Behavioral performance follows the time course of neural facilitation and suppression during cued shifts of feature-selective attention

Abstract

A central question in the field of attention is whether visual processing is a strictly limited resource, which must be allocated by selective attention. If this were the case, attentional enhancement of one stimulus should invariably lead to suppression of unattended distracter stimuli. Here we examine voluntary cued shifts of feature-selective attention to either one of two superimposed red or blue random dot kinematograms (RDKs) to test whether such a reciprocal relationship between enhancement of an attended and suppression of an unattended stimulus can be observed. The steady-state visual evoked potential (SSVEP), an oscillatory brain response elicited by the flickering RDKs, was measured in human EEG. Supporting limited resources, we observed both an enhancement of the attended and a suppression of the unattended RDK, but this observed reciprocity did not occur concurrently: enhancement of the attended RDK started at 220 ms after cue onset and preceded suppression of the unattended RDK by about 130 ms. Furthermore, we found that behavior was significantly correlated with the SSVEP time course of a measure of selectivity (attended minus unattended) but not with a measure of total activity (attended plus unattended). The significant deviations from a temporally synchronized reciprocity between enhancement and suppression suggest that the enhancement of the attended stimulus may cause the suppression of the unattended stimulus in the present experiment.

Fig. 1.

Stimulus display and illustrations of spectral and spatial patterns of SSVEP amplitudes. (A) Screenshot of stimulus display and flicker frequencies. A total of 125 red and 125 blue dots were spatially intermingled and in continuous random motion. Red dots flickered at 11.98 Hz and blue dots at 16.77 Hz. A color change of the central fixation cross indicated the color of the to-be-attended dots. (B) The grand-average amplitude spectrum obtained by Fourier analysis of SSVEP waveforms elicited under the two attentional conditions averaged across a cluster of three occipital electrodes. Peak amplitudes are located at the two stimulation frequencies. (C) Spline-interpolated isocontour maps of the grand-average SSVEP amplitudes of the mean of both attentional conditions for red (Left) and blue (Right) dots. Black circles around electrode location Oz indicate the cluster of three posterior electrodes used for statistical analysis. (D) Statistical parametric maps of the estimated cortical current-density distributions that give rise to the SSVEP-amplitude difference between attended and unattended dots for both RDKs. Scale represents t² values, and the P < 0.001 threshold for the attended versus unattended comparison corresponds to 8.34. Thresholds were corrected for multiple comparisons. Note that different scales were used.

Fig. 2.

Time courses of reaction times and SSVEP amplitudes. (A) Time course of reaction times to targets after cue onset. Reaction times became faster until the third time bin (287–429 ms). Black circles indicate midpoints of time bins. (B) Grand-average normalized SSVEP amplitudes that were collapsed over red and blue dots. Enhancement of attended stimuli began roughly 220 ms postcue and lasted for several hundred milliseconds. Suppression of unattended stimuli began roughly 360 ms postcue and lasted until the end of the trial. (C) Time course of selectivity (attended − unattended) and total activity (attended + unattended). Both measures show significant temporal changes, but only selectivity was significantly correlated with the time course of reaction times. Gray lines correspond to 95% confidence intervals (within-subjects for reaction times). Time-point zero indicates onset of the cue.