PSD and calculate RMS of PSD over EEG signal

💬 Support & Discussions
,
1

Hi everyone,
I want to calculate the average of PSD in a determined epoch, here is the code:

spectrum = epochs[62].compute_psd(method="welch",picks='C3_M2',fmin=12,fmax=30)
PSD, freqs = spectrum.get_data(return_freqs=True)
spectrum.plot()

The result would be like following:

Capture
Capture988×313 14.2 KB

As I want to take the average first I need to retrieve the Amplitude which is in (dB), like the first point which is about 123, but after I convert the PSD to dB, I have a weird data:

10 * np.log10(PSD)

which I got

array([[[ 3.13982779, -3.86176929,  6.85332206,  4.17123814,
          4.93127581,  7.4832183 ,  1.12038761,  7.68291897,
          6.40367422,  6.24721421,  6.57093037,  6.4076177 ,
          5.45691297, -1.51206827,  6.76020133,  6.69414559,
          9.51026022,  9.10557663,  5.63438618,  3.24274636,
         -5.68817689,  2.77028109,  7.11039496,  8.68978814,
          9.45466904,  5.62792485,  4.2303355 ,  4.14692724,
          3.25922632,  4.33403963,  5.39546468,  0.92007031,
          3.73590139,  5.64315007,  4.32840387,  4.41999972,
          0.74850308,  1.92540129,  5.95650245,  4.2219239 ,
          4.86983096,  7.17304789,  6.39538512,  5.62660172,
          3.48318393,  0.52069617]]])

But the element doesn’t match the figure which would be data around 116 and 128. What formulate should I use to get the y-axis values?
Also could you please help me with how can I calculate the subsurface of the PSD plot, somewhere I read I should calculate the RMS of PSD? Is this correct, and how can I calculate that?

Thanks in advance