Learn EEG and brain-computer interfaces
Plain explanations of how the signals work, the BCI paradigms, and every method between a raw recording and a result you can trust, plus how to choose software, find data and connect hardware.
The basics
What a brain-computer interface is, how EEG and fNIRS measure the brain, and where to start.
- What Is a Brain-Computer Interface (BCI)?A brain-computer interface (BCI) turns measured brain activity into commands for a computer or device. How BCIs work, invasive vs non-invasive, the main paradigms, and real uses.Read
- How EEG Works: What an EEG MeasuresHow EEG works: electrodes on the scalp measure tiny voltages from synchronised cortical neurons. What EEG can and cannot measure, the 10-20 system, artifacts, and how the signal is analysed.Read
- EEG Frequency Bands: Delta, Theta, Alpha, Beta, GammaEEG frequency bands explained: delta, theta, alpha, mu, beta and gamma brain waves, their ranges, where they appear, what they are associated with, and how band power is measured.Read
- What Is fNIRS? Functional Near-Infrared SpectroscopyfNIRS (functional near-infrared spectroscopy) measures brain activity through changes in blood oxygenation using light. How it works, HbO and HbR, fNIRS vs EEG and fMRI, and fNIRS for BCI.Read
- BCI for Beginners: How to Get Started with EEG and BCIHow to get started with brain-computer interfaces: what to learn first, a first project using a public EEG dataset, the tools you need, and the mistakes that make beginner results wrong.Read
BCI paradigms
The three classic ways a BCI turns brain activity into a command.
- Motor Imagery BCI: EEG Classification ExplainedMotor imagery BCI explained: how imagined movement changes mu and beta rhythms, ERD/ERS, the C3/C4 electrodes, CSP and EEGNet decoders, public datasets, and honest cross-subject evaluation.Read
- P300 BCI and the P300 Speller ExplainedP300 BCI explained: the oddball paradigm, the P300 speller, why trials are averaged, Xdawn spatial filtering, Riemannian ERP covariances and CNN decoders, and how to measure speller performance.Read
- SSVEP BCI: Steady-State Visual Evoked PotentialsSSVEP BCI explained: how flickering stimuli drive occipital EEG at the flicker frequency and harmonics, stimulus design, CCA, FBCCA and TRCA detection, trained decoders, and information transfer rate.Read
- Multimodal and Hybrid BCI: EEG + fNIRS + EMGHybrid and multimodal BCI: combining EEG with fNIRS or EMG, early vs late fusion, testing whether fusion actually helps, and aligning stimuli such as images, audio and video to the recording.Read
Methods
Preprocessing, artifacts, features, decoders and evaluation, one method per page.
- EEG Preprocessing Pipeline: Steps in OrderAn EEG preprocessing pipeline step by step: filtering, line-noise removal, bad channel detection, re-referencing, ICA, epoching, baseline correction and trial rejection, and where the order matters.Read
- EEG Artifact Removal: Eye Blinks, Muscle and ICAHow to remove EEG artifacts: eye blinks, eye movements, muscle, heartbeat and line noise. ICA component review, bad channel detection and trial rejection explained, with what each method costs.Read
- EEG Signal Processing SoftwareEEG signal processing software: bandpass and notch filtering, resampling, re-referencing, ICA, normalization and band-power extraction, with every parameter recorded alongside the stage it produced.Read
- EEG Feature Extraction: Which Features to UseEEG feature extraction explained: time, frequency, spatial and connectivity features, which suit which paradigm, and how to rank them without overfitting.Read
- EEG Time-Frequency Analysis: ERSP and ITCEEG time-frequency analysis explained: Morlet wavelets, STFT and multitaper, ERSP and inter-trial coherence (ITC), baseline normalisation in dB, and cluster-based permutation testing for significance.Read
- EEG Connectivity Analysis: Coherence, PLV, PLI, wPLIEEG connectivity explained: coherence, imaginary coherence, PLV, PLI, wPLI, Granger causality and graph metrics, and the volume conduction problem.Read
- EEG Data Augmentation: Methods and PitfallsEEG data augmentation for small datasets: noise, shifts, channel dropout, mixup, SMOTE and GAN or diffusion synthesis, without leaking test trials.Read
- Common Spatial Patterns (CSP) for EEGCommon Spatial Patterns (CSP) for EEG explained: variance-ratio spatial filters, log-variance features, regularized and filter-bank CSP, and the leakage trap.Read
- Riemannian Geometry for BCI: MDM and Tangent SpaceRiemannian geometry for BCI: covariance matrices, MDM, tangent space classifiers, ERP covariances for P300, and Euclidean alignment across subjects.Read
- EEGNet: The Compact CNN for EEG ExplainedEEGNet explained: the compact CNN for EEG, its temporal, depthwise and separable convolutions, why it suits small datasets, and how it compares with other models.Read
- EEG Machine LearningTrain EEG classifiers on a node canvas: CSP, Riemannian tangent space, SVM, random forest and deep architectures, with grouped splits and nested cross-validation that keep the reported number honest.Read
- Deep Learning for BCI and EEGDeep learning for BCI: EEGNet, ShallowConvNet, DeepConvNet, ATCNet, EEGConformer, LSTM, TCN and EEG transformers, trained on your own GPU from a node canvas with no training script to write.Read
- Cross-Validation for EEG: LOSO, Grouped and NestedCross-validation for EEG and BCI: why random trial splits inflate accuracy, grouped and leave-one-subject-out (LOSO) splits, nested cross-validation for tuning, time-series splits, and data leakage.Read
- BCI Domain AdaptationCross-subject and cross-session BCI domain adaptation: CORAL, MMD alignment, transfer component analysis, DANN, feature alignment and subject adaptation as blocks — free on every plan.Read
- fNIRS Machine LearningfNIRS machine learning and classification: build a chromophore-aware design matrix, train classical or deep models on a node canvas, and evaluate with grouped splits sized for small fNIRS cohorts.Read
Software, data and hardware
Choosing tools, opening files, finding datasets, and getting a model onto a device.
