![]() I’ll pick out the Fourier coefficients from one timepoint, approximately 178 ms after stimulus onset, at 10 Hz. Let’s pick a single time point from the A29 electrode. We can draw a polar plot, showing the position of each of the phase angles on that circle. When we convert to phase angles, we set the magnitude to 1, placing phase on the unit circle. Instead of taking the steps above, let’s convert our Fourier coefficients to phase angles. That corresponds to the timing of a visual ERP. You should be able to pick out that at A29 (roughly, Oz), there’s a post-stimulus power peak. This data is a bit short for demoing TFRs, but it’ll do for our purposes. Ignore all the white at the edges that’s just where the edges have been removed to avoid edge effects from the convolution process. These plots show total power at the 11 electrodes in this data. ![]() plot_tfr(demo_tfr) +įacet_wrap(~electrode) # Data is complex Fourier coefficients, converting to power. plot_tfr() does that for us automatically. Fourier coefficients can be converted to power by taking their absolute values and squaring them. ![]() Let’s have a look at raw power at all of the electrodes. To be able to calculate ITC, we’re going to need the Fourier coefficients for every single trial. Let’s see this process in action on some real data. The next step is to take the mean of these values: This normalises the values to be on the unit circle. We divide the Fourier coefficients by their absolute values: Step-by-step, we begin with the Fourier coefficients derived using the FFT: With a real signal, you’ll only see 0 or 1 if you’ve done something wrong! ITC is bounded between 0 and 1, with 1 being perfect intertrial coherence (i.e. exactly the same phase on every trial), and 0 being absolutely no intertrial coherence. jv2unSVC7O- Natalie Schaworonkow December 17, 2019 It reaches its maximum value of 1 for perfectly phase-aligned signals and becomes 0 as the phase distribution becomes uniform. It is the circular sum of phases at a certain point in time (length of red arrow). Inter-trial coherence measures phase-synchronization across trials. ![]() As the phase gets more and more misaligned across trials, ITC decreases. When the oscillation reaches the same phase at the same time on each trial, we have high ITC. On the x-axis we have time, on the y-axis we have trial number. These complex numbers represent both the magnitude and the phase of the signal.Ī fantastic visualization by Natalie Schaworonkow demonstrates the principle of intertrial coherence. The FFT returns Fourier coefficients for each combination of time and frequency - complex numbers with real and imaginary components that describe a position in a two-dimensional plane. When we perform time-frequency analysis, we are performing a moving window Fast Fourier Transform. That’s where intertrial coherence comes in! Starting at 0, we peak at 90 degrees, hit zero again on the way down at 180 degrees, trough at 270 degrees, then return to 0 at 360 degrees, where we begin the cycle again.Ī question we might have when examining oscillations in EEG or MEG data is how consistently they reach the same point in the cycle across trials. In the schematic below, I express that using degrees. The phase of the wave is a measure of how far through an oscillatory cycle the wave is. So the wave cycles from 0 up to 1, down to -1, then back up to 0. A complete cycle of of the wave is the amplitude peaking, declining to a trough, then hitting zero again. So here, you can see it has a maximum of 1, indicated by the arrow. The amplitude of the wave is the distance between 0 and the wave at a given timepoint. In the schematic below, we have a sine wave of an arbitrary frequency. Intertrial coherence (ITC) is a measure of how consistent oscillatory phase is across an ensemble of trials.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |