Features were ranked based on the median peak score over 9 subjects. A row in the table also shows the median time point when that feature reached its peak score (medians taken over 9 subjects). All features shown were significantly predicted with p < 10-5 (corrected for multiple comparisons over features, regions, and time).
Semantic features are shown in upper case, and perceptual features in lower case.
| Rank | Feature | Median time (s) |
|---|---|---|
| 1 | DOES IT USE ELECTRICITY? | 0.55 |
| 2 | CAN IT STRETCH? | 0.10 |
| 3 | White pixel count | 0.30 |
| 4 | CAN YOU HOLD IT IN ONE HAND? | 0.05 |
| 5 | DOES IT HAVE CORNERS? | 0.65 |
| 6 | IS IT MANUFACTURED? | 0.65 |
| 7 | DOES IT ROLL? | 0.15 |