Classification done honestly

You are skimming: the title, the first figure, and the short version. Switch to Read in the header for the full page, or Deep to open every deep dive.

Features

A classifier wants a short vector of numbers per trial. For EMG and oscillatory EEG: band powers or log-variances per channel, sometimes after a spatial filter. For ERPs: the downsampled voltage waveform, channels concatenated. For SSVEP: correlations with reference frequencies. Good features make simple classifiers work; no classifier rescues bad features. The most reliable way to improve a BCI is still to improve the signal.

Three classifiers that cover the field

Linear discriminant analysisLinear discriminant analysis (LDA)A simple, robust classifier that draws a straight boundary between classes; the workhorse of EEG decoding for two decades. Glossary entry fits a Gaussian to each class with a shared covariance and draws the boundary where they are equally likely. Few parameters, hard to overfit, fast, and the default in BCI for two decades. With shrinkage regularization it handles more features than trials, which is the usual EEG situation.

Common spatial patternsCommon spatial patterns (CSP)Spatial filters that maximize the variance of one class while minimizing the other, the classic feature extractor for motor imagery. Glossary entry is a feature extractor for two-class oscillatory problems: it finds spatial filters whose output variance is large for one class and small for the other. The log-variance of a few CSP components, then LDA, is the classic motor imagery pipeline.

Riemannian geometryRiemannian geometry classifiersMethods that treat each trial's covariance matrix as a point on a curved space and classify by distance there; the strongest simple baseline for EEG today. Glossary entry classifiers skip feature engineering: each trial is represented by its channel covariance matrix, which lives on a curved space where the natural distance is not Euclidean. Map the matrices to the tangent space at their mean and run logistic regression. This is the strongest simple baseline on nearly every public EEG dataset and it is three lines in pyRiemann. Learn it after LDA, and use it by default.

from pyriemann.estimation import Covariances
from pyriemann.tangentspace import TangentSpace
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline
clf = make_pipeline(Covariances('oas'), TangentSpace(), LogisticRegression())

Support vector machines, random forests, and neural networks exist too. On EEG-sized datasets they rarely beat the three above by more than noise, and they overfit more. The deep learning page is about when they do and do not help.

Now the part that matters

Leakage. Any information from the test set that reaches training is leakageData leakageInformation from the test set sneaking into training, for example by shuffling trials from one session across both; it inflates accuracy and is the most common flaw in student BCI papers. Glossary entry, and it inflates the number. Shuffling trials across sessions. Fitting the CSP filters, the normalization, or the ICA on all the data before splitting. Choosing hyperparameters on the test set. Choosing which subjects to include after seeing results. Each of these is common, each is easy to do by accident, and scikit-learn’s pipelines exist to prevent the second one: put every fitted step inside the pipeline so cross-validation fits it on training folds only.

The right splits, in increasing honesty. Shuffled trials within a session: the wrong way, useful only to check the code runs. Grouped by run or block within a session: honest within-session. Train on session one, test on session two: the cross-session number, which is what a user experiences. Leave one subject out: the cross-subject number, which is what a calibration-free product would need. Report the one that matches the claim you are making, and say which.

Chance is not 50 percent. With two classes and forty trials, a coin-flip classifier scores above 62 percent one time in twenty. The binomial distribution gives the upper bound for any trial count; the statistics page has it. Always report the chance bound next to the accuracy. With unbalanced classes, use balanced accuracy or report the confusion matrix, because 80 percent on a task with 80 percent standards is chance.

Calibration and drift. Signals change within a session (electrodes settle, attention fades) and between sessions (cap position, mood, sleep). A deployed BCI recalibrates at the start of each session and often adapts during it. If your evaluation does not include the cost of that, it is not evaluating the system.

Information transfer rate

Accuracy alone misleads: 90 percent on two targets in ten seconds is worse than 70 percent on eight targets in three. Wolpaw’s information transfer rate combines them. Bits per selection:

B = log₂N + P·log₂P + (1−P)·log₂((1−P)/(N−1))

for N targets and accuracy P, and ITR is B times selections per minute. A good SSVEP speller reaches 40 to 60 bits per minute; a P300 speller 10 to 25; motor imagery under 10. Report it, and report the time per selection it assumes.