Skip to content

nmningmei/modification-pipelines

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

141 Commits
.ipynb_checkpoints
.ipynb_checkpoints
 
 
__pycache__
__pycache__
 
 
annotations
annotations
 
 
figure I need
figure I need
 
 
Add probability to pipeline - single subject try.py
Add probability to pipeline - single subject try.py
 
 
Ellmore_data_project.py
Ellmore_data_project.py
 
 
Fast spindle detection.py
Fast spindle detection.py
 
 
Filter_based_and_thresholding.py
Filter_based_and_thresholding.py
 
 
Generate+EEG+Graphs.py
Generate+EEG+Graphs.py
 
 
Neural_Networks.ipynb
Neural_Networks.ipynb
 
 
NeurosynthIntro.ipynb
NeurosynthIntro.ipynb
 
 
Plot annotations.ipynb
Plot annotations.ipynb
 
 
README.md
README.md
 
 
Slow spindle detection.py
Slow spindle detection.py
 
 
auto running.py
auto running.py
 
 
compare auto spindles of day 1 and day 2.py
compare auto spindles of day 1 and day 2.py
 
 
compute transition matrix.ipynb
compute transition matrix.ipynb
 
 
compute transition matrix.py
compute transition matrix.py
 
 
confusion matrix individual machinelearning.png
confusion matrix individual machinelearning.png
 
 
confusion matrix individual thresholding (high).png
confusion matrix individual thresholding (high).png
 
 
confusion matrix individual thresholding.png
confusion matrix individual thresholding.png
 
 
correlation cross channel.ipynb
correlation cross channel.ipynb
 
 
cross validation - 10 - visualization.ipynb
cross validation - 10 - visualization.ipynb
 
 
deep learning in the FA and miss - fail.ipynb
deep learning in the FA and miss - fail.ipynb
 
 
eegPinelineDesign (adowaconan).py
eegPinelineDesign (adowaconan).py
 
 
eegPinelineDesign.py
eegPinelineDesign.py
 
 
eeg_compute_pipeline.py
eeg_compute_pipeline.py
 
 
example of spindle and non spindles.py
example of spindle and non spindles.py
 
 
example.py
example.py
 
 
fast spindle joint plot.png
fast spindle joint plot.png
 
 
fast spindle robust regression fit.png
fast spindle robust regression fit.png
 
 
fast spindles.png
fast spindles.png
 
 
firstdraftspindlepaper.docx
firstdraftspindlepaper.docx
 
 
for another reviewer.py
for another reviewer.py
 
 
from 6 channel baseline to 32 channels to tfc 2.py
from 6 channel baseline to 32 channels to tfc 2.py
 
 
from 6 channel baseline to 32 channels to tfc 3.py
from 6 channel baseline to 32 channels to tfc 3.py
 
 
from 6 channel baseline to 32 channels to tfc.py
from 6 channel baseline to 32 channels to tfc.py
 
 
gathering spindle and non spindle instance.py
gathering spindle and non spindle instance.py
 
 
gathering spindle instance (time frequency convol).py
gathering spindle instance (time frequency convol).py
 
 
gathering spindle instance.py
gathering spindle instance.py
 
 
get power spectral statistics.py
get power spectral statistics.py
 
 
global ASI.csv
global ASI.csv
 
 
individual performance (2 models) (high).png
individual performance (2 models) (high).png
 
 
individual performance (3 models) (high).png
individual performance (3 models) (high).png
 
 
many measures(X3).png
many measures(X3).png
 
 
many measures.png
many measures.png
 
 
model comparions.p
model comparions.p
 
 
modification pipeline - artifact rejection - adapt MNE updates.ipynb
modification pipeline - artifact rejection - adapt MNE updates.ipynb
 
 
modification pipeline 1.ipynb
modification pipeline 1.ipynb
 
 
modification pipeline 10 - threshold + compare to spindle anno + sleep stages.ipynb
modification pipeline 10 - threshold + compare to spindle anno + sleep stages.ipynb
 
 
modification pipeline 11 - sleep stages.ipynb
modification pipeline 11 - sleep stages.ipynb
 
 
modification pipeline 12 - single try part 2 - parallel run.ipynb
modification pipeline 12 - single try part 2 - parallel run.ipynb
 
 
modification pipeline 12 - single try.ipynb
modification pipeline 12 - single try.ipynb
 
 
modification pipeline 13 - threshold + compare to spindle annotation + sleep stages.ipynb
modification pipeline 13 - threshold + compare to spindle annotation + sleep stages.ipynb
 
 
modification pipeline 14-15 - spindle detection and training.ipynb
modification pipeline 14-15 - spindle detection and training.ipynb
 
 
modification pipeline 16 - functionalize comparison.ipynb
modification pipeline 16 - functionalize comparison.ipynb
 
 
modification pipeline 17 - sampling spindles and non spindles.ipynb
modification pipeline 17 - sampling spindles and non spindles.ipynb
 
 
modification pipeline 17 - sampling spindles and non spindles.py
modification pipeline 17 - sampling spindles and non spindles.py
 
 
modification pipeline 17 - single try.py
modification pipeline 17 - single try.py
 
 
modification pipeline 18 - sampling and training.ipynb
modification pipeline 18 - sampling and training.ipynb
 
 
modification pipeline 19 - single try.py
modification pipeline 19 - single try.py
 
 
modification pipeline 19.ipynb
modification pipeline 19.ipynb
 
 
modification pipeline 2.ipynb
modification pipeline 2.ipynb
 
 
modification pipeline 20 - apply classifier to one subject.ipynb
modification pipeline 20 - apply classifier to one subject.ipynb
 
