MNE-Python: Building a Bidirectional Event System for Evoked and Source Space Interactivity

MNE-Python provides powerful 3D PyVista brain viewers and 2D Matplotlib sensor plots, but these tools currently operate in isolation. This project bridges that gap by extending the mne.viz.ui_events registry to allow bidirectional interactivity between source-space and sensor-space visualizations. I will implement a decoupled event system where selecting a 2D time-point updates the 3D topography, and picking a 3D vertex or ROI (Region of Interest) instantly generates the corresponding 2D evoked time-course. This solves a major UX bottleneck for researchers validating inverse solutions and interpreting neural models. Key Deliverables: - A bidirectional event-callback architecture integrated into mne.viz. - Synchronized UI updates for mne.viz.Brain and mne.Evoked.plot_* (topomaps and joint plots). - A prototype "Interactive Dipole" tool for real-time forward modeling (Xfit replacement). - Comprehensive pytest fixtures for the new event classes and a sphinx-gallery tutorial demonstrating the interactive workflow.