Semantic Wavelet-Induced Frequency Tagging

The Semantic wavelet-induced frequency-tagging (SWIFT) method was designed to selectively tag high-level visual areas (while constantly activating low-level visual areas).

The method for creating the SWIFT sequences is described in detail elsewhere (Koenig-Robert & VanRullen, 2013; Koenig-Robert et al., 2015). In brief, sequences are created by cyclic wavelet scrambling in the wavelets 3D space, allowing to scramble contours while conserving local low-level attributes such as luminance, contrast and spatial frequency. First, wavelet transforms are applied based on the discrete Meyer wavelet and 6 decomposition levels. At each location and scale, the local contour is represented by a 3D vector. Vectors pointing at different directions but of the same length as the original vector represent differently oriented versions of the same local image contour. Two such additional vectors were randomly selected in order to define a circular path (maintaining vector length along the path). The cyclic wavelet-scrambling is then performed by rotating each original vector along the circular path. The inverse wavelet transform is then used to obtain the image sequences in the pixel domain. By construction, the original unscrambled image appeared once in each cycle. The original image is identifiable briefly around the peak of the embedded image, as has been demonstrated psychophysically (Roger Koenig-Robert et al., 2015).