Judging orientation does not elicit the Spatial-Numerical Association of Response Codes: Evidence against the Neural Overlap Hypothesis

According to the neural overlap hypothesis (Fias et al., 2001) magnitude information is automatically extracted from digits when there is a sufficient degree of neural overlap between structures dedicated to the processing of relevant and irrelevant information. Specifically, since numbers are processed in parietal areas, processing a visual feature with high parietal involvement (orientation) should interact with number magnitude, while this interaction should be limited when processing features with minimal parietal involvement (colour). This seems to be confirmed by the fact that previous studies detected a SNARC effect with orientation judgment but very limited evidence was found with colour judgment. Although the neural overlap hypothesis has been highly influential, only a few studies attempted at replicating the original findings. In the present study, we aimed at replicating a key experiment introduced by Fias and colleagues (line orientation judgment), testing both digits and letters. Based on the neural overlap hypothesis, a SNARC effect should be found only with digits while it should be weak or absent with letters, due to different degrees of neural overlap with parietal areas. In two experiments, by testing participants both online and in the lab, we consistently failed to detect a SNARC effect for both numbers and letters. Aggregated analysis on N = 155 participants confirmed the absence of the effect. Our evidence cast doubts on the neural overlap hypothesis and suggest that a small SNARC effect might be detected in non-semantic tasks only by testing very large samples, as recently demonstrated with colour judgment.