Abstract
Does the detection of image blur require attention? Perhaps not, since image blur affects saccade target selection; with the eyes avoiding blurred extrafoveal image regions (Geisler, Perry, & Najemnik, 2006; Loschky & McConkie, 2002). Conversely, attention affects spatial resolution in many psychophysical tasks (Carrasco, 2011), so cognitive load could decrease blur sensitivity by reducing available attentional resources. To test this hypothesis, we had subjects detect image blur while looking at real-world scenes in preparation for an easy picture recognition memory task, while simultaneously carrying out an N-back task to manipulate cognitive load. Subjects were required to detect gaze-contingent blur that was presented on every 7[sup]th[/sup] fixation (using an EyeLink 2000) at four different retinal eccentricities (0, 3, 6, and 9 degrees), under four levels of cognitive load (N-back: 0, 1, 2, and 3-back), and two control conditions (single-task blur detection both with and without to-be-ignored N-back letter presentations). Cognitive load was varied between blocks of trials, with order counter-balanced in a 6x6 Latin Square. We measured blur detection thresholds using an adaptive threshold estimation procedure (SIAM; Kaernbach, 1990) for each retinal eccentricity and level of cognitive load. In Experiment 1, an auditory N-back task was presented, to avoid visual interference with the blur detection task. Results showed no effect of cognitive load on blur thresholds. However, there was an effect of eccentricity, with blur thresholds monotonically shifting to lower spatial frequencies as eccentricity increased. In Experiment 2, the N-back task was presented visually, with each letter presented at the point of gaze, to test whether a foveal stimulus was required to produce an effect of cognitive load on blur thresholds. Results replicated those of Experiment 1. Thus, in both experiments, we find evidence that blur detection was unaffected by the level of attentional resources available.
Meeting abstract presented at VSS 2013