Preferential signal pathways during the perception and the imagery of familiar places: a DCM study

The ability to perceive and remember relevant information about the space around us is necessary to correctly navigate in the environment. Perceiving and imagining visual stimuli, such as buildings and landscapes, activate similar content-dependent brain areas spanning from occipital to temporo-medial regions. However, how these areas interact with each other during perceptual stimulation and visual imagery of familiar scenes, especially when recollecting their spatial location, remains still unknown. In this study, we combined functional magnetic resonance imaging (fMRI) and resting-state functional connectivity (rs-fc) to assess the spontaneous fluctuation of signal among regions entailing scene-processing, the primary visual area (V1) and the hippocampus (HC), responsible for the retrieval of stored information. Afterward, we performed an effective connectivity analysis using Dynamic Causal Modelling (DCM) to evaluate whether the dynamic couplings among these regions differ between perception and imagery. The present study is based on a re-analysis of BOLD data from two datasets. The first one included fMRI resting-state data of 127 subjects, some of which (44 subjects) also underwent a functional localizer for scene-selective regions: the Parahippocampal Place Area (PPA), the Retrosplenial Complex (RSC) and the Occipital Place Area (OPA). The second dataset included data from 14 healthy subjects who participated in a study (Boccia et al., 2017), consisting of three scans of a scene perception-imagery experiment and two localizer scans for scene-selective regions. In the perception-imagery experiment, participants were asked to recall the spatial position of buildings within the University campus, in both perceptual and imagery trials. Specifically, pictures of the buildings were presented in the perceptual trials, whereas written labels were displayed on the screen in imagery trials. Firstly, we performed (rs-fc) analysis on the first fMRI dataset to study the functional connectivity profile at rest of scene selective regions, the primary visual area V1, and the hippocampus. Concurrently, we used the results of rs-fc analysis as constraints for studying the effective connectivity among the aforementioned regions using the Dynamic Causal Modelling (DCM) (Friston et al., 2003). Indeed, we applied DCM on BOLD data of the second dataset with the aim of inferring whether the neural communication among the beforehand regions differs during perception and imagery of familiar places. For each subject and region, we extracted the first eigenvariate of adjusted data. We specified and inverted a DCM with a full model among regions. Finally, we used Parametric Empirical Bayes (PEB) (Friston et al., 2016) analysis to estimate parameters at the group level. The whole-brain analysis revealed a consistent separation in two portions of the parahippocampal place area (PPA) – the anterior and the posterior PPA – in our sample. The rs- fc analysis revealed a connectivity pathway similar to the one observed in macaques and different preferential connectivity linking the anterior and the posterior PPA with the retrosplenial complex (RSC) and the occipital place area (OPA), respectively. Afterward, we performed an effective connectivity analysis using Dynamic Causal Modelling (DCM) to evaluate whether the dynamic couplings among these regions differ between perception and imagery. We found a strong and positive effect of the hippocampus on RSC during the retrieval of imagined landmark positions and a strong effect of occipital regions on both RSC and pPPA during the perception of familiar places.