, 2010). Conversely, the ventromedial ATL has strong connections with visual processing regions in ventral posterior temporal cortex (Binney et al., 2012) and shows greater activation when participants make semantic decisions to pictures relative to words (Visser et al., 2012). This visual semantic bias suggests that a C > A effect might be expected in this area, since concrete words are more strongly associated with visual experiences. It is also important to note that other parts of the ATL are equally responsive to all meaningful stimuli, no matter which modality they are presented in. PET and recent distortion-corrected

fMRI studies have identified an area in the inferior temporal and fusiform gyri (which we

term here the ventral ATL) that responds equivalently to spoken words, written words, pictures and non-verbal sounds Ixazomib cell line (Spitsyna et al., 2006, Vandenberghe et al., 1996 and Visser et al., 2012). Hypometabolism in this region has been linked to multi-modal semantic deficits in patients with semantic dementia (Butler et al., 2009 and Mion et al., 2010) and it has been proposed that the ventral ATL acts as a multi-modal convergence Smad inhibitor “hub” that integrates information from modality-specific sites across the brain to form conceptual representations (Binney et al., 2012 and Patterson et al., 2007). While a number of recent neuroimaging studies have demonstrated activation in ventral ATL for concrete concepts (Peelen and Caramazza, 2012, Robson et al., 2014, Visser et al., 2012 and Visser and Lambon Ralph, 2011), we

are not aware of any studies reporting activation in this area for abstract words. This may be RANTES in part due to susceptibility artefacts that make it difficult to obtain reliable signal in this area with standard, gradient-echo fMRI (Devlin et al., 2000 and Visser et al., 2010). While special steps can be taken in image acquisition and processing to combat this problem (e.g., Embleton et al., 2010 and Halai et al., 2014), the vast majority of fMRI studies do not do so and have reduced sensitivity to activation in the ventral ATLs. It is important to address the question of ventral ATL involvement in abstract concepts because, in common with much of the literature on semantic cognition, implemented computational models of the hub theory have focused exclusively on concrete concepts (Lambon Ralph et al., 2007 and Rogers et al., 2004). As a consequence, Shallice and Cooper (2013) have recently proposed that a separate system is required to meet the different challenges of representing abstract concepts. Furthermore, some researchers have proposed that ATL atrophy in semantic dementia primarily affects visual feature knowledge and, as a consequence, has a disproportionate effect on understanding of concrete words (Bonner et al., 2009 and Libon et al., 2013).