continuous complex gaussian wavelet transform  (MathWorks Inc)

Journal: bioRxiv

Article Title: Fluctuations of fMRI activation patterns reveal theta-band dynamics of visual object priming

doi: 10.1101/148635

Figure Lengend Snippet: (A) Slow trend and surrogate data of one participant. The slow trend is the slow trend of congruent condition in the FFA, there were 18 different slow trends; the surrogate data were generated by adding a Gaussian curve peaked at 400 ms and white noise (different for each participant) to the slow trend for each participant. Thus, 18 sets of surrogate data were generated. Subsequent analyses of these surrogate data are identical to how we analyzed the real data. (B) Left: Averaged surrogate data ( n =18, mean ± SEM), smoothed (60 ms bin) as a function of mask-to-probe SOA (200-780 ms in steps of 20 ms). Middle: Slow trends averaged across participants. Right: Average smoothed-and-detrended data, extracted by subtracting slow trends shown in Middle from smoothed (60 ms bin) data shown in left (thick lines). (C) Average spectrum for detrended data (extracted by subtracting slow trends from the surrogate data without smoothing). The statistical threshold of significance ( p < 0.05, multiple comparison corrected) calculated by performing a permutation test was shown with a dashed line.

Article Snippet: To assess MVPA classification accuracies as a function of time (mask-to-probe SOA) and frequency, the detrended temporal profile for each condition was transformed using the continuous complex Gaussian wavelet (order = 4; e.g., FWHM =1.32 s for 1 Hz wavelet) transforms (Wavelet toolbox, MATLAB), with frequencies ranging from 1 to 25 Hz in steps of 2 Hz.

Techniques: Generated, Comparison