 
modification pipeline 21 - 3 criteria.ipynb
modification pipeline 21 - 3 criteria.ipynb
 
 
modification pipeline 22- single try.py
modification pipeline 22- single try.py
 
 
modification pipeline 23 - RMS detecting spindles.ipynb
modification pipeline 23 - RMS detecting spindles.ipynb
 
 
modification pipeline 23 - RMS detecting spindles.pdf
modification pipeline 23 - RMS detecting spindles.pdf
 
 
modification pipeline 3.ipynb
modification pipeline 3.ipynb
 
 
modification pipeline 4.ipynb
modification pipeline 4.ipynb
 
 
modification pipeline 5.ipynb
modification pipeline 5.ipynb
 
 
modification pipeline 6.ipynb
modification pipeline 6.ipynb
 
 
modification pipeline 7.ipynb
modification pipeline 7.ipynb
 
 
modification pipeline 8.ipynb
modification pipeline 8.ipynb
 
 
modification pipeline 9.ipynb
modification pipeline 9.ipynb
 
 
neural network using only numpy - version 1.ipynb
neural network using only numpy - version 1.ipynb
 
 
neural network using only numpy - version 2 - adding momentum.ipynb
neural network using only numpy - version 2 - adding momentum.ipynb
 
 
new data (adowaconan).py
new data (adowaconan).py
 
 
new data comparison (high resolution).png
new data comparison (high resolution).png
 
 
new data.py
new data.py
 
 
optimization results.py
optimization results.py
 
 
osf_test.py
osf_test.py
 
 
permutation test used in the current study.py
permutation test used in the current study.py
 
 
plot manual annotation and compare auto annotations.py
plot manual annotation and compare auto annotations.py
 
 
power plot with fast spindles.py
power plot with fast spindles.py
 
 
power plot with slow spindles.py
power plot with slow spindles.py
 
 
preprocessing - remove artifacts by ICA.ipynb
preprocessing - remove artifacts by ICA.ipynb
 
 
respond to reviewer.py
respond to reviewer.py
 
 
run to have spindle results.py
run to have spindle results.py
 
 
sleep annotation.p
sleep annotation.p
 
 
slow spindle joint plot.png
slow spindle joint plot.png
 
 
slow spindle robust regression fit.png
slow spindle robust regression fit.png
 
 
slow spindles.png
slow spindles.png
 
 
slow waves.py
slow waves.py
 
 
spindle detection.py
spindle detection.py
 
 
spindle example and comparison on power spectral.py
spindle example and comparison on power spectral.py
 
 
spindle result_fast.csv
spindle result_fast.csv
 
 
spindle result_slow.csv
spindle result_slow.csv
 
 
ssp preprocessing for sleep data.py
ssp preprocessing for sleep data.py
 
 
stats of annotations.py
stats of annotations.py
 
 

Repository files navigation

modification-pipelines

progress of how the pipeline changes and how logistic regression can be used to detect spindles in EEG channel data

Pipelines for analyzing EEG data and detect spindls, sleeping stages, and k-complex.

Entering the phase using threshold detection algorithm to detect spindles in 6 EEG channels.

  1. Threshold detection algorithm
  2. ASI capturing alpha,beta,spindle activity in terms of power density

ICA parameters:

  1. lowpass 200 Hz (however, literature suggests also highpass at 1 Hz to advoid overfitting ICA)
  2. apply notch filter at 60 Hz
  3. MNE ICA: iterration = 3000, fixed random state
  4. artifact dectection is based on channles "LOC" and "ROC", and run automatically
  5. rejection parameters: EEG: 80 - 160 depending on subjects; tstep: 2 seconds; EOG criteria: 0.4; skewness: 2; kurt: 2; variance:2
  6. bandpass 0.1-50 Hz

Detecting spindles:

  1. bandpass slow/fast spindle range (10-12Hz/12-14Hz) (Begmann et al., 2012)
  2. select channels: F3, F4, C3, C4, O1, O2
  3. use a moving window to compute root-mean-square (RMS): Gaussian window, standard deviation = windown length / .68 / 2, window length = 2000 samples, convolution using central part of convolution of the same size
  4. compute the harmonic mean of the RMSs of the 6 channels and call it the mean channel
  5. compute RMS for the mean channel
  6. compute trimmed mean and trimmed standard deviation (5%) on the data after the first 100 seconds and before the last 30 seconds for both the individual channels and the mean channel
  7. lower_threshold = mean + lower_threshold * standard deviation (Begmann et al., 2012)
  8. higher_threshold = mean + higher_threshold * standard deviation
  9. post threshold parameter: segments that is above the threshold and duration of the segments is in between 0.5 - 2 secs
  10. determining spindles: find spindles in AT LEAST (>=) 3 channels AND find spindle in average channel at the similar time stamp (deviate < 1 second)

Power spetral density analysis

  1. delta 1: 0-2 Hz
  2. delta 2: 2-4 Hz
  3. theta: 4-8 Hz
  4. alpha: 8-12 Hz
  5. beta: 12-20 Hz
  6. slow spindle: 10-12 Hz
  7. fast spindle: 12.5-14.5 Hz

compute power spectral density of all these frequency bands using a moving window with lenght of 10 seconds and overlapping half of the window size. power spectral density is rescaled by 10*log10

Assumptions:

  1. alpha drop == awake to stage 1 sleep
  2. alpha drop and beta drop == stage 1 to stage 2 sleep
  3. alpha drop and beta drop, and slow/fast spindle increase == generating spindles

To capture my assumption: compute ASI

About

No description or website provided.

Topics

machine-learning eeg sleep

Resources

Readme
Activity

Stars

0 stars

Watchers

2 watching

Forks

2 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages