Applied Research Projects#
Project-style examples that target a concrete scientific question – age regression, p-factor prediction, sex classification, P300 transfer, clinical-catalog summary – with realistic data sizes, runtimes, and limitations. Difficulty 2-3; assumes the Start Here trio and the core workflow track.
These are not first-week tutorials. They are scaffolds for your own analyses: each script frames a research question, picks an appropriate evaluation regime, runs a defensible baseline, and surfaces the limitations honestly so you know what would have to change before any result here could be cited. Compared with tutorials, these projects emphasise labels, splits, baselines, and reporting rather than the individual EEGDash API calls.
What you will learn:
How to frame an EEG-from-population study (age, sex, p-factor) as a single regression or classification problem with an honest baseline.
How to choose between a feature pipeline and a deep model based on data size and the question being asked.
How to apply transfer learning across paradigms (P300 transfer across subjects and sessions) without leaking labels.
How to summarize a clinical catalogue (subjects, sessions, conditions, hours of recording) for inclusion in a paper.
How to write up the limitations section that an EEG paper actually needs (Cisotto and Chicco 2024; Pernet et al. 2019 for BIDS).
Treat each script as a starting point for your own work, not a prescriptive recipe.
