Synergistic effect of combined temporal and spatial expectations on visual attention

Abstract

We developed a naturalistic behavioral task to investigate the influence of spatial and temporal expectations on attentional orienting to moving targets. In this task, the movement of an object before its disappearance under an occluding barrier generated expectations concerning the location and/or time of its reappearance. Four different trial types were intermixed, each inducing a different state of expectation: no expectation, only spatial expectation about the location of reappearance, only temporal expectation about the moment of reappearance, and combined spatial and temporal expectation. The behavioral validity of the task was shown by the fact that all expectation conditions produced significantly shorter reaction times than the control state of no expectation. Spatial attention modulated early perceptual analysis in extrastriate areas, as demonstrated by significant enhancement of the visual P1 component. Temporal attention alone had no effect on P1 but instead modulated response-specific components. However, when spatial and temporal attention were combined, the enhancement of perceptual processing was significantly augmented, leading to a greater enhancement of the P1 component than by spatial attention alone. Perceptual analysis reflected by the P1 component correlated significantly with reaction times. In summary, event-related potentials revealed the presence of individual modulatory effects attributable to spatial and temporal expectation as well as synergistic effects indicative of an interaction of the two. This synergistic effect is likely to play a critical role in directing attention to the reappearance of a temporarily occluded moving target, a process of obvious importance in everyday situations.

Figure 1.

Task and behavioral results. a, The three pictures represent the task at three different stages. First, a red ball (dark gray) appeared at the left side of a black screen (white) and moved across the screen in steps (its previous trajectory is indicated by the dashed line). After reaching the gray occluding band, the ball disappeared underneath for two steps. After reappearance, it contained a black dot on 50% of trials. This was only present for the reappearance step. Subjects were required to respond if the black dot was present. b, This expands on the movement of the ball across the screen. The task manipulated spatial and temporal expectations factorially. The four resulting trial types were randomly intermixed: ST, S, T, and N. To create spatial expectation, the ball moved with a straight-line trajectory across the screen (S+ conditions; left column). When there was no spatial expectation, the ball still moved from the left to the right but was placed randomly on the vertical axis at each step (S– conditions; right column). Temporal expectation was created by steps of the same time period (550 ms; T+ conditions; top row). When there was no temporal expectation, each step took a different length of time (200–900 ms) (T-conditions; bottom row). The upright lines below each display represent schematically the duration of each time step. c, A bar graph indicating the median reaction time averaged over subjects (n = 24) for the four conditions. Error bars indicate SE.

Figure 2.

Enhanced attentional modulation of visual components by combined spatial and temporal expectations. Visual components P1 and N1 from grand-averaged waveforms (n = 24 subjects) are shown for the lateral posterior electrodes analyzed as well as midline occipital electrode Oz for the four attentional conditions: combined spatial and temporal expectations (black line), spatial expectation alone (blue line), temporal expectation alone (red line), and no expectation (dashed line). The grand-averaged waveform for the horizontal eye channel is also shown. The polarity of the waveforms is plotted with positive values upward in this and the following figure. The P1 component, showing a main effect of spatial expectation and an interaction between spatial and temporal expectation is highlighted at electrode PO7. The region of the P1 is enlarged in the inset directly below electrode PO7 to show its enhancement by spatial expectation relative to no expectation and its further enhancement by combined spatial and temporal expectations. The topographies of the attentional modulations on P1 (110–130 ms) are shown in the bottom panel (using data referenced to the average of the mastoids), plotted using Advanced Source Analysis software (Advanced Neuro Technology, Enschede, The Netherlands). Topographies are shown on a realistic three-dimensional head model from a posterior and superior perspective, with the left side of the head shown on the left. The voltage values for each expectation condition minus the no-expectation condition (ST-N, S-N) are plotted at each electrode location and interpolated at intermediate sites. Positive voltage is plotted in red, and negative voltage is plotted in blue. The same voltage values are used in both cases to enable comparison of the magnitude of the effects. Contralateral P1 amplitudes were enhanced by spatial expectation and further enhanced by combined spatial and temporal expectation. No modulation by temporal expectation occurred during this time period. The locations of the electrodes from which waveforms are shown are shaded white on the topographies. HEOG, Horizontal electrooculogram.

Figure 3.

Distinct ERP modulations by spatial versus temporal expectation alone. Grand-averaged waveforms (n = 24 subjects) combined for midline and flanking electrodes over frontal to parietal sites for go (solid line) and no-go (dashed line) trials in the presence (thick line) and absence (thin line) of spatial expectations (left column) and temporal expectations (right column). ERPs for midline and flanking electrodes were averaged together, as indicated by the line joining them in the diagram of the electrode montage. Open arrows point to the main effect of response requirement (go versus no-go trials) on the frontal N2 component. The thick filled arrow points to the main effect of spatial expectation on the amplitude of the P3 component. The thin arrow points to the shortening of the peak latency of the P3 component in go trials.