New graduate students often arrive at an EEG project without knowing where to start or how many skills the work requires. This is the path I recommend.
- Work through the Luck textbook or ERP Bootcamp to learn the basic EEG vocabulary.
- Learn enough Python to read files, manipulate DataFrames, call functions, make plots, and understand error messages.
- Try MNE-Python with real or example data. Inspecting recordings gives you a reason to keep learning Python.
- Build a mock analysis for one subject from your project. Understand each step before you scale it up.
- Learn Git before the analysis starts to grow.
- Learn Bash if you need command-line tools, an HPC system, or an automated pipeline.
- Generalize the analysis across subjects once the one-subject version is understandable and repeatable.
- Keep revising the workflow as you learn.
This takes months. A novice does not learn Python or EEG analysis in a week.
EEG fundamentals
Start with one of these:
- An Introduction to the Event-Related Potential Technique, often called the Luck textbook
- My usual recommendation for someone new to EEG.
- Budget about 20 to 30 focused hours for its roughly 370 pages. Stop to take notes and look up unfamiliar terms.
- ERP Bootcamp
- A more structured course with videos and quizzes.
- Budget about 15 hours.
Either route should give you enough vocabulary to discuss recordings, event markers, epochs, referencing, filtering, artifacts, ICA, experimental design, quality control, and subject-level analysis. You do not need to memorize every ERP component before opening a data file.
Technical skills
Programming skill comes from practice. A university course gives students about 11 weeks to learn one language in a formal setting. Set your expectations accordingly.
- Python is my first recommendation. Learn to work with files, functions, DataFrames, plots, packages, and errors. Those skills transfer directly into MNE-Python and most scientific Python tools.
- Bash becomes necessary when you automate jobs or work on an HPC system. Learn paths, pipes, scripts, environment variables, and how to read logs.
- R is worth learning when your lab already relies on it for statistics or reports. Otherwise, I recommend Python first.
- MATLAB remains common in older EEG workflows, so you may need to read it. I no longer recommend it as a first language. Its licensing creates problems when students leave institutional access, and it is a poor default for work meant to be reproduced or redistributed.
Software and workflow tools
Learn these while building the one-subject analysis. Reading about them without a real dataset will only get you so far.
- BIDS gives EEG projects shared rules for file names, folders, metadata, events, participants, and derivatives. Learn it before your project invents an undocumented folder layout. Budget 3 to 6 hours for the basics.
- Git tracks changes to scripts and configuration. Learn commits, branches, remotes, and how to compare versions. Budget 4 to 8 hours for the basics, followed by regular use until they stick.
- MNE-Python reads, preprocesses, plots, and analyzes EEG data in Python. Budget 10 to 20 hours for its beginner material. A complete study-specific pipeline will take much longer.
- EEGLAB is useful when a lab or an older project already depends on it. Learn enough to inspect and reproduce those workflows. I would not choose it as the default for a new project.