Abstract
According to the neural overlap hypothesis (Fias et al., 2001), numerical magnitude may influence performance in a non-semantic task when the processing of task-relevant and task-irrelevant information recruits overlapping neural structures. Specifically, because number and orientation processing both engage parietal regions, the account predicts a SNARC effect during orientation judgments, whereas effects should be weaker in tasks with less parietal involvement, such as colour judgments. However, only a few studies have attempted to replicate the original findings. Here, we conducted a large-sample replication of the line orientation task introduced by Fias et al. (2001, Experiment 4) and included letters as a complementary condition. Across two experiments, conducted online and in the laboratory, we consistently failed to detect a SNARC effect for numbers or a SNARC-like effect for letters. A cumulative analysis likewise provided no evidence for either association. These findings fail to replicate a key result supporting the neural overlap hypothesis and suggest that shared parietal processing may not give rise to spatial-numerical associations as robustly as previously assumed.