The association between amyloid burden and language outcomes in the wisconsin registry for Alzheimer's prevention (WRAP)


      The early diagnosis of Alzheimer’s Disease (AD) depends upon clinical manifestations of difficulties with learning and memory which negatively impact activities of daily living (Dubois et al., 2007). Beta-amyloid (1-42) (Aβ42) accumulation, may be a first major stage of presymptomatic AD (Sperling et al., 2011). Language deficits, based on retrospective analysis and prospective cohort studies, may also be present years or decades before diagnosis (Snowdon et al, 1996; Garrard et al 2005; Forbes-McKay and Venneri, 2005; Ahmed et al 2013). Given these findings, we investigated whether rates of amyloid burden would be associated with language outcomes from both standardized tests and non-standardized connected speech samples in a healthy, at-risk cohort.


      Participants were recruited from the Wisconsin Registry for Alzheimer’s Prevention (WRAP), a longitudinal cohort (n>1500) enriched for positive family history of AD. The present analyses included 188 WRAP subjects (mean age=61; 68% female; 38% APOE-4+) who had undergone [C-11]Pittsburgh compound B ([C-11]PiB) positron emission tomography (PET) scan, a Fluorodeoxyglucose ([F-18] FDG PET) scan and a 3.0 Tesla magnetic resonance imaging (MRI) scan. Neuropsychological language measures from their most concurrent WRAP visit were used; a subset of 48 individuals had transcribed speech samples from a picture description task and were used for the discourse analysis measures.


      We constructed a series of linear models using separate regions of interest (ROI) as predictors, including one global measure, an average of 8 ROIs. Adjusting for the covariates of age, sex and baseline literacy, we found significant associations between areas of the frontal cortex (Inferior Frontal Gyrus-Right (p = .04) and Left (p=.05), Anterior Cingulate- Right (p=.05) and Left (p=.04), and Orbitofrontal Cortex Right (p=.03) and Left (p=.05) and the syntactic complexity measure from the connected speech samples. No other language outcomes were significantly associated with the hypothesized ROIs.


      We found an association between amyloid burden in areas of the frontal cortex and a spontaneous speech measure of syntactic complexity. This preliminary analysis suggests that discourse analysis may be a sensitive and informative measure of early cognitive change, particularly when paired with other biomarkers of AD pathology such as amyloid burden.
      Table 1Regions of Interest and Language Outcomes from WRAP
      Region of InterestBoston Naming TestPhonemic FluencySemantic FluencySyntactic Complexity - DiscourseSemantic Units - DiscourseSpeech Fluency (Mazes)
      Inferior Frontal Gyrus0.030.861.070.300.030.864.560.04*0.340.560.590.45
      Superior Temporal Gyrus0.000.980.020.900.140.711.100.300.130.720.940.34
      Anterior Cingulate0.110.752.*0.000.950.200.66
      bal 8 Regions of Interest0.040.840.760.380.120.732.710.110.080.790.250.66
      Superior Medial Gyrus0.000.990.020.900.010.950.130.730.130.720.080.78
      itoFrontal Medial Cortex0.030.860.670.410.010.914.510.04*0.340.560.130.72
      Regions reported here are a sum of left and right hemispheres