Fig. 2: Results of representational similarity searchlight analysis.

Analyses (n = 17 subjects) tested for the effects of sample, operation, operand, and result predictors on brain activity pattern dissimilarity in a multiple regression approach. Analyses were initially performed for each subject per time bin/acquired volume during a prolonged delay period where subjects computed the result of an operation and subsequently kept it in memory, with effect size estimates (β) then being averaged across early, middle and late parts of the delay period (right) before performing surface-based group analyses in FreeSurfer (two-tailed one-sample t-tests, correction for multiple comparisons by permutation at cluster level, p_FWE < 0.05, cluster forming threshold p < 0.01). The color scale represents uncorrected −log(p) values within the clusters surviving correction. Results are projected onto a flattened cortical average surface (https://mri.sdsu.edu/sereno/csurf/), with superimposed lines representing the regional borders of the Human Connectome Project (HCP-MMP1) parcellation³⁰. For better orientation, the location of major anatomical landmarks (MFG middle frontal gyrus, IFG inferior frontal gyrus, Cent S central sulcus, IPS intraparietal sulcus, STS superior temporal sulcus, OTS occipitotemporal sulcus, Calc S calcarine sulcus) is indicated. Source data are provided as a Source Data file.