Eyetracking Reveals Aberrant Visual Search During Confrontation Naming of Alzheimer’s Disease and Primary Progressive Aphasia

• Main Authors: Ashley Bauer, Jinyi Hung
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2015-04-01
• Series: Frontiers in Psychology
• Subjects: Alzheimer's disease; eye tracking; primary progressive aphasia; Semantic Dementia; confrontation naming
• Online Access: http://journal.frontiersin.org/Journal/10.3389/conf.fpsyg.2015.65.00050/full

Introduction Anomia is a core diagnostic marker for Alzheimer’s Disease (AD) and Primary Progressive Aphasia (PPA). There persists controversy as to the locus of the associated naming impairment. Upon one view, anomia reflects loss of core semantic processes that support object knowledge. An alternative perspective holds that anomia reflects access-like impairments. Since aspects of a picture are not equally meaningful for naming, the feature-based spatial precision (e.g., does a patient look at an animal’s eyes instead of its legs?) as well as the temporal resolution of eyetracking can offer uniquely sensitive insights into the etiology of progressive anomia. Methods Participants included a heterogeneous sample of AD (n=4), Semantic variant PPA (n=3), and Logopenic PPA (n=2). Diagnoses were established by an experienced behavioral neurologist (MG) in accord with published criteria (Gorno-Tempini et al., 2011; McKhann et al., 2011). Participant eyetracking characteristics were modeled relative to the performance of neurotypical controls (n=11; Mean age = 21.64, SD = 2.20). Participants named stimuli from the short form BNT as eye movements were tracked via a 120Hz Red-M Eyetracker. We windowed analysis to the first 1000ms post picture presentation and focused on a priori areas of interest (AOI) demarcated by neurotypical controls (e.g., controls fixate on the doorknob in the ‘house’ picture). We did so by subdividing each BNT image into a 16x16 grid subtending approximately 2 cm squared. A single cell within this matrix was treated as a target AOI if >60% of controls fixated on that region (i.e., at least 7 out of 11). We conducted a between-group target AOI analysis examining 1) N-subjects who looked to target AOIs; 2) Entry time to the first fixation within a target AOI. Results Controls naming accuracy was near ceiling (range = 13-15) whereas patients showed a range of impairment on the BNT (range = 1-15). Patients fixated on target AOIs (M = 3.41, SD = 1.94) significantly less than controls (M = 8.22, SD = 1.16) [t(48) = 16.07, p<.01, d = 4.2]. Patients also showed significantly delayed entry times to gaze at the target AOIs (M = 3652.11 ms, SD = 3900.30) relative to controls (M = 712.56 ms, SD = 442.00) [t(46) = -5.44, p<.01, d = 0.75]. Discussion Neurotypical adults show a highly efficient visual search pattern, fixating rapidly on one key semantic feature (e.g., eyes of a face) prior to naming. In contrast, patients often do not look to the same key feature or show significantly delayed entry times to those features. These results are consistent with the claim that patients with AD and PPA experience reduced top-down support for visual search during confrontation naming. These findings suggest that the locus of anomia in AD and the temporal lobe variants of PPA lies within the degraded semantic systems.