Abstract
Human intraparietal sulcus (IPS) has been shown to be retinotopically organized (Swisher et al., 2007). However, the reliable identification of IPS visual field maps in individual subjects requires the collection of a substantial amount of data. Connectivity fingerprint models have proven to be a successful method for predicting area-level organization from structural and resting-state connectivity measures (Saygin et al. 2011; Osher et al., 2015; Tavor et al., 2016; Tobyne et al., 2018). However, these methods have not yet been successfully applied to predict finer scale functional organization, such as that revealed by retinotopic mapping procedures. Here, we applied a connectivity fingerprint model to the Human Connectome Project 7T dataset (N = 181) to characterize individual intra-areal IPS retinotopic organization. A Bayesian regression model was trained to predict vertex-wise IPS polar angle preferences from patterns of resting-state functional connectivity. Model performance was evaluated using a split-half cross-validation scheme. Model predictions were found to accurately reflect IPS polar angle organization in individual subjects. Predicted polar angle maps further allowed for the delineation of boundaries between adjacent intraparietal retinotopic regions, IPS0-3. This work presents a method for potentially examining the retinotopic organization in individuals for which the collection of sufficient retinotopic mapping data is difficult. These findings also extend connectivity fingerprinting methods to predict fine-scale gradients in functional organization within a cortical area.
Acknowledgement: R21 EY027703-01A1 NSF GRFP DGE-1247312