This toolbox adapts Representational Similarity Analysis (RSA) for MEG data, addressing fMRI-based limitations in temporal resolution. Focused on implementing peer-reviewed, high-impact methods from MEG research, it prioritizes temporal dynamics and source localization accuracy.
utils.similarity_metrics.py: Multiple similarity metrics with citable references (accessible via help()).utils.transformers.py: Tools for constructing model RDMs from pretrained LLMs.
I will add some tutorial soon.
All implementations derive from rigorously validated studies—no novel methods proposed. Use judiciously; scientific/technical issues may be reported via GitHub issues. Contributions to error fixes are prioritized.