EEG extraction

[PhD-GOAT]

In this example (https://github.com/raphaelvallat/yasa/blob/master/notebooks/01_spindles_detection.ipynb), you can see the plotting using the txt file. But in Yasa, I wonder if there is a function to extract EEG every 30 seconds or if there is something to extract from mne.

data_N3_no-spindles_30sec_100Hz.txt

[umair-hassan]

MNE Raw data also works with YASA spindle_detect or sw_detect methods. For example see here:

yasa.spindles_detect(data, sf, ch_names=chan, multi_only=True, remove_outliers=True)

Here data is MNE raw date object
chan="C3" for example (depending on channel label in your imported raw data, similarly there are other flags).

I am posting an example code below from my notebooks:

raws_c3 = raw.pick_channels(['C3'])
sp = yasa.spindles_detect(raws_c3, 200,ch_names="C3")
sp.summary()
sp.plot_average(ci=None, palette=['tab:grey']);
sp.summary()

[PhD-GOAT]

import mne
import yasa
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from yasa import sleep_statistics

raw = mne.io.read_raw_edf('C:/Users/SC4001E0-PSG.edf', preload=True)
print(raw)
print(raw.ch_names)

data = raw.get_data()

raw_np = raw.pick_channels(['EEG Fpz-Cz'])
sp = yasa.spindles_detect(raw_np, 100, ch_names="EEG Fpz-Cz")
sp.summary()
sp.plot_average(ci=None, palette=['tab:grey'])
sp.summary()

I tried raw.get_data, but there is an error that says 0 cannot be divided.

[umair-hassan]

The you will have to pick the channel of your intetrest using raw data and give it to spindle_detection function in the same way as the text file was getting it. You could perhaps attach one sample data set on drive or here directly and I may give it a try ?

Best, Umair

[PhD-GOAT]

GitHub is not allowed to upload the edf file. SC4001E0-PSG. I'm going to link the edf. (https://physionet.org/content/sleep-edfx/1.0.0/sleep-cassette/#files-panel

[umair-hassan]

Was the issue solved? Just checking in before I jump on it again. Thanks!

[PhD-GOAT]

No, yasa says Spectral Analysis works, but it doesn't seem to be able to visualize or anything else. Although it is possible to read the raw data, edf, it is thought that there is a problem in the algorithm. If want to use it for Yasa, have to handle the edf file, which makes no sense.