Subjects had to press and hold the down-arrow key to begin the movie presentation,

which started 100 ms after the key press, and were instructed to release the key at the moment they recognized the second person. Movies involving the same stimulus pair in either direction were never shown in consecutive mTOR inhibitor trials. Finally, according to the subjects’ responses, for each pair we selected three morphed images (M1, M2, and M3) giving an ambiguous perception: M2 was the one estimated to give the most ambiguous perception to the subject—i.e., the image that corresponded to the mean response time, averaging the presentations going from A to B with the ones going from B to A; M1 and M3 were closer to pictures A and B, respectively, and were between three to eight frames away from M2 (the exact number of frames was heuristically selected in each case to give an ambiguous image but with a slight bias toward recognition of one or the other image). The morphed pictures were created using the software SmartMorph, after rescaling and cropping the images with Photoshop. Images were presented at the center of the laptop screen and covered about 1.5° of visual

angle. After determining the morphs eliciting an ambiguous perception, subjects performed the adaptation paradigm, in which the perception of the ambiguous images was biased by first showing Bumetanide one of this website the two original pictures used to generate the morphs (Figure 1A). The basic idea is that, when shown a morphing between pictures A and B, subjects are more likely to recognize it as picture B if the morphed image is preceded by a presentation of picture A (the adaptor) and vice versa (Webster et al., 2004 and Leopold et al., 2005). This effect has been attributed to diminished responses of feature-selective neurons after previous stimulation by

the adaptor (Leopold et al., 2005). In the first eight sessions, the adaptor image (either picture A or B of each pair) was shown for 1 and 1.5 s (first six and following two sessions, respectively), but a better perceptual bias was later obtained when using a longer presentation (4 s) of the adaptors (Figure 1C), which was used in the remaining 13 sessions. For each picture pair, a total of six to eight presentations of each morph (M1, M2, and M3) preceded by an adaptation to picture A, and an equal number of times preceded by an adaptation to B, were shown in pseudorandom order. Each trial started with a fixation cross displayed at the center of the screen for 500 ms. After a random delay between 0 and 100 ms, the adaptor picture was presented (for 1, 1.5, or 4 s) and, following a blank of 300 ms, one of the morphed images was shown for 200 ms.