Functional connectivity of the scene processing network at rest does not reliably predict human behaviour on scene processing tasks

The perception of scenes is associated with processing in a network of scene-selective regions in the human brain. Prior research has identified a posterior-anterior bias within this network. Posterior scene regions exhibit preferential connectivity with early visual and posterior parietal regions, indicating a role in representing egocentric visual features. In contrast, anterior scene regions demonstrate stronger connectivity with frontoparietal control and default mode networks, suggesting a role in mnemonic processing of locations. Despite these findings, evidence linking connectivity in these regions to cognitive scene processing remains limited. In this preregistered study, we obtained cognitive behavioural measures alongside resting-state fMRI data from a large-scale public dataset to investigate interindividual variation in scene processing abilities relative to the functional connectivity of the scene network. Our results revealed substantial individual differences in scene recognition, spatial memory, and navigational abilities. Resting-state functional connectivity reproduced the posterior-anterior bias within the scene network. However, contrary to our preregistered hypothesis, we did not observe any consistent associations between interindividual variation in this connectivity and behavioural performance. These findings highlight the need for further research to clarify the role of these connections in scene processing, potentially through assessments of functional connectivity during scene-relevant tasks or in naturalistic conditions. Significance Statement Our ability to process scenes is crucial for interacting with our environment as it allows us to extract spatial, contextual, and navigational information. However, the mechanisms by which the scene network in the human brain supports these abilities remain poorly understood. To investigate this, we compared behavioural measures of scene processing with resting-state functional connectivity within the scene network. Extensive individual variability was evident in scene recognition, spatial memory, and navigational abilities. However, contrary to our preregistered hypothesis, we did not observe any consistent associations between task performance and the resting-state functional connectivity of the scene network. These results suggest that future research employing task-related or naturalistic designs may be necessary for elucidating the neural basis of scene perception.