versus temporal (all pcorrected > 0.05, aHPC: χ2 (1) = 4.30, pHPC: χ2 (1) = 4.14, pPHG: χ2 (1) = 0.11). Additionally, the node hierarchies in the right panels of Figure 4A,B emphasize a similar level of connectivity and betweenness centrality for the different conditions. Contrary to our successful memory retrieval analysis, spatial and temporal retrieval networks exhibited similar density; a chi-squared test confirmed that the number of connections is balanced between the spatial and the temporal conditions (χ2 (1) = 1.38 p = 0.24, Fig. 4C, right), arguing that differences in regional connectivity could not be accounted for by lower degrees of connectivity in one condition. Again, differences in regional activation did not account for the pattern of node connectivity as the two were uncorrelated (spatial – temporal: ρ = −0.17, p = 0.4120; temporal – spatial: ρ = −0.09, p = 0.6740). Overall, these findings support the idea that spatial and temporal retrieval contribute somewhat independently to the correct retrieval effect we reported above, acting as subnetworks that overlap within the larger network involved in episodic memory retrieval.
