Chapter 2: Importing and Inspecting Data¶
Continuing from Chapter 1 — Installation and First Launch.
With EEGLAB and ERPLAB installed, this chapter loads a recording and gets familiar with how EEGLAB represents it internally.
Importing a recording¶
File → Import data → Using EEGLAB functions and plugins offers importers for most common EEG file formats: Biosemi BDF, Neuroscan CNT, EGI raw/simple binary, generic EDF, and a generic ASCII/float importer for anything else. Pick the matching importer, point it at your file, and EEGLAB loads the recording into a new dataset.
The EEG struct¶
Everything EEGLAB knows about the current recording lives in one MATLAB structure, conventionally called EEG. If you've read the MNE-Python track, this is the same role MNE's Raw object plays — a single container for the signal and everything describing it:
EEG.data— the sample-by-channel numeric arrayEEG.chanlocs— channel names and (once set up) their 3D positionsEEG.event— the list of event markers, covered in Chapter 5EEG.srate— the sampling rate
You rarely touch EEG directly in the GUI — menu items read and write its fields for you — but recognizing it helps when reading EEGLAB documentation or error messages, and matters directly once you start scripting (see the note on EEG.history in Chapter 3).
Channel locations¶
Most menu operations that plot topographic maps need to know where each electrode physically sits. Edit → Channel locations loads a location file — a standard .ced/.elp/.locs file matching your electrode cap, or one of EEGLAB's bundled standard 10-20 layouts. Once loaded, Plot → Channel locations confirms every channel landed in a sensible spot before you rely on any scalp-map plot later.
Looking at the data¶
- Plot → Channel data (scroll) opens a scrolling multi-channel view — the everyday way to eyeball a recording for obvious problems (flat channels, huge artifacts) before doing anything else.
- Plot → Channel spectra and maps shows the power spectrum per channel plus a scalp topography at a chosen frequency, useful for spotting line noise or a dead channel that scrolling alone might miss.
Multi-subject studies, and saving your work¶
A single EEG dataset covers one recording. When a project has many subjects and conditions, File → Create study groups multiple datasets into a STUDY — EEGLAB's structure for group-level and condition-level analysis. This track stays at the single-dataset level throughout, but it's worth knowing STUDY exists once your own project grows past one file.
Whatever you've done to a dataset, File → Save current dataset as writes it out as a .set file (metadata) plus a .fdt file (the raw numeric data) — EEGLAB's native format, and the one this track's final chapter reads back into MNE-Python.