- EEG Analysis Tools ComparedA practical comparison of EEG analysis tools — EEGLAB, FieldTrip, MNE-Python, Brainstorm and BCILattice — by what they are actually good at, and how to choose between them.Read
- BCI Software and Tools ComparedBCI software compared: OpenViBE, BCI2000, MNE-Python, pyRiemann, MOABB, Braindecode and BCILattice, grouped by what each tool is actually for.Read
- Free EEG Analysis SoftwareFree EEG analysis software options: EEGLAB, MNE-Python, FieldTrip, Brainstorm and the BCILattice free plan. What each costs you in setup and skills, and what BCILattice's free plan includes.Read
- No-Code EEG Analysis and Machine LearningEEG analysis and machine learning without code: visual preprocessing, ICA, feature extraction, model training and honest evaluation, with Python optional.Read
- EEGLAB Alternative for EEG Decoding and MLAn EEGLAB alternative for decoding: no MATLAB, machine and deep learning built in, leakage-safe evaluation, and EEGLAB .set files import directly.Read
- MNE-Python Alternative with a Graphical InterfaceAn MNE-Python alternative with a GUI: BCILattice is built on the same Python stack and runs EEG preprocessing, decoding and evaluation without scripting.Read
- EEG File Formats: EDF, BDF, GDF, FIF, VHDR, SET, SNIRFEEG and fNIRS file formats explained: EDF and EDF+, BDF, GDF, FIF, BrainVision VHDR, EEGLAB SET, Neuroscan CNT, SNIRF and CSV. What each stores, which systems write them, and how to open them.Read
- Public EEG and BCI Datasets: Where to Find ThemWhere to find public EEG and BCI datasets: motor imagery, P300, SSVEP and fNIRS datasets, BCI Competition data, PhysioNet, OpenNeuro, MOABB, and the BCINexus dataset library with paradigm metadata.Read
- EEG Headset Software: Analyse Recordings from Your DeviceAnalyse recordings from your EEG headset: OpenBCI, BrainVision, BioSemi, Neuroscan and consumer devices via EDF, BDF, VHDR, CNT or CSV, plus LSL (beta).Read
- Real-Time BCI: From Offline Decoder to Online UseReal-time BCI explained: why offline accuracy overstates live use, pseudo-online evaluation, false alarms, latency, bit rate, and Lab Streaming Layer.Read
BCILattice and BCINexus
The product pages: what the desktop app and the publishing platform do for each kind of study.
- BCI Research PlatformA BCI research platform covering the whole study: import EEG, fNIRS and EMG recordings, preprocess, extract features, train decoders locally, then publish for review and citation.Read
- BCI Analysis SoftwareBCI analysis software with eleven reviewable steps: ICA review, channel quality, epoching, normalization, time-frequency, channel selection, dimensionality reduction and 71 feature families.Read
- EEG Analysis SoftwareEEG analysis software for researchers: ICA artifact review, channel quality scoring, epoching, topographic activation maps, ERSP/ITC time-frequency and named feature extraction — offline, on your machine.Read
- fNIRS Analysis SoftwarefNIRS analysis software that treats HbO and HbR as separate physiology: SNIRF import, Beer-Lambert conversion, short-channel regression, chromophore-aware steps and haemodynamic band presets.Read
- EMG Analysis SoftwareEMG analysis software with the standard activation measures — RMS, mean absolute value, waveform length, slope sign changes and Willison amplitude — plus EMG band presets, envelope extraction and classification.Read
- Reproducible BCI ResearchMake BCI research reproducible: every processing stage keeps the parameters it ran with, evaluation guardrails catch leaky splits, and a published study carries the pipeline rather than describing it.Read
- Neuroscience Research WorkflowOne workflow for a neuroscience lab: import and manage recordings, run analysis and training in a shared study structure, review results together, and publish without re-assembling anything.Read