Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Cortical Thickness as a Moderator of the Time Course of Pseudoword Rejection in Individuals with Alzheimer’s Disease and Mild Cognitive Impairment
Poster Session C, Thursday, October 1, 10:45 am - 12:45 pm, Wangari Maathai
JungMoon Hyun1, Eve Higby2, Alexandre Nikolaev3,4; 1Auburn University, 2California State University, East Bay, 3University of Eastern Finland, 4University of Helsinki
Introduction: A correct “No” response to a pseudoword in lexical decision is often treated as the endpoint of a failed search for a matching stored word representation. However, pseudowords partially overlap with real words, and this overlap may activate familiar lexical forms before a final “No” response is made. We asked whether the timing of pseudoword rejection is shaped by position-specific orthographic overlap with real words and whether these time-varying effects are moderated by cortical thickness in regions implicated in Alzheimer’s disease (AD) and language processing. Methods: We analyzed lexical decision performance in 62 older Finnish-speaking adults aged 57–85: 17 healthy controls, 24 with mild cognitive impairment (MCI), and 21 with AD. This study analyzed performance for 177 pseudowords, yielding 10,024 correct rejections and 483 false alarms with finite reaction times. For each pseudoword, we queried a Finnish lexical corpus for real words sharing its initial, middle, or final letter sequence. We derived two measures at each position: overlap size (the number of real-word neighbors) and overlap frequency (the frequency of those neighbors). Lexical decision was modeled as a time-to-event process using Cox proportional hazards models. Models also included demographic (e.g., age, education) and orthographic (e.g., word length, bigram frequency) covariates. Structural MRI estimated cortical thickness in left fusiform, middle temporal, supramarginal, and inferior frontal regions, and bilateral entorhinal cortex. These measures were entered as moderators of time-varying orthographic overlap effects. Results: Individuals with AD or MCI were less accurate than healthy controls on pseudoword decisions and slower in their correct pseudoword rejections. In survival models pooling all three groups, initial letter-sequence overlap between pseudowords and Finnish words was the strongest source of time-varying evidence. Position-specific orthographic overlap frequency initially slowed correct rejection, consistent with a word-like pull from frequent onset neighbors, but later shifted toward facilitation. Position-specific orthographic overlap size showed a complementary pattern, with an early tendency to support rejection followed by later competition. Cortical thickness analyses showed the strongest group differences in bilateral entorhinal cortex, with AD showing marked thinning relative to controls and MCI. Importantly, cortical thickness did not simply predict overall speed; rather, it moderated when position-specific orthographic overlap cues influenced lexical decisions. Greater entorhinal thickness amplified the time-varying effect of initial overlap size for correct rejections. For erroneous “Yes” responses, it supported earlier use of positional orthographic regularities and reduced the prolonged pull of high-frequency real-word neighbors. Left inferior frontal thickness modulated later frequency-based effects, consistent with a role in resolving lexical competition. Left supramarginal gyrus thickness suggested a possible involvement in orthography-to-phonology mapping. Conclusions: Pseudoword rejection is not simply a failure of lexical access. It reflects an evolving competition between structural orthographic constraints and partial activation of familiar word forms. AD-related cortical thinning, especially in the entorhinal cortices, appears to alter the time course in which orthographic overlap cues are used, shifting lexical decisions away from early structure-based rejection and toward prolonged vulnerability to the pull of high-frequency real-word neighbors.
Topic Areas: Disorders: Acquired, Computational Approaches