Naming in Frontotemporal Dementia and its neural correlates

The study sought to understand the basis for naming problems in the behavioural form of frontotemporal dementia (bvFTD) through direct comparison with semantic dementia (SD) and examination of neural correlates. The data addresses the question whether naming problems in bvFTD arise secondary to patients' frontal executive disorder or occur independently as additional linguistic deficits. The data show test scores for 71 patients with bvFTD and 32 with SD on the Graded Naming test (a published challenging test of naming), the Manchester naming test (a less demanding test of picture naming) and a word-picture matching test that uses the same stimuli as the Manchester naming test (10 pictures of animals, 10 of fruits/vegetables, 10 articles of clothing and 10 household objects). The nature of naming errors is indicated for each test item and are summarised as separate variables. Naming errors are classified in terms of the following types: semantic (e.g. "dog" for rabbit), superordinate category (e.g. "animal" for rabbit), associative (this includes descriptions of an object's function, gestural demonstrations of its use and associated information e.g. "in Australia" for kangaroo), visually related misidentifications, omissions (which includes generic responses that convey insufficient information to convey recognition) and viable alternative responses (e.g. "coat" for jacket). The data also shows independent (blind) visual ratings of magnitude of atrophy based on coronal magnetic resonance scan images. Different regions of the frontal and temporal lobes are rated in accordance with the scheme defined by Davies et al Neuroradiology, 2009, 51, 491-503. Background clinical data includes scores on standard executive tests: verbal fluency, Weigls block sorting and Brixton spatial anticipation test. BvFTD patients, as expected, perform better than the SD group on the language test, but deficits are identified in about half of patients. There are strong inverse correlations between naming scores in both groups and atrophy in temporal lobe structures, particularly the temporal pole and fusiform gyrus, whereas comprehension in SD, measured by word picture matching, correlates more strongly with posterior than anterior temporal lobe atrophy. Error analysis shows a significant relationship in both groups between associative-type responses and temporal pole atrophy, whereas 'don't know' responses, suggestive of a loss of conceptual knowledge, correlates with more posterior temporal regions. There is some correlation in bvFTD between naming and executive performance but not frontal lobe atrophy. The data support the view that naming problems in bvFTD can arise independently of 'frontal' executive deficits and highlight clinical overlap with bvFTD. The findings have implications for the hub-and-spoke model of semantic memory and argue against the notion of the temporal pole as a semantic hub.