DOI: 10.1002/alz.082444 ISSN: 1552-5260

A neuropsychological investigation of social‐semantic knowledge in frontotemporal dementia

Matthew A Rouse, Ajay D Halai, Siddharth Ramanan, Karalyn Patterson, James B. Rowe, Matthew A Lambon Ralph
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



The loss of semantic memory is a key feature of frontotemporal dementia (FTD), and is associated with atrophy of the anterior temporal lobes (ATLs). The role of the left and right ATLs in supporting social‐semantic knowledge is unclear. To address this, we developed a social‐semantic battery comprising exemplars of social concepts such as person‐specific knowledge, abstract social concepts, and understanding of social norms. Performance on this battery was directly compared with non‐social semantic knowledge, measured using the Cambridge Semantic Test Battery.


48 people with FTD (behavioural‐variant FTD = 26, semantic dementia = 22), and 19 age‐matched healthy controls were recruited. Participants completed both semantic batteries, alongside tests of general cognition and executive function. Participants also underwent a 3T T1‐weighted structural MRI scan. All neuropsychological tasks were entered into a principal component analysis (with varimax rotation) to explore graded variations in cognitive performance in FTD. Voxel‐based morphometry was conducted to explore grey matter correlates of social and non‐social semantic knowledge. The relative contributions of the left and right ATL to social‐semantic knowledge were tested by including ATL volume and asymmetry indices as predictors in a multiple regression analysis.


People with SD and bvFTD were impaired in both social‐and non‐social knowledge tasks. The principal component analysis indicated separate components for semantic memory and executive function. Both social and non‐social semantic tasks loaded onto the semantic memory component (Figure 1). People with SD scored significantly worse than bvFTD on the semantic memory component (t = 5.18, p<0.0001), while people with bvFTD scored worse than SD on the executive function component (t = 3.97, p<0.001) (Figure 2). Semantic memory factor scores were correlated with grey matter in the bilateral ATLs (Figure 3). There were no significant clusters associated with scores on the executive function component. Multiple regression revealed that performance on the semantic memory component was predicted by the magnitude but not the asymmetry of ATL atrophy.


Both social and non‐social forms of semantic knowledge are impaired by FTD, and this degradation is associated with bilateral ATL atrophy. These results suggest a shared representation for all types of semantic memory across the ATLs.

More from our Archive