`spkit.eeg`.ICA_filtering¶

spkit.eeg.ICA_filtering(X, winsize=128, ICA_method='extended-infomax', kur_thr=2, corr_thr=0.8, AF_ch_index=None, F_ch_index=None, verbose=True, window=['hamming', True], hopesize=None, winMeth='custom')¶

Aftifact Removal using ICA

ICA Filtering algorithm uses following three criteria for removing artifacts for EEG.

Kurtosis based artifacts - mostly for motion artifacts:
- parameter kur_thr is used as threshold on kurtosis of ICs. Any IC above kur_thr is removed. As higher kurtosis of component, more peaky it is.
Correlation Based Index (CBI) for eye movement artifacts:
- CBI method [R3cb126c15dac-1] computed the comparaision of power in prefrontal electrodes with frontal eletrodes, in IC. A component, that stisfy the criteria, is considered as component capturing eye-blink activity and removed.
- For applying CBI method, index of prefrontal (AF - First Layer of electrodes towards frontal lobe) and frontal lobe (F - second layer of electrodes) channels needs to be provided.
- For case of 14-channels Emotiv Epoc
  
  ch_names = [‘AF3’,’F7’,’F3’,’FC5’,’T7’,’P7’,’O1’,’O2’,’P8’,’T8’,’FC6’,’F4’,’F8’,’AF4’]
  
  Pre-frontal Channels =[‘AF3’,’AF4’],
  
  Fronatal Channels = [‘F7’,’F3’,’F4’,’F8’]
  
  then suplied index of channels are as follow;
  
  AF_ch_index =[0,13]
  
  F_ch_index =[1,2,11,12]
- [1]
  https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3812628/
Correlation of any IC with many EEG channels:
- If any IC is correlated corr_thr % (80%) of elecctrodes, is considered to be artifactual
- Similar like CBI, except, not comparing fronatal and prefrontal but to all

Parameters:

X: array,(n,ch)

input signal (n,ch) with n samples and ch channels

winsize: int, default=128

window size to process, if None, entire signal is used at once

ICAMed = {‘fastICA’,’infomax’,’extended-infomax’,’picard’}

method of ICA Decomposition

kur_thr: + scalar, default=2

threshold on kurtosis of IC component, as per (1) criteria (see above)

Note

CBI Method if AF_ch_index or F_ch_index is None, CBI is not applied.

AF_ch_index: list, None

(AF - First Layer of electrodes towards frontal lobe)
example [0,13], as for [‘AF3’,’AF4’] for Emotiv 14 channel data
check above details for suplying accurate indices, as per your data

Changed in version 0.0.9.7: Default is set to None, instead of [0,13]

F_ch_index: list, None

(F - second layer of electrodes)
example [1,2,11,12], as for [‘F7’,’F3’,’F4’,’F8’] for Emotiv 14 channel data
check above details for suplying accurate indices, as per your data

Changed in version 0.0.9.7: Default is set to None, instead of [1,2,11,12]

corr_thr: float [0, 1] (deafult 0.8)

threshold to consider correlation, higher the value less IC are removed and vise-versa
if None, this is not applied

Note

Criteria (3) if corr_thr is None then Criteria (3) is not applied.

Returns:

XR: array,(n,ch)

Filtered signal, same size as X, (n,ch)

See also

ATAR: Automatic and Tuanable Artifact Removal Algorithm
ATAR_1Ch: ATAR for single channel
ATAR_mCh: ATAR for multiple channel
CBIeye: Correlation Based Inded

References

[1]

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3812628/

Examples

#sp.eeg.ICA_filtering
import numpy as np
import matplotlib.pyplot as plt
import spkit as sp
X,fs, ch_names = sp.data.eeg_sample_14ch()
X = sp.filterDC_sGolay(X, window_length=fs//3+1)
t = np.arange(X.shape[0])/fs
Xc1 = sp.eeg.ICA_filtering(X.copy(),winsize=128, ICA_method='extended-infomax',kur_thr=2,corr_thr=0.8,
                        AF_ch_index = [0,13] ,F_ch_index=[1,2,11,12])

Xc2 = sp.eeg.ICA_filtering(X.copy(),winsize=128, ICA_method='infomax',kur_thr=2,corr_thr=0.8,
                        AF_ch_index = [0,13] ,F_ch_index=[1,2,11,12])

Xc3 = sp.eeg.ICA_filtering(X.copy(),winsize=128, ICA_method='picard',kur_thr=2,corr_thr=0.8,
                        AF_ch_index = [0,13] ,F_ch_index=[1,2,11,12])

sep=200
plt.figure(figsize=(10,6))
plt.subplot(221)
plt.plot(t,X+np.arange(X.shape[1])*sep)
plt.xlim([t[0],t[-1]])
plt.yticks(np.arange(X.shape[1])*sep,ch_names)
plt.title(r'$X$: EEG')
plt.xlabel('time (s)')
plt.grid()
plt.subplot(222)
plt.plot(t,Xc1+np.arange(14)*sep)
plt.xlim([t[0],t[-1]])
plt.title(r'$X_{c1}$: with Extended-infomax')
plt.yticks(np.arange(Xc1.shape[1])*sep,ch_names)
plt.xlabel('time (s)')
plt.grid()
plt.subplot(223)
plt.plot(t,Xc2+np.arange(14)*sep)
plt.xlim([t[0],t[-1]])
plt.title(r'$X_{c2}$: with Infomax')
plt.yticks(np.arange(Xc2.shape[1])*sep,ch_names)
plt.xlabel('time (s)')
plt.grid()
plt.subplot(224)
plt.plot(t,Xc3+np.arange(14)*sep)
plt.xlim([t[0],t[-1]])
plt.title(r'$X_{c3}$: with Picard')
plt.yticks(np.arange(Xc3.shape[1])*sep,ch_names)
plt.xlabel('time (s)')
plt.grid()
plt.suptitle(r'ICA Based Artifact Removal')
plt.tight_layout()
plt.show()