Event-related brain potential correlates of visual awareness

Abstract

Electrophysiological recordings during visual tasks can shed light on the temporal dynamics of the subjective experience of seeing, visual awareness. This paper reviews studies on electrophysiological correlates of visual awareness operationalized as the difference between event-related potentials (ERPs) in response to stimuli that enter awareness and stimuli that do not. There are three candidates for such a correlate: enhancement of P1 around 100 ms, enhancement of early posterior negativity around 200 ms (visual awareness negativity, VAN), and enhancement of late positivity (LP) in the P3 time window around 400 ms. Review of studies using different manipulations of awareness suggests that VAN is the correlate of visual awareness that most consistently emerges across different manipulations of visual awareness. VAN emerges also relatively independent of manipulations of nonspatial attention, but seems to be dependent on spatial attention. The results suggest that visual awareness emerges about 200 ms after the onset of visual stimulation as a consequence of the activation of posterior occipito-temporal and parietal networks.

Introduction

Brain imaging methods have been used during the last 10–15 years to study neural correlates of visual awareness. Results suggest that activation along the ventral visual pathway from V1 to temporal cortex correlates with changes in contents of visual awareness (Bar et al., 2001, Beck et al., 2001, Logothetis, 1998, Moutoussis and Zeki, 2002, Pins and ffytche, 2003, Tong et al., 1998, Vanni et al., 1996). In addition, visual awareness is often accompanied by activation of frontal and parietal areas which are typically associated with attention (Beck et al., 2001, Haynes et al., 2006, Lumer and Rees, 1999, Naghavi and Nyberg, 2005, Pins and ffytche, 2003, Rees and Lavie, 2001). At present, it is not clear whether these areas are specifically related to visual awareness or whether they just reflect the fact that aware stimuli tend to draw more attention than unaware ones. Because of their relatively poor temporal resolution, current brain imaging methods do not reveal the temporal course of neural processing or the temporal order in which the areas are activated during aware perception.

Fine-grained information about the temporal dynamics of neural processing can be obtained by recording electroencephalography while the participant is performing a cognitive task. The electrical potential changes that are time-locked with sensory or cognitive events are called event-related brain potentials (ERPs). They seem to be an ideal method to directly track the time course of neural processing with millisecond precision during cognitive tasks. The ERP waveform (Fig. 1) consists of positive and negative deflections, called ‘peaks’ or ‘components’. They are labelled with “P” to indicate positivity and “N” to indicate negativity relative to a reference electrode, followed by a number to indicate the order of occurrence or the timing of the peak (e.g., N2 refers to the second negative peak; N200 refers to a negative peak occurring 200 ms after the stimulus onset). The early peaks in the ERP waveform, such as P1 around 100 ms after stimulus onset, are generally thought to reflect sensory (exogenous) responses to the stimulus, because they are influenced by external stimulus factors (e.g., luminance) and elicited by stimuli independent of the task the participant is doing (Luck, 2005). Already the P1 wave can be, however, modified by internal factors such as attention (Luck and Ford, 1998). The later a peak occurs in the waveform, the more likely it is to reflect higher (‘endogenous’) cognitive processes and not the physical properties of the stimulus. For example, the P3 around 300–400 ms after stimulus onset is a typical endogenous component as it depends on internal rather than external factors (Donchin and Coles, 1988, Luck, 2005).

The research on attention during the last 30 years demonstrates the usefulness of the ERP approach. Attentional selection has been shown to have well-documented, characteristic effects on ERPs. For example, the P1 and N1 amplitudes are enhanced by spatial attention (Hillyard and Anllo-Vento, 1998, Luck and Ford, 1998). Stimuli that fall into an attended location elicit larger P1 and N1 amplitudes at posterior scalp sites than do stimuli falling into the same location when this location is not attended to. These early effects have been interpreted to show that attention modulates processing at an early sensory stage. Another well known phenomenon is selection negativity (SN). Selective attention to target objects or features, as compared to nontargets, is reflected in ERPs as a negative amplitude shift at posterior electrode sites around 200 ms after the stimulus onset. SN provides a high-resolution measure of the time at which a particular feature is discriminated and selectively processed in the brain (Hillyard and Anllo-Vento, 1998, Proverbio and Zani, 2003). N2pc (N2-posterior-contralateral) is an other negative attention-related component that occurs after 200 ms. It occurs over the contralateral hemisphere relative to the visual field of a target stimulus, when attention is deployed to the target within an array of distractros. N2pc has been interpreted to reflect allocation of attention to targets (Eimer, 1996), focusing of attention (Hopf et al., 2000), or deployment of attention to reduce interference between an attended target and nearby distractors (Luck and Ford, 1998, Luck and Hillyard, 1994).

Recently, also visual awareness has been studied using ERPs. Such research has progressed along several separate lines, with each study typically focusing on a specific subfield, usually determined by the experimental technique or by the perceptual phenomenon that has been used to manipulate awareness (e.g., masking, change blindness, or binocular rivalry). The results are scattered in the literature and rarely compared with each other. There is a need for a review which draws together and integrates the results from different subfields and summarizes and tries to identify whether there is a common ERP marker of visual awareness. In order to do that, we here review the studies focusing on ERP correlates of visual awareness that have been published within the last 10 years, that is, during the time when consciousness has been an accepted topic of research in cognitive neuroscience. We include studies manipulating the contents of visual awareness with different forms of masking, contrast level, attentional blink, change blindness, and bistable perception. We exclude early ERP studies on visual masking (e.g., Jeffreys and Musselwhite, 1986), some of which have been reported already during 1960s and 1970s, because the standards of EEG recording, data presentation and statistical analyses in these studies do not meet modern requirements, making it difficult to relate their findings to those in the more recent studies.

Studies that have measured ERPs separately in aware and unaware condition without a comparison of the ERPs between these conditions are not reviewed, because they lack the crucial information for the purposes of the present review. It is precisely the electrophysiological difference between the conditions that can be operationalized as the correlate of visual awareness. We focus on three electrophysiological differences in standard ERPs (enhanced P1, enhanced negativity in N1–N2 range, enhanced P3) which have been suggested to correlate with visual awareness (see Table 1). Usually such differences overlap with the classical ERP components (e.g., Fig. 1, left) and are observed most clearly in difference waves (e.g., VAN and LP in Fig. 1, right). We do not review results from other electrophysiological measures, such as time–frequency analyses of EEG (Ohla et al., 2007), as they are not directly comparable to ERPs and would thus need a review of their own.

The review of recent studies shows that the most reliably and consistently observed ERP correlate for subjective visual awareness of a stimulus, as compared with a stimulus that does not enter awareness, seems to be an increase of negativity at posterior recording sites around 200 ms. We have called this effect visual awareness negativity, VAN (e.g., Koivisto and Revonsuo, 2003) (see Fig. 1). Because it is important not to confound the electrophysiological effects of visual awareness with those of attention, the last part of the paper presents a review of studies that have focused on the relationship between attention and ERP correlates of visual awareness.