Date of Award

May 2019

Degree Type


Degree Name

Master of Science



First Advisor

Ira Driscoll

Committee Members

Christine L Larson, Deborah E Hannula


aging, Alzheimer's disease, biomarker, hippocampus


Markers of cognitive impairment are needed to distinguish normal from pathological aging prior to the onset of clinical symptomology to improve Alzheimer’s disease (AD) treatment or prevention efforts. AD pathology is believed to develop years or even decades prior to diagnosis in medial temporal lobe subregions that provide input to the hippocampus (Braak & Braak, 1991), disrupting the ability of the hippocampus to bind individual elements of an experience to form cohesive memory representations. Eye movement behavior is a sensitive index of learning and effects of memory on eye movements have been shown to emerge rapidly (within 500-750ms of stimuli onset) in healthy, cognitively unimpaired individuals, regardless of task demands (Hannula, Ryan, Tranel, & Cohen, 2007). Importantly, memory-based viewing effects are completely absent from viewing patterns in patients with hippocampal amnesia (Hannula et al., 2007). Thus, eye movements measured during hippocampus-dependent tasks may detect subtle changes in cognition due to early pathology and may permit successful differentiation of normal from pathological aging prior to diagnosis. Fifty healthy, community-dwelling middle-aged (N = 22; age 40-55) and older adults (N = 28, age 65-80) were recruited. The Montreal Cognitive Assessment was used to identify individuals performing in the Mild Cognitive Impairment (MCI) range (“at risk”), MCI being a prodrome of dementia. Participants completed the Scene Face Pair Task (SFPT), where they were studied an arbitrary set of unique scene-face pairs. During the test phase, participants viewed 3-face displays superimposed on top of previously studied scenes while eye movements were recorded. Participants were asked to indicate during the test phase whether the associate (i.e. the matching face of the previously studied scene) was present or absent in the 3-face display. All groups distinguished target-present from target-absent displays, although recognition performance was worse in the “at risk” group. While all three groups directed viewing disproportionately to the matching face within 500-750ms of 3-face display onset, memory-based viewing throughout the full time course of 3-face display in the “at risk” group was less robust. These results suggest that the SFPT, combined with explicit recognition and eye movement indices of performance, may be able to detect early deficits associated with pathological aging.