Abstract
Selective attention improves perception and memory, and designates the target of motor responses, including saccades. Dual-task studies have demonstrated superior perceptual performance at the saccadic goal, implying that a single attentional filter is used by saccades and perception (Gersch et al., 2004; VSS 2006). Does this same filter also determine the input to visual memory?
We examined visual memory during the performance of sequences of saccades. Observers made brisk sequences of oblique saccades across a 5×5 array of circles (diam 1°) following a path designated by a color cue. During a randomly-selected intersaccadic pause, an array of 25 letters (one in each circle) was presented. A post-cue in one of the central 9 circles designated the letter to be reported at the end of the trial.
Saccadic planning had dramatic effects. Performing the saccadic sequences impaired memory in that fewer letters were remembered during intersaccadic pauses (∼2.1 letters) than during steady-fixation control trials (∼3.4 letters). Memory for the target of the upcoming saccade was best (67% correct), exceeding even performance at fixation (43% correct). Performance at locations that were not saccadic targets was near chance levels. A pre-trial cue, disclosing the location of the letter to be reported, succeeded in broadening the attentional window, but did not abolish the effects of saccades. The pre-cue also led to more saccadic errors.
These results support a single attentional filter for both perceptual and motor tasks, and extend the domain of this filter to visual memory. The effects of this filter on memory were, if anything, more pronounced than the effects observed in perceptual tasks. To avoid an overly-narrow attentional field in natural tasks, and remember information off the path of saccades, it may be necessary to adopt a flexible saccadic strategy that sacrifices saccade frequency or accuracy, as needed.
Supported by NIH (EY15522)