eeg-decoding

Perform multivariate decoding on EEG epochs with cross-validated scoring and permutation significance.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/dengzhe-hou/auto-eeg-analysis --skill eeg-decoding
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: eeg-decoding
Source: https://github.com/dengzhe-hou/auto-eeg-analysis/tree/main/skills/eeg-decoding
Command: npx skills add https://github.com/dengzhe-hou/auto-eeg-analysis --skill eeg-decoding

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, scikit-learn, mne.

What problem does it solve?

This Skill turns EEG epochs into multivariate decoding results so you can answer whether (and when) information about conditions or classes is represented in the brain.

Core Features & Use Cases

  • Sliding time-point decoding: train and test at each time point to produce time-resolved decoding curves (e.g., face vs object).
  • Temporal generalization (train time × test time): reveal whether neural representations are transient, sustained, or reactivated over time.
  • Searchlight decoding & CSP decoding: localize informational content across channels (searchlight) or decode oscillatory patterns (CSP) for frequency-band focused questions.
  • Use Case: If your analysis plan claims that two conditions differ in discriminable neural patterns between 100–200 ms, use decoding to quantify peak decoding time, chance level, and statistical significance.

Quick Start

Use the eeg-decoding skill to decode a cognitive contrast by running decoding on your epoched data and producing decoding outputs for each claim in ANALYSIS_PLAN.md.

Frequently Asked Questions about eeg-decoding

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run multivariate pattern analysis on EEG epochs to test if two conditions are distinguishable?

Multivariate decoding on EEG epochs quantifies whether conditions are distinguishable over time and channels by applying cross-validated sliding time-point decoding, producing time-resolved accuracy curves and permutation-based significance.

What is temporal generalization in EEG decoding and when do I need it?

Temporal generalization trains a decoder at one time point and tests it across all others, generating a train-time × test-time matrix that reveals whether neural representations are transient, sustained, or reactivated.

Can I decode oscillatory EEG patterns using Common Spatial Patterns with MNE and scikit-learn?

Yes, CSP decoding applies Common Spatial Patterns to extract band-limited oscillatory features, integrating MNE for epoch handling and scikit-learn for cross-validated classification to decode frequency-band focused neural patterns.

How do I prevent data leakage when running spatial searchlight decoding on EEG data?

Cross-validated spatial searchlight decoding enforces no data leakage by properly partitioning training and testing sets within each spherical cluster of channels, ensuring decoding scores reflect true informational content.

Do I need preprocessed EEG epochs to perform sliding time-point decoding?

Yes, sliding time-point decoding requires preprocessed EEG epochs as input, along with an analysis plan specifying decoding claims, to produce cross-validated scores and statistical significance outputs.

What is the best way to localize informational content across EEG channels during cognitive tasks?

Searchlight decoding localizes informational content by applying a moving spherical spatial cluster across channels, quantifying whether specific scalp regions contain discriminable neural patterns for the contrasted conditions.