Frites

Frites analyzes neurophysiological data using information-theoretical measures to extract cognitive brain networks and perform robust group-level statistical inference on MEG and intracranial recordings.

Key Features:

• Information-Theoretical Metrics: Implements non-parametric permutation-based analyses of non-negative information measures derived from information theory, machine learning, and distance metrics.
• Statistical Framework: Supports fixed- and random-effect group models with test- and cluster-wise corrections for multiple comparisons.
• Multi-Level Inferences: Performs analyses at multiple scales including local brain regions, inter-areal functional connectivity, and network properties to extract task- and behavior-related effects across populations.
• Numerical Simulations: Uses numerical simulations to compare the accuracy of ground-truth retrieval across different group models.

Scientific Applications:

• Cognitive brain-network extraction: Extracts task- and behavior-related cognitive brain networks from neurophysiological recordings.
• Group-level analysis of MEG and intracranial recordings: Performs group-level statistical inference on MEG data and non-uniform intracranial recordings.
• Functional connectivity and network analysis: Evaluates inter-areal functional connectivity and network-level properties using information-theoretical measures.

Methodology:

Numerical simulations compare ground-truth retrieval accuracy across different group models; analyses employ non-parametric permutation testing, fixed- and random-effect models, and test- and cluster-wise multiple comparisons corrections.