Visual attention and memory under central vision loss

Abstract

Efficient goal-directed interaction with our environment depends on attentional selection. In normal viewing attentive stimulus processing is carried out with high-resolution foveal vision and attentional selection is closely tied to eye movements. Does the loss of central vision impair higher-level visual memory functions that depend on the efficient deployment of attention? This research question is highly relevant to patients with age-related macular degeneration (AMD) who are forced to rely on extrafoveal vision for all processes typically carried out in foveal vision. The experiments presented in this thesis were designed to examine whether encoding of objects into visual long-term memory (VLTM) or efficient attentional guidance by acquired contextual memory suffers when central vision is impaired. Experiments were performed by patients with AMD and normal-sighted younger observers under sco-toma simulation. Patients were tested monocularly with their worse eye and under binocular vision. Recording of eye movements allowed for the additional analysis of exploration behavior. Performance and gaze parameters were compared to respectively matched unimpaired control groups. The first set of experiments examined encoding of natural objects into visual long-term memory, a process that crucially depends on attentive foveation. Change detection sensitivity was measured af-ter free scene exploration. A highly salient cue capturing attention to a non-target location in half of the trials ensured that change detection relied on memory. In Experiment 1, patients’ monocular and binocular change detection sensitivity was preserved, not significantly correlated with visual impair-ment, and went along with normal fixation parameters. A simulated gaze-contingent central scotoma in Experiment 2 complicated visual exploration and impaired memory performance in unexperienced observers. Attentive encoding of objects into VLTM thus depends on attentive foveal vision as long as saccadic rereferencing to an extrafoveal retinal location as a focus of attention has not yet developed. The second set of experiments investigated incidental visual learning of repeated distractor configu-rations during visual search. Contextual cueing leads to faster detection of the target and more efficient exploration in repeated search displays. In Experiment 3, search with a simulated central scotoma eliminated contextual cueing and scan paths became less efficient. In Experiment 4 AMD patients showed reduced contextual cueing with binocular vision. The effect diminished with their worse eye and was correlated with visual acuity. Experiment 5 demonstrates that a simulated central scotoma leads specifically to impaired guidance of attention by learned contexts. When the scotoma was removed after an initial learning phase, contextual cueing reinstated. A simulated annular peripheral scotoma on the other hand, prevented contextual learning by interfering with the integration of invariant local configurations with the global display layout. The experiments presented in this thesis provide evidence that loss of central vision can affect higher-level visual memory processes that depend on attentive foveation in normal vision in different ways. While encoding of objects into VLTM can be compensated when eye movements are efficiently reref-erenced to an extrafoveal retinal location as a focus of attention, automatic guidance of attention in familiar environments remains impaired. Rereferencing may require more top-down controlled visual exploration and compete for visuospatial working memory capacity. These results may contribute to improve training regimes that help patients to better cope with their visual deficits in everyday life.