Project Milestones

Note

The purpose of the milestones is to help you monitor your progress throughout the semester.

For each milestone, each group is supposed to prepare 1-2 slides (no more than specified, please!) and upload them on ILIAS until the deadline (10:00 Monday before the respective session). See the milestone page on ILIAS for a template slide.

Attending the Wednesday meetings for milestones is mandatory. During the meeting, group members will present the group’s progress, and we can discuss open questions.

Depending on your project, you may choose to prioritize certain course videos. The videos recommended for each milestone are listed below.

Important

Check out the schedule for the actual dates.

Goals to be reached:

  • You selected a paper and a dataset to replicate
  • You extracted the central hypotheses of the paper

Slides: Prepare 1 slide showing:

  • Teamname + Members
  • The general idea of the experiment
  • The main research question
  • The main analysis you will perform (ERP, Time-Frequency, Source-space, others)
Note

πŸ‘‰ Research questions can often be answered by a β€œyes” or β€œno”.

Goals to be reached:

  • Map out the required analysis pipeline of the authors
  • Compare this to the analysis pipeline you want to do (it should not be identical!)
  • Downloaded + Loaded the data. Show a first impression of the continuous data of one subject.

Slides: Prepare 1 slide showing:

  • The central research question (one sentence, very brief)
  • A table contrasting the authors and your pipeline
  • A plot of the continuous EEG data with events marked
Note

❗ Not all authors do a good job to clearly communicate how their pipeline looked like. Add notes and ask, when you are unsure!

πŸ‘‰ It is totally fine to adapt your processing pipeline as you go & experience problems / successes and get feedback.

Lectures needed to be watched for this milestone (for everyone):

  • 02 An overview of Event Related Potentials (even if you’re not doing ERP for your final analysis, this is important for understanding and preprocessing)
  • 03 Preprocessing 1: Filter in time and space
  • 04 Preprocessing 2: Data Cleaning
  • 05 Preprocessing 3: Blind source separation

Lectures recommended depending on the analysis used by your paper:

  • ERP -> 02 An overview of Event Related Potentials
  • Linear regression/Encoding -> 06 Analysis 1a: Multiple Linear Regression; 07 Analysis 1b: Encoding Models; 08 Analysis 1c: Multiple Comparison Correction
  • Time frequency analysis -> 09 Analysis 2: Time-Frequency
  • Decoding -> 10 Analysis 3: Decoding

Goals to be reached:

  • Implemented a preprocessing pipeline
  • One single subject is analyzed

Slides: Prepare 1 slide showing:

  • The table with the authors and your improved pipeline - highlight changes.
  • First results of one subject (e.g. an ERP, TRF or similar)
Note

Lectures needed depending on the analysis used by your paper:

  • ERP -> 02 An overview of Event Related Potentials
  • Linear regression/Encoding -> 06 Analysis 1a: Multiple Linear Regression; 07 Analysis 1b: Encoding Models; 08 Analysis 1c: Multiple Comparison Correction
  • Time frequency analysis -> 09 Analysis Time-Frequency
  • Decoding -> 10 Analysis 3: Decoding

Goals to be reached:

  • Your pipeline worked for all subjects
  • Further outliers / problems identified

Slides: Provide 1-2 slides with results of your sanity checks.

Sanity checks could be (depending on project, come up with your own!)

  • ICA topographies
  • ERPs including baseline
  • Comparisons to author results

Goals to be reached:

  • You should by now have replicated (or not!) the results of the authors
  • You now think of an additional analysis (e.g. ERP, decoding, linear model, time-frequency, source-space) to run on this dataset.

The analysis you propose here along with the feedback you receive will guide your further work on the project (beyond answering your main research question) between milestone 5 and the time of handing in your project.

Slides: Prepare 2 slides listing:

  • (Slide 1) the current state of replication vs. not
  • (Slide 2) your ideas on what to look at next
Note

If you haven’t, watch all the other videos on analysis methods for inspirations!