Behavioral Neuroscience © 2018 American Psychological Association 2018, Vol. 132, No. 5, 317–338 0735-7044/18/$12.00 http://dx.doi.org/10.1037/bne0000261 Heterogeneity in Human Retrosplenial Cortex: A Review of Function and Connectivity Elizabeth R. Chrastil University of California, Santa Barbara Retrosplenial cortex (RSC) is an important information hub in the brain and several mental disorders demonstrate RSC dysfunction, but its role is still largely unclear. Although researchers in many cognitive domains have recognized the importance of RSC, a broader synthesis of RSC function across cognitive domains is lacking. This review examines human RSC function across several cognitive domains, considering both specific cognitive functions and the RSC subregions in which that function occurs. Overall, this review found evidence for a functional gradient across the anterior-posterior axis of RSC involving several cognitive domains. Within the cognitive realm of navigation, RSC is important for path integration (including head direction), landmark processing, and the transformation between viewpoints. The related cognitive domain of scene processing encompasses information about place recognition and spatial context. Both navigation and scene processing are localized to more posterior subregions of RSC. Episodic memory (particularly episodic recall), mental imagery, and self-referential processing tend to be supported by anterior portions of RSC. The heterogeneity of RSC function is consistent with RSC anatomy and connectivity found in animal models. Finally, this review examines several common themes that emerged, including mental imagery and self-referential processing. Both the functional heterogeneity and the common themes of RSC function could provide new avenues for research and insight into the numerous mental disorders characterized by RSC dysfunction. Keywords: episodic memory, mental imagery, navigation, scene perception, self-referential processing Retrosplenial cortex (RSC) has been associated with many This review has two primary goals. First, this review aims to diverse functions. In animal models, studies of RSC function are highlight functional differences across subregions of RSC derived largely focused on spatial navigation as well as spatial and con- from human neuroimaging studies, arguing for a functional gradi- textual memory. In humans, RSC is known to be involved in a ent across the anterior-posterior axis of RSC. Similar gradients in wide variety of cognitive functions, including scene perception the hippocampus (Kjelstrup et al., 2008; Poppenk, Evensmoen, (Epstein & Higgins, 2007; Park, Intraub, Yi, Widders, & Chun, Moscovitch, & Nadel, 2013) and medial entorhinal cortex (Gio- 2007; Summerfield, Hassabis, & Maguire, 2010), spatial naviga- como, Zilli, Fransén, & Hasselmo, 2007) relate to the spatial scale tion (Auger, Mullally, & Maguire, 2012; Chrastil, Sherrill, Has- of processing, while gradients of cognitive function have been selmo, & Stern, 2015; Sherrill et al., 2013; Wolbers & Büchel, observed in parahippocampal cortex (Baldassano, Beck, & Fei-Fei, 2005), episodic and autobiographical memory (Steinvorth, Corkin, 2013; Baldassano, Esteva, Fei-Fei, & Beck, 2016) and precuneus & Halgren, 2006; Summerfield, Hassabis, & Maguire, 2009; Vin- (Peer, Salomon, Goldberg, Blanke, & Arzy, 2015). Thus, wherever cent et al., 2006), head direction (Baumann & Mattingley, 2010; possible, anterior/posterior and medial/lateral distinctions will be Marchette, Vass, Ryan, & Epstein, 2014; Shine, Valdés-Herrera, made throughout the review (see Figure 1). Second, this review Hegarty, & Wolbers, 2016), pain and emotional processing (Kucyi aims to reach across cognitive domains to examine the functional et al., 2014; Luo et al., 2014), and mental imagery (Boccia et al., commonalities throughout the entire RSC region. Part of this This document is copyrighted by the American Psychological Association or one of its allied2015 publishers. ; Summerfield et al., 2009). Previous research has taken a process will be to examine RSC contributions to all of its disparate This article is intended solely for the personal use ofdomain-specific the individual user and is not to be disseminated broadly. approach to understanding RSC, yielding consid- functions and to consider what aspects they share, with suggestions erable insight into RSC function, but also preventing a synthesis of where future work could test these potential unifying functions. across domains into a broader understanding of the region. The review will propose that the strongest commonalities are found in self-referential processing, imagery, and memory. This review is designed for researchers of both human and This article