Fig. 4. Examining relationships between typicality, memorability, and semantic and visual content.

(A) Visualizing the correlation of DCNN-based typicality and memorability for all 1854 concepts in terms of an early layer (layer 2) and late layer (layer 20) allows for the observation of an overall positive relationship between early and late layer typicality scores across the concepts (r = 0.253, P = 2.504 × 10⁻²⁸). A chi-square analysis of the four quadrants of the scatterplot demonstrated significantly more concepts than chance showed a pattern where the most memorable items were prototypical in terms of both early and late layer features (χ² = 38.046, P = 6.909 × 10⁻¹⁰). Contrastingly, we find significantly fewer concepts that demonstrate “mixed” patterns where more memorable items demonstrated early layer prototypicality and late layer atypicality (χ² = 8.454, P = 0.004) or the opposite pattern (χ² = 20.286, P = 6.668 × 10⁻⁶). We found no significant difference from chance for concepts where the most memorable items were atypical across both early and late layer features (χ² = 8.399, P = 0.553). This suggests that, in general, memorable concepts tend to be both visually and semantically prototypical. (B) Example concepts that fell into each quadrant of the scatterplot seen in (A). Note that in this figure, THINGS database images were replaced by similar looking images from the public domain images available in THINGSplus (48).