neural-decoding-analysis

To prevent information leakage in fMRI MVPA decoding, implement bias-free cross-validation strategies by strictly separating training and test data at the subject level. This skill provides decision logic and parameter guidelines to structure your pipeline and avoid circular analysis.

Designing a temporal generalization matrix for EEG involves training classifiers on specific time windows and testing across others. This skill provides methodological guardrails to correctly interpret these matrices and apply appropriate statistical thresholds for MEG and EEG modalities.

Representational similarity analysis (RSA) compares neural similarity matrices rather than training explicit classifiers like standard MVPA. Use this skill's decision trees to choose between RSA, MVPA, and encoding models based on your systems neuroscience research questions.

Yes, the methodological guidance supports research planning for multivariate neural decoding and RSA across fMRI, EEG, and MEG modalities. It provides parameter recommendations and statistical inference rules tailored to the specific neuroimaging modality you are using.

Use permutation testing for neural decoding statistical inference to rigorously validate classifier accuracy without assuming data normality. This skill offers expert guidelines on setting statistical thresholds and adhering to established neuroimaging reporting standards for robust inference.