eeg-spectral

What problem does it solve?

Spectral-domain EEG analysis is slow and error-prone when you need defensible PSD, band-power metrics, and aperiodic+periodic decomposition (e.g., specparam/FOOOF) rather than just raw FFT plots.

Core Features & Use Cases

• Power spectral density (PSD): computes resting-state (continuous) and task-related (epoched) PSD using Welch by default, with configurable frequency bounds and FFT resolution.
• Band power extraction: derives canonical (delta/theta/alpha/beta/gamma) or user-specified bands with clear output for absolute vs dB-style reporting decisions.
• Aperiodic+periodic decomposition: optionally runs specparam/FOOOF to estimate 1/f aperiodic parameters and detect periodic peaks, including goodness-of-fit reporting.
• Individual Alpha Peak Frequency (IAPF): optionally extracts IAPF per subject (from specparam peaks or PSD peak-detection) to support individualized alpha-band definitions.
• Publication-ready outputs: writes spectral plans and subject/group artifacts plus figure-stage PSD plots, topomaps, and IAPF/IAPF-related visuals.

Quick Start

Ask the agent to run /eeg-spectral projects/my-study to compute PSD and band power from the appropriate preprocessing/epoch stage, and enable specparam/FOOOF with optional IAPF extraction if requested.