Vision 2005;5(8):369. doi: 10.1167/5.8.369.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: In the past electrophysiological investigations of visual processing in the CNS have almost exclusively relied on stimulus-locked averaging. Often all activity that showed no first or higher order correlation with the stimulus was considered additive noise. Clearly, this view is much too limiting. As endogenous activity of the CNS must interact with visual input in the highly nonlinear neural substrate, the resulting neural responses must carry visual information that is no longer stimulus locked and lost in conventional signal processing. We hypothesize that it represents the ‘active’ component in visual processing. The purpose of this project is to demonstrate and characterize such visually modulated endogenous activity (VMEA) in the human VEP.
Methods: In this first study the stimulus consisted of slow, pseudorandom pattern presentations using different base intervals up to 500 ms. In the first step, we subtracted all response component (kernels) that correlated directly with changes in features of the stimulus. The noise-like residue, presumably Demodulation requires application of a nonlinear operation to the response signal before it is correlated, again with the visual stimulation. We are using the formation of auto-products of the EEG signal immediately preceding the stimulus with various lags as well as rectification for demodulion.
Results: At slow stimulation rates when the stimulus transitions are individually perceived, the VMEA dominates. At fast stimulation rates, when the stimulus appears noise-like, the directly correlated response components dominate and the VMEA becomes negligible in size. The main VMEA components appear at an implicit time of around 300ms and another at 375 to 450 ms.
Conclusions: The response residue, after subtraction of all stimulus-locked components, indeed carries stimulus related information. These VMEA components are well defined and can be derived with excellent signal-to-noise ratios suggesting a large information content.
This PDF is available to Subscribers Only