Skip to main content
. 2019 Apr 15;9:6033. doi: 10.1038/s41598-019-42429-x

Figure 4.

Figure 4

Temporal generalization. (A,B) Generalization of image decoding across time for high and low memorable images. The trained SVM classifier on MEG data at a given time point t (training time) was tested on data at all other time points t’ (testing time). The resulting decoding matrices were averaged within high or low memorable scene images and over all subjects. White contour indicates significant decoding values. Statistical tests used a cluster-size permutation procedure using cluster defining threshold P < 0.05, and corrected significance level P < 0.05 (n = 15).