Figure 9
From: Multivariate EEG analyses support high-resolution tracking of feature-based attentional selection
Reconstructing the neural signature of attentional capture for target positions that were never presented during Experiment 2. (A) The target positions that are reconstructed: top, bottom, left and right. (B) The constructed channel responses that are associated with these positions using the CTF from Fig. 8A (see methods for details). Top is in between target position 8 and 1, so the channel response to top is constructed by averaging the hypothetical channel response to position 8 and position 1. Similarly, right is created from averaging channel responses to 2 and 3 etc. Any position on the circle can be constructed using a weighted average of channel responses. Left, right, bottom and top weights were reconstructed using the product of the constructed channel responses and the weight matrix at 260–270 ms. Next, the left versus right pattern was generated by subtracting the left from the right pattern (C), likewise the top versus bottom pattern was created by subtracting the bottom pattern from the top pattern (D). Note the similarity with the left-right and top-bottom patterns from Experiment 1. All patterns were normalized across space to allow direct comparisons. Thick electrode dots belong to a cluster that has p < 0.01 under cluster based permutation testing.
