Abstract
The cortical mechanisms of object learning are not well understood. A number of studies have examined the fusiform face area (FFA) as a potential expertise area with mixed results (e.g. Gauthier 1999; Op De Beeck et al., 2006), while other studies have found training-induced changes in the lateral occipital gyrus, a generic object processing region (e.g. Grill-Spector 2000; Kourtzi et al., 2005). However, there has been little focus on the effect of learning within areas that, before learning, respond preferentially to the trained object class. Given that training can result in changes to the tuning properties of task-responsive neurons (Schoups et al 2001), the strongest effects of learning might be expected in regions that are reliably responsive to the trained object class prior to training. The current fMRI experiment focuses on houses, an object class that is associated with three main regions of preferential neural activity (e.g. Epstein et al., 2007) : the parahippocampal place area (PPA), the retrosplenial cortex (RSC), and the temporal occipital sulcus (TOS). Houses also elicit strong behavioural perceptual learning (Husk et al., 2007). BOLD responses to house stimuli were measured during a 1-back matching task before and after 5 days of psychophysical house discrimination training. The behavioural improvement in the 1-back matching task correlated significantly with increases in PPA and RSC, but not TOS, activity. No such correlations were observed for activity within FFA or any retinotopically-defined visual areas. These results suggest that object discrimination training modifies activity within existing object-selective cortical networks without recruiting additional regions.
This work was supported by NSERC Discovery Grants 42133 and 105494 and Canada Research Chairs to PJB and ABS, and an NSERC PGS-D award to JSH.