finding the frequency components of Steady state Visual Evoked Potential signal

hello

i have a task to find the frequency of visual stimulation of SSVEP signal.

My task in detail:

In a Brain computer Interface experiment, data from a subject were recorded. Eight electrodes were attached at locations Pz,PO3,O1,Oz,O2,O9 and O10. The 10s of data were recorded and digitized with a sampling rate of Fs=2048 Hz.

1)Load the Matlab File(attached) trial02.mat and find the frequency components of the acquired EEG signals using the FFT. The attached matlab file is an array of Nt X Nch. Where Nt is the no of samples and Nch is the no of Electrodes. PLot the FFt for the electrode Oz. which Frequency was used for visual stimulation?

2)Construct a bandpass filter with cutoff frequencies of 10 and 40 Hz. pLot the bode diagram(filter Response) of this digital filter.

3)Apply the bandpass filter to the EEG signals and plot the new frequency spectrum for the electrode Oz.

how do i proceed..can someone explain?

my knowledge on this: At first the EEG signal should be filtered to allow signals between 10 and 40 Hz.then its Frequency components can be found using FFT and it can be plotted.Then every frequency has to be normalised ...

But i dunno how and where to start

Answer 1

[Star Strider]

You need to read the documentation on the fft function. (The R2015a version is to me more straightforward, so I’m linking to it here. Pay special attention to the code between the top two plot figures.) For filter design, my filter design procedure is described in How to design a lowpass filter for ocean wave data in Matlab?. For a bandpass filter, the stopband frequencies have to be ‘outside’ the passband frequencies. If you define your passband as ‘Wp’ with Nyquist frequency ‘Fn’, as a first estimate, define the stopband frequencies ‘Ws’ as:

Wp = [Wlow  Whigh]/Fn;
Ws = Wp .* [0.5  1/0.5];

You may have to tweak ‘Ws’ to get the filter characteristics you want. Be certain to use the second-order-section (SOS) implementation for stability, and always check the filter characteristics with the freqz function, using the sampling frequency ‘Fs’ as an additional argument.