Reproduce “Spatially Masked Regression Reveals Local and Distributed Predictability in Electrophysiological Recordings” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Maryam Ostadsharif Memar, Nima Dehghani

This study developed the Spatially Masked Regression (SMR) framework to quantify the balance between local and distributed predictive information in electrophysiological recordings. The framework revealed that while local activity significantly predicts a channel's signal, a substantial portion is also predictable from broader, distributed networks, even when immediate neighbors are excluded, and demonstrated modality-specific differences in predictability and transferability between EEG and iEEG.

Key insights

  • Electrophysiological signals exhibit hierarchical spatial dependencies, where local neighbors provide concentrated predictive information, but optimal reconstruction also requires integrating broader, non-local inputs.
  • EEG signals display higher reconstructability and cross-subject transfer compared to iEEG due to greater spatial mixing and standardized montages, indicating EEG operates in a more spatially redundant regime.
  • The sustained predictability of a target signal despite strict local masking suggests that individual channels reflect a distributed dynamical system, indicating genuine distributed predictive content beyond biophysical local redundancy.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.