Emotion-induced blindness effect from erotic images is not caused by distinctive low-level image features

Poster Presentation: Saturday, May 17, 2025, 2:45 – 6:45 pm, Pavilion
Session: Attention: Reward

Jeff Saunders1, Ho Ming Chan1; 1University of Hong Kong

Erotic images produce strong temporal attention effects, which have been interpreted as emotion-induced blindness (EIB). However, the attentional effects of erotic images could alternatively be due to distinctive image properties. We distinguished these possibilities by comparing the effects of erotic images and neutral images with closely matched image properties. We used a convolutional neural network to generate pairs of erotic and non-erotic neutral images that have the same distributions of medium-level features. The erotic images showed nude people in a sexual context, while the non-erotic images depicted common objects without recognizably erotic content but with closely matched image features (e.g., 'nude chair'). We tested whether these two types of images produced similar distraction. We also varied the colors of the background images to test whether color distinctiveness affects initial capture of attention. Subjects performed an RSVP task with distractor images presented before target images with lag of 2 or 8. The distractor was either an erotic image, a matched neutral image, or a background image. Background images were presented with original colors (unmatched background) or were recolored to have the same color histogram as the distractor (matched background). We found that the erotic images produced a much larger decrease in target detection at lag 2 than the matched neutral images, and the effect was the same regardless of whether the background stream was color matched. The matched neutral distractors had a smaller but detectable effect when background images were landscapes, and no effect in a follow-up experiment that used background images from the same categories. Our results demonstrate that erotic images cause a strong EIB effect even when image properties are carefully controlled, and suggest that distinctive image features have little effect. We conclude that the attentional effects of erotic images are primarily due to the erotic content.