Electronic Thesis and Dissertation Repository

Thesis Format



Master of Science




Ansari, Daniel


Previous studies using fMRI adaptation to investigate the neural substrate of symbolic number processing have found ratio-dependent responses in regions of the parietal cortex, suggesting that number symbols are coded by overlapping neuronal populations: the larger the ratio between two numerals, the more their representation overlap. The current study analyzed the distributed patterns of activation associated with numerals presented during this task. I could not find substantial evidence supporting the ratio-dependent structure of the similarity space predicted by the univariate adaptation analyses. I also failed to find evidence in favor of the alternative model that similarities were driven by lexical frequency. These null results were confirmed by Bayesian analysis showing substantial support for the null. These findings do not align with the theory of ratio-dependent overlapping representation of number symbols and challenge previous interpretations of the adaptation literature.

Summary for Lay Audience

How are symbolic numerals represented in the brain? Some theories have suggested that numerals are represented in a number line according to their magnitude. Numerals that are closer in this number line are represented in the brain more similarly than numbers that are further away. In other words, the way our brain represents the numeral “3” is more similar to the representation of numeral “4” than numeral “9”. Other authors have suggested that magnitude is not as important in the association between numerals that are established in the brain. These authors have suggested that other factors like frequency are more important. Said differently, this alternative theory poses that numerals that are seen together more frequently in real life are more similar, disregarding their magnitude. In this thesis we tested those competing theories. We could not find evidence for either of them. However, because the first alternative has been highly influential in this research field, our results lead to rethink about these theories. Future studies should develop more advanced models that may determine how numerals are organized in the brain.