Fig. 6: ECoGs during the imagery task.
From: Voluntary control of semantic neural representations by imagery with conflicting visual stimulation
a Schematic of the imagery task. After the presentation of the word or landscape image (nonimagery period), the subjects imagined the word image during the presentation of the landscape image, or imagined the landscape image during the presentation of the word image (imagery period). For all combinations of images, the trials were repeated in randomized order, resulting in 50 trials. To obtain modulation of R(vinferred, vword) due to imagining the word image (ΔZword), the R(vinferred, vword) during the presentations of landscape images were Fisher z-transformed and averaged within the nonimagery period and the imagery period to calculate the difference between them (\(\overline{z(R({v}_{inferred},\,{v}_{word}))}\) during imagery period – \(\overline{z(R({v}_{inferred},\,{v}_{word}))}\) during nonimagery period). Similarly, ΔZlandscape was evaluated from the R(vinferred, vlandscape) during the presentation of word images. b Results of time–frequency decomposition for the ECoGs of the electrode indicated by the black arrow in (c). The upper and middle panels represent the time–frequency maps of ECoGs while watching the image of words (left) and landscapes (right) during the nonimagery period and imagery period, respectively. The difference between these two maps of each column is shown in the bottom panel. c Powers in the high-γ band from 0 to 1 s after the presentation of word images (left) and landscape images (right) were subtracted between the two periods (imagery period – nonimagery period) to be shown on the cortical surface at the location of each subdural electrode with color-coding.
