Word picture verification performance reveals auditory comprehension deficits in primary progressive aphasia

Word/picture verification has been found to be a sensitive measure of lexical-semantic abilities in post-stroke aphasia and reveals information about disruptions in semantic and phonological processing. Exploration of the nature of auditory comprehension deficits using word/picture verification has not been replicated in primary progressive aphasia (PPA). We tested 108 in-dividuals with PPA [logopenic variant PPA (lvPPA), n = 50; nonfluent agrammatic PPA (nfaPPA), n = 36; semantic variant PPA (svPPA), n = 22] on a spoken word/picture verification task and a spoken word/picture matching task. The spoken word/picture matching task identified 22 (20%) of individuals as impaired, whereas the spoken word/picture verification task identified 51 (47%) of individuals as impaired (two-tailed p < 0.001). Errors on spoken word/picture verification were due to semantic rather than phonologic foils in lvPPA (nouns, p < 0.001; verbs, p < 0.001), svPPA (nouns, p < 0.001; verbs, p < 0.001), and for nouns (p = 0.001) but not verbs in nfaPPA. Spoken word/picture verification was a more sensitive measure of single-word auditory comprehension deficits in PPA than word/picture matching. The error pattern, consistent with the distributed model of semantic knowledge, suggests that degradation of semantic represen-tations for both nouns and verbs can occur in lvPPA, which may help to distinguish lvPPA versus nfaPPA and guide treatment for anomia.

Automated summary

Word/picture verification is a sensitive measure of lexical-semantic abilities in post-stroke aphasia and reveals disruptions in semantic and phonological processing. However, its application in primary progressive aphasia (PPA) has not been replicated. In this study, we tested PPA patients on spoken word/pictures tasks and found that spoken word/picture verification was more sensitive in identifying auditory comprehension deficits compared to word/picture matching. The error pattern in verification tasks suggests degradation of semantic representations in lvPPA, which can help distinguish different PPA variants and guide treatment for anomia.