Here’s where I showcase the different parts of my thesis

Previous Simulation Analysis/Visualization Code - R

This code represents my initial attempt to build a simulation experiment. I broke down the code into two sections, the first focusing on the Type-I error rate and the second focusing on the Type-II error rate/Power.

Previous Simulation Results - Tableau

Initially I developed a simulation experiment that used average ERP waveforms as the population dataset.

Visualizing the Experimental Methodology - R

Following some feedback I went back to the drawing board to develop a new simulation methodology that would use a population dataset containing single trials.

Simulation Code - Matlab

I then started to focus on building the simulation code in Matlab that will be used to generate the results of the experiment.

At this point in time I’m testing whether the simulation code works properly.

Analysis/Visualization Code - R

In “simulation procedures v3.0 & v3.1”, I was loading thousands of csv files into R to generate the visualizations and analysis.

Notes: In “simulation procedure v4.0 & v4.1”, I merged all csv files into one large csv file using the command prompt: copy *.csv merged.csv. When loaded into R, this merged.csv file is ~ 96 million rows. The population data contains ~ 1800 single trials. It might be worthwhile to reduce the population data size to ~ 1500 single trials to see how the distributions change.

  • Simulation Procedure v4.0
    • ~1800 single trials
    • 20 participants
    • adding a standardized effect size (e.g., 0.5 microvolts for each 0.1 increase in effect size)
    • stimulus-locked averaging methodology data only
  • Simulation Procedure v4.1
    • ~1800 single trials
    • 20 participants
    • adding a standardized effect size generated using the population data’s standard deviation
    • stimulus-locked averaging methodology data only

Notes: In “simulation procedure v5.0”, I’m going to add a few tables to understand the anova distributions. I may try to visualize this data in Tableau.

Simulation Results - Tableau

After running the most recent simulation procedure (v4.0), I visualized the distirbutions of the P3 amplitudes and p-values.

Notes: In the P3 Distribution with Overlapping Conditions dashboard I noticed that I was generating a control condition for each of the effect sizes when really I only needed one control condition, that being, the one generated using an effect size of 0. To fix this issue I simply deleted the other control conditions being produced by effect sizes 0.1 and up.

Simulation Code - Matlab

After completing the simulation code to produce the stimulus-locked ERP data and testing the whether the results made sense generally, I moved on to complete the simulation code to produce the RIDE ERP data.