EEG Signal Processing Guide
Electroencephalography (EEG) measures brain-related electrical activity from electrodes placed on the scalp. Unlike ECG, EEG usually does not produce a single repeating waveform. The useful patterns are often easier to interpret in the frequency domain.
A polished EEG workflow should help users move from raw multi-channel recordings to band power, Welch PSD, spectrograms and exportable results without forcing them to set up MATLAB, Python or EEGLAB for a quick review.
Why EEG Analysis Starts in the Frequency Domain
Raw EEG can look random in the time domain because brain activity, muscle artifact, eye movement and electrical noise are all mixed together. Frequency analysis separates the recording into interpretable bands so you can compare slow-wave activity, alpha rhythm, beta activity and higher-frequency components.
Understanding EEG Frequency Bands
EEG band definitions vary slightly between labs, but the following ranges are widely used for exploratory analysis and education:
- Delta, 0.5-4 Hz: slow activity often associated with deep sleep and high-amplitude slow waves.
- Theta, 4-8 Hz: activity often reviewed during drowsiness, light sleep, meditation or memory-related tasks.
- Alpha, 8-13 Hz: a prominent rhythm often seen in posterior channels during relaxed eyes-closed conditions.
- Beta, 13-30 Hz: faster activity commonly reviewed during attention, cognitive load or motor-related tasks.
- Gamma, 30+ Hz: high-frequency activity that requires careful artifact review because it can overlap with muscle noise.
Welch PSD vs. Raw FFT
A single Fast Fourier Transform (FFT) can produce a noisy, high-variance spectrum from an EEG segment. Welch Power Spectral Density (PSD) improves readability by splitting the data into overlapping windows, applying a window function, calculating spectra and averaging the results.
That averaged PSD makes it easier to compare alpha, beta, theta and delta power across channels or conditions. BioSignal Workstation uses this approach in the spectral analysis workflow so EEG band estimates are smoother and easier to interpret.
Visualizing EEG with Spectrograms
A PSD plot summarizes average frequency power across a selected recording. A spectrogram adds time, showing when specific rhythms appear, fade or shift during an experiment.
For example, an eyes-open to eyes-closed recording may show stronger activity near 10 Hz after the eyes close. A spectrogram makes that transition visible as a time-localized alpha change rather than a single average number.
How to Analyze EEG Data Online
BioSignal Workstation gives students, researchers and builders a fast way to inspect EEG exports before moving into heavier analysis environments.
- Upload the EEG file: import CSV, TXT, XLS or XLSX data into the BioSignal Workstation analyzer.
- Apply the EEG preset: start with a high-pass filter for slow drift and a 50/60 Hz notch filter for powerline noise.
- Open spectral analysis: view Welch PSD and compare standard EEG frequency bands.
- Generate a spectrogram: inspect how frequency power changes during task blocks, rest periods or artifacts.
- Export results: save filtered data and reports for documentation or follow-up analysis.
EEG Analyzer FAQ
Can I calculate EEG band power online?
Yes. The EEG analyzer can estimate band power from supported files using browser-based spectral analysis and Welch PSD.
Which EEG bands should I review first?
Start with delta, theta, alpha and beta bands for most educational or exploratory workflows. Gamma can be useful, but it needs stronger artifact review because high-frequency muscle activity can contaminate it.
Can I use this instead of EEGLAB or MNE?
Use it for fast inspection, filtering, PSD, spectrograms and reports. For advanced source localization, event-related pipelines or publication-grade statistics, EEGLAB, MNE or custom scripts may still be appropriate.