![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
Background: The potential differential impact of Alzheimer's disease (AD) across semantic categories/domains (i.e., living/nonliving) has been debated extensively in the past 30 years. An important methodological consideration in this area is the issue of whether category effects are genuine or a by-product of intrinsic properties of items, i.e., nuisance variables (NVs).
Aims: The study's aim was to investigate whether NVs versus semantic domain of words generated in a semantic fluency task better predict semantic impairment.
Methods & Procedures: We examined semantic fluency performance of demographically matched AD patients and controls. AD patients were longitudinally examined over a 2-year period. Norms of all NVs known to differ across domains were obtained for each word generated; influence of domain was also studied.
Outcomes & Results: AD patients generated fewer words than controls. However, both groups showed fairly similar performance: more familiar words, with earlier acquisition, and representing concepts with higher manipulability were generated. Domain exerted similar influence on semantic fluency performance in controls and patients at initial evaluation, but it did not influence performance as the disease advanced. In contrast, the role of familiarity increased with disease progression.
Conclusions: The role of NVs—especially familiarity—appears to be, comparatively, a more relevant predictor of longitudinal deterioration than semantic domain. Knowing which variables within a semantic fluency task longitudinally predict cognitive decline is a useful clinical tool and could possibly be a cognitive marker to improve accuracy of neuropsychological evaluation of patients with probable AD.
We wish to thank I.N.S. and John L. Woodard for all their help. We thank Mary Beth Spitznagel, who reviewed the English version of this work, for her help and her valuable comments and kindness. Julia Mayas collaborated in recruiting some healthy participants. We are also indebted to Professors Keith Laws and Dolores Luna, for their valuable observations. This research was supported by Grant AP 2003-3064 to PRM from the Spanish Ministry of Education and Science.
Notes
1Earlier data analyses showed that typicality was highly correlated with participants' overall performance. This is a regular outcome of the semantic fluency task, given that it requires production of representative exemplars of a category (Battig & Montague, Citation1969; Kiran & Thompson, Citation2003b). Additionally, typicality demonstrated substantial multicollinearity with others NVs. Typicality was thus eliminated from analyses, to reduce the potential for data contamination (Berry & Feldman, Citation1985).
