Automatic Artifact Rejection with Dynamic Reliability

 

Step # 4 - Use the Automatic Artifact procedure to select “Good” & “Reliable” artifact free and representative samples of EEG as a template for quantitative analysis.

 

4a- For automatic template selections Click Edit > Automatic Selection.   We will use the 1st 2 seconds of selected EEG as a template of “Good” EEG, this is only for illustration and to note that the first 2 seconds is generally not a good period to select from.   Press the left mouse button and drag it across the first 2 seconds of EEG.  De-select using the right mouse button.  A “good” EEG sample  is split-half reliability ≥ .95, test re-test reliability > 0.90 and Edit Time  ≥  60 seconds.  

 

 

4b-  The Default multiplier is 1.0.   Click O.K. and the FFT and Z scores of the edited selections will be displayed.   Cut Off Multiplier = 1.0 is a RMS amplitude match for each 2 seconds of EEG that are equal to or less than the RMS amplitude of the user selected  EEG template.

4c - Edit time is now 21 seconds and Reliability is improved.   Visually Re-scan the EEG to de-select segments that may have artifact and to select Good and representative EEG segments that may have been omitted.

 

 

 

4d-  To change the template selection of “Good” EEG, highlight “Clear Automatic Selections” and then repeat the automatic selection process by clicking Edit > Automatic Selections.  To change the Cut-Off highlight “Clear Automatic Selections” and repeat steps 4a and 4b.

 

 

The Cut Off Multiplier determines the amplitude of match to the “Good EEG” template.  When the cut off multiplier = 1.25 then the match will be 25% larger peak-to-peak amplitude than the template amplitude, if it is 1.5 then the match will be 50% larger amplitude than the template.  The Cut Off Window is the delay between automatic selections.   A value of 0.25 = 250 millisecond minimum gap between automatic selections and a value of 0.5 = 500 millisecond minimum gap between automatic selections, etc.

 

4e- Click O.K. using the 1.25 Multiplier Cut-Off and see a larger sample of EEG of 2:22 minutes and Reliability Increased.    Note that the EEG Spectrum and Z Scores did not change much.   Scroll through the EEG Selections as an Expert would and look for Epileptic discharges and other possible abnormal features of the EEG that may have been excluded.  Change to 1.5, 1.75, etc. to test reliability and then click Edit .  Save your selections by clicking Edit > Save Edit File, e.g., Template-nonartifact 30 seconds or Template-nonartifact 2min or Template-All EEG without any artifact rejection.    Compare these three artifact free files to the compare the Z scores and the Dynamic FFT to different selections in order to understand the influence of your selections may or may not have upon the Z scores and FFT .  Finally, click Edit > Select All (the worse case scenario in which artifact is not excluded).  Conclusion: High Theta (5 Hz – 7 Hz) and Low Beta (15 – 25 Hz) are pervasive and representative of the artifact free EEG.

 

 

 

Press the left mouse button and scan the Z scores while observing the Z score values at the left margin of the EEG tracings for the different frequencies of the spectrum