was published Online First August 30, 2018. animal models of RSC structure and function, with an emphasis on I thank Chantal Stern, Sean Tobyne, Rachel Nauer, and Allen Chang for anatomical and functional variation across the RSC region. The their many insightful conversations about RSC. A portion of these ideas scholarly traditions and history of RSC research in animals and were presented at the UC Irvine International Conference on Learning and humans differ (Knight & Hayman, 2014), and each tradition can Memory, 2018. Correspondence concerning this article should be addressed to Elizabeth offer insight to the other. Researchers of human RSC can gain R. Chrastil, Department of Geography, University of California, Santa insight from anatomical and structural connectivity studies in Barbara, 1832 Ellison Hall, UCSB, Santa Barbara, CA 93106-4060. E- rodents and nonhuman primates, as well as some key lesion and mail: [email protected] electrophysiological recording studies (for detailed analysis of 317 318 CHRASTIL Figure 1. Anatomy of the human retrosplenial cortex (RSC) region. Human anterior RSC corresponds to ventral RSC in the rodent, and posterior in the human corresponds to rodent dorsal. Left: sagittal view of brain near the midline. Right: Axial view. Subregions are shown in different colors and described from anterior to posterior. Red (medium dark gray): posterior cingulate isthmus, very often included as part of anterior RSC. Blue (darkest gray): Anterior RSC, the most anatomically defined section of RSC. Orange (lightest gray): Parietal- occipital sulcus, part of posterior RSC and part of the functionally defined retrosplenial complex. Green (medium light gray): Extreme posterior RSC, part of the functionally defined retrosplenial complex. This subregion does not include any anatomical RSC, but does include part of the calcarine sulcus. See the online article for the color version of this figure. rodent behavior and RSC function, see Clark et al., in press; Smith Histological studies provide a more fine-grained approach to et al., in press; Todd & Bucci, 2015). Those using animal models defining RSC, making distinctions between cell types and layers. can benefit from the wide variety of complex cognitive tasks used Granular cortex is located in Brodmann areas (BA) 29a-c in in human neuroimaging, as well as from observing RSC dysfunc- anterior RSC (corresponding to ventral in the rodent), whereas tion in clinical populations. dysgranular cortex is found in BA30 in the posterior (dorsal) The review begins with an overview of RSC anatomy and region (Kobayashi & Amaral, 2000; Vann et al., 2009). Histology connectivity based on animal models. It then proceeds through also suggests that much of RSC is buried into the sulcus of the several primary functions of RSC as explored through human corpus callosum (Braak, 1979; Kobayashi & Amaral, 2000). Area neuroimaging, including navigation, scene perception, episodic 30v in the depths of the anterior calcarine sulcus and area 23 in the memory, and body/self-related processing. Throughout, differ- nearby posterior cingulate are often included in discussions of ences between anterior and posterior RSC will be highlighted, and greater RSC (Kobayashi & Amaral, 2000; Morris, Petrides, & the evidence supporting an anterior-posterior gradient of function Pandya, 1999), although this has been debated (Vann et al., 2009). will be examined in a separate section. Finally, the review will Unfortunately, cytoarchitecture cannot typically be resolved from examine several potential candidates for unifying RSC function standard-resolution MRI (although see Fischl et al., 2008). Thus, across cognitive domains. most in vivo human MRI imaging includes BA29, BA30, and a large swath of BA23. Anatomy Employing functional definitions of RSC in human fMRI can In humans, RSC is a three-dimensional diagonally oriented circumvent many of these problems, and can be very useful for This document is copyrighted by the American Psychological Association or one of its allied publishers. region (see Figure 1), such that the anterior portion also tends localizing function within individuals for further analysis, such as This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. to be superior and medial, and the posterior portion tends to be for multivoxel pattern analysis (MVPA). However, this approach inferior and lateral. Because of this oblique orientation, these is less useful for generalizing across studies or cognitive domains terms are often used interchangeably—as this review will do. because they restrict interpretation to an a priori function. In The splenium of the corpus callosum and parietal-occipital addition, BOLD registration with functional definitions