摘要：Controversy persists regarding the locus of anomia in semantic variant primary progressiveaphasia (svPPA). One possible explanation holds that naming impairment reflects the loss of anamodal conceptual system. An alternative account holds that lexical and/or perceptual accessproblems compromise naming. Eyetracking can potentially provide a sensitive index of featurerepresentation and online search during visual confrontation naming. In previous work, we foundthat healthy older adults fixated on diagnostic features (e.g., eyes, fins), whereas patients with svPPAoften fixate on irrelevant, non-diagnostic features (e.g., color, edges, or discontinuities). Theseanalyses informed us of “ where” people tend to look but complementary information is necessaryregarding “ when” people look to particular regions. We analyzed patterns of visual search from fourpatients diagnosed with svPPA (mean age=66.8) (Gorno-Tempini, 2011). and contrasted lookingcharacteristics for pictures that were anomic (forgotten) relative to named (known) (N=180). Thecore semantic degradation hypothesis predicts that there exists a strong correlation between namingand knowing, such that anomic items should show unique patterns of visual search (e.g., morefixations to irrelevant features). Patients were seated in a quiet room in in front of a 56” plasmascreen television and a 120 Hz table-mounted infrared eyetracking unit (RED/SMI). Participantsnamed common pictures as we monitored eye movements. We predicted that patients with svPPAwould show longer average fixation duration for anomic relative to named pictures. We furtherhypothesized that patients would perform in a manner consistent with core semantic degradation inthe context of preserved executive search. That is, patients would not backtrack or searchinefficiently despite the fact that they attend to features that are not informative of an object’sidentity. The top two panels of Figure 1 illustrate the observed trends for one patient who named“ umbrella” successfully and another who failed to name the same item. For the larger set of picturestimuli attempted (N=548), patients showed longer fixation durations for anomic than named itemst(546)=4.5, p<.001] (Figure 1, bottom panel). svPPA patients showed no significant differences inbacktracking rates for anomic relative to named items [MWU, p=.52, retain null]. As predicted,svPPA patients exhibited well-organized search patterns suggestive of intact executive processes.Moreover, prolonged fixation on unnamed items is consistent with the hypothesis that anomia inthese patients reflects the loss of core semantic knowledge.
