Fig. 2: Evidence of egocentric-like coding in the medial parietal cortex.

a Morphing trajectories subtending only 7 possible angular conditions to the goal (“egocentric conditions”). b Adaptation approach, modeling fMRI BOLD signal as a function of the log time passed between two presentations of the same egocentric condition. c Results of the adaptation analyses, one-sided t-test as implemented in SPM12, thresholded at p < 0.001 at the voxel level, corrected for multiple comparisons at cluster level with p < 0.05. d Pearson’s correlation (two-sided) between adaptation effect in the medial occipito-parietal cluster and percentage of Hit during the recall task. e Control analyses in occipito-parietal cluster, considering (i) only “off-target” trials, (ii) “blue” and (iii) “green” contexts separately, (iv) whether the goal was on the “left vs. right” without considering the angular distance, and egocentric-like responses after controlling for (vi) distance from the goal or (vii) repetition of the same starting point. For all the control analyses mentioned above, Supplementary Figs. 8–12 report the unbiased whole-brain statistical analyses, while this panel reports participants’ scores and their average in the main occipito-temporal cluster without statistics, to avoid circularity. n = 40 participants. a.u. = arbitrary units. f Cross-context decoding approach, where a classifier is trained to distinguish egocentric conditions in one context and is then tested in the other one, and vice versa (see Methods). Results of a one-sided t-test (SPM12) are thresholded at p < 0.001 at voxel level, corrected for multiple comparisons at cluster level with p < 0.05. g Anatomical intersection between the brain maps resulting from the adaptation analysis and the cross-context decoding. h MDS of the egocentric conditions. Patterns of activity are extracted from the cluster in the precuneus (peak from intersection analysis). i We used the DeepMReye toolbox⁴¹ to estimate gaze behavior from the eyeballs’ BOLD signal. Trial-wise eye movements were correlated with the “true” position of the goal in conceptual space. j Fisher-transformed Pearson’s correlation score was significant at the group level (n = 40, two-sided t-test) for horizontal eye movements. k The effect was present in both contexts (two-tailed t-test), although to a weaker extent, without differences between the two. l Regression line showing the linear relationship between egocentric angles of goal position in conceptual space and leftwards vs. rightwards eye movements (n = 40). *p < 0.05; ns not significantly different.