Skip to main content
. 2012 Nov;50(13):3107–3121. doi: 10.1016/j.neuropsychologia.2012.07.007

Fig. 5.

Fig. 5

Decoding scene representations in the medial temporal lobe (Bonnici et al., 2012). (A) The two original scenes are displayed at the top, along with full set of morph stimuli below (note the labels A and B were never used in the actual experiment). (B) Timeline of a single trial comprised a stimulus duration of 2.5 s during which the participant registered their decision. Participants then indicated their confidence in that decision during the next 3 s, from a choice of sure, fairly sure, and very sure. There was then a 2 s rest before the start of the next trial. (C) Group data showing the choice behaviour for each of the morph stimuli, which clearly followed a sigmoid profile. (D) Average classification accuracy values for the original (100%) scenes displayed for the hippocampus (HC), entorhinal cortex (EC) and posterior parahippocampal gyrus (PHG). For all three regions classification accuracy was significantly above chance, although the HC classifier performed significantly better than the EC and PHG classifiers (*p<0.05). (E) Average classifier accuracy values for the 50% morphed scenes (i.e., perceptually ambiguous). Classifier performance was significantly above chance in all three regions, with the HC and PHG classifiers both out-performing that of the EC (*p<0.05).