Neuropathologic Associations of Learning and Memory in Primary Progressive Aphasia

IMPORTANCE The dementia syndrome of primary progressive aphasia (PPA) can be caused by 1 of several neuropathologic entities, including forms of frontotemporal lobar degeneration (FTLD) or Alzheimer disease (AD). Although episodic memory is initially spared in this syndrome, the subtle learning and memory features of PPA and their neuropathologic associations have not been characterized. OBJECTIVE To detect subtle memory differences on the basis of autopsy-confirmed neuropathologic diagnoses in PPA. DESIGN, SETTING, AND PARTICIPANTS Retrospective analysis was conducted at the Northwestern Cognitive Neurology and Alzheimer's Disease Center in August 2015 using clinical and postmortem autopsy data that had been collected between August 1983 and June 2012. Thirteen patients who had the primary clinical diagnosis of PPA and an autopsy-confirmed diagnosis of either AD (PPA-AD) or a tau variant of FTLD (PPA-FTLD) and 6 patients who had the clinical diagnosis of amnestic dementia and autopsy-confirmed AD (AMN-AD) were included. MAIN OUTCOMES AND MEASURES Scores on the effortless learning, delayed retrieval, and retention conditions of the Three Words Three Shapes test, a specialized measure of verbal and nonverbal episodic memory. RESULTS The PPA-FTLD (n = 6), PPA-AD (n = 7), and AMN-AD (n = 6) groups did not differ by demographic composition (all P > .05). The sample mean (SD) age was 64.1 (10.3) years at symptom onset and 67.9 (9.9) years at Three Words Three Shapes test administration. The PPA-FTLD group had normal (ie, near-ceiling) scores on all verbal and nonverbal test conditions. Both the PPA-AD and AMN-AD groups had deficits in verbal effortless learning (mean [SD] number of errors, 9.9 [4.6] and 14.2 [2.0], respectively) and verbal delayed retrieval (mean [SD] number of errors, 6.1 [5.9] and 12.0 [4.4], respectively). The AMN-AD group had additional deficits in nonverbal effortless learning (mean [SD] number of errors, 10.3 [4.0]) and verbal retention (mean [SD] number of errors, 8.33 [5.2]), which were not observed in the PPA-FTLD or PPA-AD groups (all P < .005). CONCLUSIONS AND RELEVANCE This study identified neuropathologic associations of learning and memory in autopsy-confirmed cases of PPA. Among patients with clinical PPA syndrome, AD neuropathology appeared to interfere with effortless learning and delayed retrieval of verbal information, whereas FTLD-tau pathology did not. The results provide directions for future research on the interactions between limbic and language networks.