abstract
- Inverted and contrast-reversed faces are identified less accurately and less rapidly than normal, upright faces. The effects of inversion and contrast-reversal may reflect different sampling strategies and/or different levels of internal noise. To test these alternative hypotheses, we used a combination of noise-masking and response-consistency techniques to measure the internal noise and high-noise efficiency associated with the identification of upright, inverted, and contrast-reversed faces. We found that both face inversion and contrast-reversal reduced efficiency, but did not change internal noise.