Please enable JavaScript.

Coggle requires JavaScript to display documents.

Midterm Prep, Lecture 2, Lecture 3, Lecture 4, Lecture 5 - Coggle Diagram

- - - - ~50-1500 uV RMS
        
        Peak-to-Peak often 0.5-10 mV, depending on contraction
    - - Amplitude rises w/ contraction
        
        Not linearly at extremes
    - - Burst-like during dynamic tasks
      - Sustained during isometrics
    - - iEMG resolves single Motor Unit Action Potentials (MUAPs)
      - sEMG reflects composite activity
  - - - MUAPs
    - - ~ 50 uV - 5 mV
    - - 20 - 500 Hz
        
        Surface
      - Up to 10 kHz
        
        Needle
  - - - Ag/AgCl gel electrodes
    - - Needle + fine wire
  - - - f = 0-20 Hz
      - Amplitude > 5 mV
      - From cable & electrode-skin movement
- - - - Passband needs to be 5-15 Hz
        
        To maximize QRS energy
    - - Pan-Tompkins Method
  - - - <0.003 Hz
    - - 0.0033-0.04 Hz
    - - 0.04-0.15 Hz
    - - 0.15-0.40 Hz
    - - Balance index
- - - - Low frequency
        
        0.05-1.0 Hz
        
        Needs HPF
    - - 0.1-10 Hz
    - - 20-1000 Hz
      - Peak Energy 30-300 Hz
      - Can mask P waves
  - - - f = 5-30 Hz
      - Peak = 3-5 Hz
    - - f = 8-50 Hz
      - Peak = 10-40 Hz
    - - f = 0-10 Hz
      - Peak = 1-2 Hz
    - - f = 150 - 250 Hz
- - - - Brain activation during specific tasks
    - - Spontaneous brain activity patterns
    - - Online processing for neurofeedback
- - - - The signals are independent of one another
    - - At most one subcomponent is Gaussian
        
        Others have non-gaussian distribution
      - Gaussian = Symmetrical, bell shaped curve
  - - - Mixtures of independent signals tend toward gaussian distributions
        
        So we seek non-Gaussian projections to find sources
    - - If signals extracted from mixtures are independent & non-Gaussian, they must be source signals
- - - - i.e. for L=3, Level 1 is from 0 to 0.33, Level 2 is from 0.33 to 0.67, Level 3 is from 0.67 to 1
    - - Usually m=2, which would be counting the occurances that like 1 -> 1 or 1 -> 3 happen
    - - H = -sum(p x log_2(p))
- - - - PC1 would capture the max variance
        
        PC2 would capture the next largest variance orthogonal (perpendicular) to PC1
        
        PCs are uncorrelated & ordered by variance (PC1 > PC2)