Top-down effects on early visual processing in humans: A predictive coding framework

Abstract

a b s t r a c tAn increasing number of human electroencephalography (EEG) studies examining the earliest component of the visual evoked potential, the so-called C1, have cast doubts on the previously prevalent notion that this component is impermeable to top-down effects. This article reviews the original studies that (i) described the C1, (ii) linked it to primary visual cortex (V1) activity, and (iii) suggested that its electrophysiological characteristics are exclusively determined by low-level stimulus attributes, particularly the spatial position of the stimulus within the visual field. We then describe conflicting evidence from animal studies and human neuroimaging experiments and provide an overview of recent EEG and magnetoencephalography (MEG) work showing that initial V1 activity in humans may be strongly modulated by higher-level cognitive factors. Finally, we formulate a theoretical framework for understanding top-down effects on early visual processing in terms of predictive coding.

State-dependent attention modulation of human primary visual cortex: A high density ERP study

“…Increasing perceptual load at fixation produces a reduction of the perceptual encoding, taking place in V1, of a peripheral task-irrelevant distractor stimulus, in agreement with previous ERP and imaging results (Rauss et al, 2009;Schwartz et al, 2005), as well as attention selection models (Lavie, 2005;Rauss et al, 2011a). However, this amplitude modulation of the C1 component with perceptual load was no longer evident when an increase in state anxiety or negative affect was transiently induced.…”

“…These effects therefore translate flexible filtering mechanisms taking place in V1 during early sensory processing. Tables and Table captions Low with the electrophysiological hallmark of this early retinotopic visual ERP component (Clark et al, 1995;Jeffreys and Axford, 1972;Rauss et al, 2011a and2011b). Following the C1, a P1m was recorded for the stimuli presented in the upper visual field, while a more lateralized P1 was recorded in response to stimuli presented in the lower visual field.…”

“…The C1 has for a long time been described as being resistant to modulatory effects exerted by distant fronto-parietal attention control regions onto lower tier visual cortex (Clark and Hillyard, 1996;Handy et al, 2001;Martinez et al, 1999). More recently, systematic C1 amplitude changes have been reported with top-down cognitive manipulations (see Rauss et al, 2011a for a recent review), including perceptual learning and expertise (Bao et al, 2010;Jin et al, 2010;Pourtois et al, 2008), emotional valence (Eldar et al, 2010;Halgren et al, 2000;Pourtois et al, 2004;West et al, 2011), and featurebased or spatial attention (Karns and Knight, 2009;Kelly et al, 2008;Proverbio et al, 2010;Zani and Proverbio, 2009). With regards to attention, not only increases of the C1 to attended visual stimuli were shown, but also substantial reductions of this same early visual component to unattended or task-irrelevant stimuli were evidenced, suggesting flexible and adaptive gain control mechanisms exerted by putative fronto-parietal networks onto lower tier visual cortex, 2009, 2011b) and emotion control (Stolarova er al., 2006) by themselves yield amplitude modulations of the C1 component during visual perception, we hypothesized a possible combined effect of the two factors influencing this early visual evoked component.…”

Occipital and orbitofrontal hemodynamics during naturally paced reading: An fNIRS study

“…Obleser & Kotz, 2010). Such effects are explainable by a predictive coding framework, which suggests that visual representations of expected events can become active previous to stimulus exposure (Rao & Ballard, 1999;Rauss et al, 2011). Similarly, the free energy principle suggests that an iterative search for a plausible 'cause' for the unexpected event can explain such a greater occipital activation (Friston, 2010).…”

“…The cognitive system uses context information to predict upcoming stimuli (Bar et al, 2006;Friston, 2010;Rao and Ballard, 1999;Rauss et al, 2011;Summerfield & Egner, 2009). When a stimulus is presented, it can either match these predictions or not.…”