#32 · Tools & infrastructure

Open tools and estimator corrections target analysis fidelity in EEG, PET, fMRI

The common thread is the construction and correction of analysis pipelines for human neuroimaging, where analytic choices directly shape reported effects. MLE-Toolbox packages end-to-end EEG and MEG workflows in an open-source MATLAB GUI, spanning preprocessing, source localization, connectivity, phase-amplitude coupling, and machine learning. BrainPET Studio offers an open-source desktop GUI for atlas-based regional PET quantification in MNI space that avoids FreeSurfer and is validated against ADNI. The IQ-Motion preprint is a different kind of contribution: it applies an errors-in-variables estimator to ABIDE-I and suggests that pooled OLS overstates the IQ-motion slope in autism fMRI confound correction, framing the inflation as potentially driven by site-correlated measurement uncertainty. What is new across the set is explicit attention to estimator bias and accessible, reproducible tooling, rather than a shared modality or mechanistic finding, since the items differ in modality and aim.