Abstract
It is commonly acknowledged that visual imagery and perception rely on the same content-dependent brain areas in the high-level visual cortex (HVC). However, the way in which our brain processes and organizes previous acquired knowledge to allow the generation of mental images is still a matter of debate. Here, we performed a representation similarity analysis of three previous fMRI experiments conducted in our laboratory to characterize the neural representation underlying imagery and perception of objects, buildings and faces and to disclose possible dissimilarities in the neural structure of such representations. To this aim, we built representational dissimilarity matrices (RDMs) by computing multivariate distances between the activity patterns associated with each pair of stimuli in the content-dependent areas of the HVC and HC. We found that spatial information is widely coded in the HVC during perception (i.e. RSC, PPA and OPA) and imagery (OPA and PPA). Also, visual information seems to be coded in both preferred and non-preferred regions of the HVC, supporting a distributed view of encoding. Overall, the present results shed light upon the spatial coding of imagined and perceived exemplars in the HVC.
Highlights
Mental imagery corresponds to the human ability to access perceptual information from memory to create a complex and sophisticated mental experience of objects, people, or places (Farah 1989; Kosslyn 1980)
In a first preliminary experiment (Boccia et al 2015), we found that topological mental images—namely, images in which it is possible to navigate (Guariglia and Pizzamiglio 2006, 2007)—activate the same scene-selective brain regions required for perception of landmarks in the high-level visual cortex (HVC), i.e., the parahippocampal place area (PPA) and the retrosplenial complex (RSC); whereas, nontopological images—namely, images in which it is not possible to navigate (Guariglia and Pizzamiglio 2006, 2007)— activate a different set of brain areas
In a second functional magnetic resonance imaging (fMRI) experiment (Boccia et al 2017), using multi-voxel pattern classification (Kriegeskorte and Bandettini 2007a, b), we found that item-specific information from perceived landmarks was re-instantiated during mental imagery of the same landmarks in scene-selective regions (i.e., PPA and RSC) as well as in the HC
Summary
Mental imagery corresponds to the human ability to access perceptual information from memory to create a complex and sophisticated mental experience of objects, people, or places (Farah 1989; Kosslyn 1980). Brain Structure and Function (2021) 226:1511–1531 by Haxby and colleagues (Haxby et al 2001), in which the authors used correlations between response patterns as an index of similarity, it has become clear that the representations of faces and objects in the HVC may be widely distributed These authors found that the category of the presented stimulus (e.g., a face) could be identified from the distributed pattern of activity in the HVC even after excluding the area that maximally responded to that category (e.g., FFA) from the analysis, and even when limiting the analysis to regions maximally responding to another category (e.g., PPA). Different perceptual categories share the neural space when they share common attributes (O’Toole et al 2005)
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call