Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Relative Linguistic Impairment Assessment in Neurodegenerative Diseases Reveals Subtle Deficits and Aids Differential Diagnosis
Poster Session E, Friday, October 2, 11:00 am - 1:00 pm, Wangari Maathai
Grant M. Walker1, Zoe Ezzes2, Lisa Wauters3, Willa Keegan-Rodewald4, Siddarth Ramkrishnan4, Ryan Newbury4, Nicoletta Biondo4, Maria Luisa Mandelli4, William W. Seeley4, Howard J. Rosen4, Zachary Miller4, Maria Luisa Gorno Tempini4, Gregory Hickok1; 1University of California, Irvine, 2University of California San Diego / San Diego State University, 3University of Texas Austin, 4University of California San Francisco
The Relative Linguistic Impairment (RLI; Walker et al., 2024) assessment was developed to evaluate specific cognitive processes involved in picture naming, including word retrieval (lexical) and phonological encoding (sublexical), using stroke data. This study investigated whether the RLI assessment produces comparable patterns of impairments in neurodegenerative disorders. We evaluated 116 consecutive patients referred for neurodegenerative conditions between September 2023 and May 2025, including 15 patients with nonfluent, 16 with logopenic, and 10 with semantic variant primary progressive aphasia (PPA). Additional diagnoses included Alzheimer’s disease, unspecified PPA, amyotrophic lateral sclerosis, corticobasal syndrome, frontotemporal dementia, mild cognitive impairment, Parkinson’s disease, progressive supranuclear palsy, and traumatic brain injury. Participants completed the RLI assessment and the 15-item Boston Naming Test (BNT). Repeated-measures ANOVA was used to test main effects and interactions among diagnosis (nonfluent, logopenic, semantic, or other), error type (phonological or semantic), and RLI stimulus type (lexical or sublexical). Barnard's exact tests identified individual differences in accuracy rates between stimulus types (p < .1). A BNT cutoff score ≤11 was considered clinically significant, and a control sample (N = 20) established normative RLI scores. Structural 3T MRI data were available for 105 patients, with tissue volumes estimated using the Brainnetome atlas and correlated with RLI accuracy, controlling for stimulus type, age, sex, and intracranial volume. Results showed significant main effects of diagnosis (p = .003) and error type (p = .037), but not stimulus type alone (p = .082). Critically, the interaction between error type and stimulus type was significant (p < .001), indicating that stimulus type influences error patterns. A significant three-way interaction among error type, stimulus type, and diagnosis (p = 0.029) demonstrated that diagnosis affects how error types respond to stimulus manipulations. Stimulus-driven changes in semantic error rates were significant across all PPA variants, while changes in phonological error rates were only significant in logopenic PPA; phonological error rates remained high in nonfluent PPA and low in semantic PPA. Four participants showed significant sublexical impairment and eleven showed significant lexical impairment. Notably, eleven participants (9.5%) had abnormal RLI scores despite normal BNT scores. Neuroimaging correlations for lexical impairment involved bilateral inferior temporal gyrus (ITG), fusiform, and parahippocampal gyrus; left anterior superior temporal gyrus; and right middle temporal gyrus, amygdala, and hippocampus. Sublexical impairment was associated with bilateral precuneus; left posterior superior temporal gyrus, posterior inferior frontal gyrus, inferior parietal, inferior postcentral, and lateral occipital cortex; and right inferior frontal gyrus and superior parietal lobe. Patients with substantially poorer performance on lexically challenging stimuli had reduced tissue volume in right superior frontal gyrus, ITG, fusiform, and insula, while poorer performance on sublexically challenging stimuli was associated with reductions in left precentral, postcentral, posterior superior temporal, inferior parietal, and lateral occipital regions, and right precuneus and thalamus. The RLI assessment provides clinically meaningful information about stimulus-driven naming behaviors linked to distinct neurodegenerative patterns. Importantly, the RLI detected naming impairments even in individuals within normal limits on a standardized picture naming measure, suggesting increased sensitivity to subtle language dysfunction.
Topic Areas: Methods, Disorders: Acquired