Concurrent recording of steady-state and transient event-related potentials as indices of visual-spatial selective attention

Abstract

Objectives: The present study investigated the effects of spatial attention on concurrently recorded visual event-related potentials (ERPs) and steady-state visual evoked potentials (SSVEPs) to isoluminant color changes embedded in rapidly flickering stimuli.

Methods: EEG was recorded while subjects attended to flickering LEDs in either the right or left visual hemifield and responded by a button press to isoluminant color changes (targets).

Results: Target isoluminant color changes elicited the typical ERP components P1, N1, N2 and P3, which were enhanced for attended targets compared to unattended targets. Consistent with previous findings, SSVEP amplitude was enlarged for attended flicker stimuli at posterior electrode sites contralateral to the attended visual hemifield. In addition, significant correlations were found between the N1, N2 and the SSVEP attention effects, whereas no such correlations were found between the P1, P3 and SSVEP attention effects.

Conclusions: Results suggest that the SSVEP and ERP reflect partially overlapping attentional mechanisms that facilitate the discriminative processing of stimuli at attended locations.

Introduction

When attention is directed towards a particular location in the visual fields, stimuli presented at that location typically elicit enlarged early event-related potential (ERP) components in relation to stimuli at unattended locations. The components most consistently enhanced by spatially focused attention are the P1 (with a latency of 80–130 ms after stimulus onset) and the N1 (latency 140–200 ms) waves, which have been localized to sources in extrastriate visual cortex – reviewed in Mangun (1995) and Hillyard and Anllo-Vento et al. (1998). These amplitude modulations have been interpreted as evidence of an early sensory gain control mechanism that enhances the signal-to-noise ratio of visual inputs within the ‘spotlight’ of spatial attention and facilitates their transmission to higher cortical areas where feature processing and pattern identification take place (Hillyard et al., 1998). This higher processing of relevant features and patterns is indexed by longer-latency visual ERPs that include selection negativity (150–300 ms), N2 (200–300 ms) and P3 or P300 (300–500 ms) components.

The aforementioned ERPs are elicited as transient components in response to the presentation of individual stimuli. In contrast, a visual stimulus that is presented repetitively at a rate of 6–8 Hz or more will typically elicit a steady-state visual evoked potential (SSVEP), which is a continuous, oscillatory response of the visual cortex having the same fundamental frequency as the initiating stimulus (Regan, 1989). The SSVEP has certain advantages for studying spatial attention in that it is rapidly quantifiable in the frequency domain and provides a continuous measure of the focusing and shifting of attention among items in a visual display. In previous studies we found that the SSVEP elicited by a flickering stimulus was substantially increased in amplitude when attention was directed to its location. This attention-related enhancement was observed for SSVEPs in the frequency ranges of 8–12 Hz (Morgan et al., 1996, Hillyard et al., 1997) and 20–28 Hz (Müller et al., 1998a, Müller et al., 1998b), which correspond to the ‘low’ and ‘medium’ frequency bands described by Regan (1989). These SSVEP amplitude modulations were attributed to a sensory gain control mechanism that acts at the level of extrastriate visual cortex and amplifies visual inputs within the spotlight of attention.

In the studies of Müller et al., 1998a, Müller et al., 1998b) the stimuli consisted of two vertical rows (hereafter called ‘bars’) of 5 flickering LEDs each. One bar was presented to the left visual field (flickering at 20.8 Hz) and the other to the right visual field (flickering at 27.8 Hz; see Fig. 1). At the start of each trial a central cue directed the subject to attend to the flickering bar in one field and to ignore the bar in the opposite field. For most of the trial the LEDs in both bars flickered in red, but at irregular intervals two of the 5 LEDs in one of the bars (left or right at random) simultaneously changed from red to green for a period of 144 ms and then reverted to red. A target was defined as a simultaneous color change of the top and bottom LEDs in the bar in the attended field, to which the subject responded by a button press. Thus, the design of Müller et al. allowed for recording not only SSVEPs to the two continuously flickering bars but also transient ERPs to the brief color-change targets within the attended and unattended bars.

In the present study, we investigated whether spatial attention would influence transient ERPs to the color change stimuli as well as the SSVEP to the ongoing flicker in the paradigm of Müller et al. (1998b). It is well known that isoluminant color changes elicit robust ERPs (Regan, 1989), but only one study to our knowledge has investigated the effects of spatial attention on ERPs to such a stimulus (Wijers et al., 1997). Wijers and coworkers found that green letters appearing against an isoluminant gray background elicited P1 and N1 components that were delayed by 40–50 ms in relation to those elicited by similar luminance-increment stimuli. These delayed P1 and N1 components were enhanced in amplitude when attention was directed to the location of the isoluminant stimuli, which supported the idea that spatial attention acts as a sensory gain control on a visual inputs. In the present study, we wanted to find out whether isoluminant color changes in a rapidly flickering sequence would elicit early ERP components that were similarly enhanced by spatial attention. A further aim was to find out whether attentional modulations of these early transient ERPs were correlated with attention-related amplitude changes of the concurrently recorded SSVEPs to the flickering sequence. Such a relationship would suggest that spatial attention acts very generally to control the gain of both transient and steady-state inputs at an early stage of processing in the visual pathways.