Fig. 5

Latent relationship between the superiority illusion and the Ponzo illusion. A Weighted total feature importance values (the products of machine learning feature importance and PCA loadings) between the models were comparable in PC1 but dissociable in PC2 and PC3. The results indicate that codes related to SI were overlapping yet dissociable between the Ponzo illusion magnitude (d’) and metacognitive performance (log M-ratio). B Generic (machine learning irrelevant) relationships between the three PCA scores and d’ (top row) and between the three PCA scores and log M-ratio (bottom row). Transparent dots represent individual data points. Transparent lines represent linear regression fit using ordinary least squares