Implicit short- and long-term memory direct our gaze in visual search

Abstract

Visual attention is strongly affected by the past: both by recent experience and by long-term regularities in the environment that are encoded in and retrieved from memory. In visual search, intertrial repetition of targets causes speeded response times (short-term priming). Similarly, targets that are presented more often than others may facilitate search, even long after it is no longer present (long-term priming). In this study, we investigate whether such short-term priming and long-term priming depend on dissociable mechanisms. By recording eye movements while participants searched for one of two conjunction targets, we explored at what stages of visual search different forms of priming manifest. We found both long- and short- term priming effects. Long-term priming persisted long after the bias was present, and was again found even in participants who were unaware of a color bias. Short- and long-term priming affected the same stage of the task; both biased eye movements towards targets with the primed color, already starting with the first eye movement. Neither form of priming affected the response phase of a trial, but response repetition did. The results strongly suggest that both long- and short-term memory can implicitly modulate feedforward visual processing.

Keywords: Eye movements; Implicit memory; Long-term memory; Priming; Visual search.

Fig. 1

a The experimental design (up to block 3) of the experiments used to investigate long-term priming. The colors indicate the trial type distribution in each block. This was equal for both targets in Neutral blocks, but 80 % of targets were ‘bias-colored’ during Biased blocks. Long-term priming is defined as speeded RTs for biased-color targets during neutral blocks induced by this bias. Sub-blocks I–III were defined to explore long-term priming at a finer time scale. The bottom graph depicts this effect (not real data). b Schematic illustration of a search display as used in this study. The task is to search for a red or green diamond (see Methods for further detail)

Fig. 2

a Average response time (RT) as a function of experiment block, separately computed for target type (Bias or Other) and its relation with the previous trial (Repeat or Switch). In each block, shorter RTs are found for target repetitions than for switches (short-term priming). After the bias blocks, which are indicated by shaded areas, bias-color targets also yield faster RTs than other-colored targets (long-term priming). b z-scored RT of trials from Block 3 and 5, corrected for a priori color biases (see text for details), plotted across sub-blocks and separated for trials with a response repetition versus a response switch. This figure again shows the long- and short-term feature priming effect, and illustrates that the long-term priming effect does not attenuate across sub-blocks. In addition, it depicts a response priming effect: repeated responses yield shorter RTs than response switch trials. In these and all subsequent graphs, error bars (or shaded error ribbons) reflect Cousineau-Morey 95 % confidence intervals (Baguley 2011)

Fig. 3

Conform Fig. 2, but using the number of fixations as a dependent measure. a Number of fixations recorded from the onset of the search display to the manual response (excluding the initial central fixation) as a function of target type (bias/other) and target repetition/switch relative to the previous trial. b Number of fixations across sub-blocks, z-scored and corrected for a priori color biases, plotted across sub-blocks and separated for response repetition- and switch-trials. Note the similarity between these data and the pattern of results for RT (Fig. 2): again, a difference was found between repetition and switch trials (short-term priming), and between bias color trials and other color trials (long-term priming). However, the repetition of responses does not seem to affect the number of fixations in a trial

Fig. 4

a The percentage of first fixations in each neutral block (1, 3, 5) that fall on the target color, as a function of whether the target color is the biased/other color, as well as a repetition/switch. Both short-term and long-term priming affect the first fixated color in a trial. b Average time between the onset of the final target fixation on a trial and the response, as a measure of the response phase. This phase is not affected by either short- or long-term feature priming effects (although this latency is affected by the response repetition, described in the text)

Fig. 5

Long- and short-term priming throughout the search, within an epoch of 600 ms following display onset. a Throughout the experiment, the eyes fixate items sharing their color with the previous target color more than items with the other potential target color (here: distractor color). This difference arises at the first eye movement and persists at least throughout the entire epoch. b Long-term priming similarly biases participants’ gaze to items with the bias-target color from the first eye movement onwards. Black bars under the graphs mark timepoints with significant differences between the two colors, determined by the permutation test with TFCE

Fig. 6

Data as in Fig. 5, but during a 750-ms epoch locked to the response. Long-and short-term priming do not seem to affect the response phase of a trial. If either form of priming were to affect the response phase, this would have been reflected by the ‘Switch’ curve (in a) and the ‘Other’ curve (in b) leading with respect to their counterparts. However, these curves follow identical time courses