INTENSE Module
INTENSE: Information-Theoretic Evaluation of Neuronal Selectivity
A framework for analyzing neuronal selectivity to behavioral and environmental
variables using information theory, particularly mutual information analysis
with rigorous statistical testing. Cell-cell pairwise analysis
(compute_cell_cell_significance) produces adjacency matrices suitable
for Network-based spectral and topological analysis.
Information-Theoretic Evaluation of Neuronal Selectivity (INTENSE) provides tools for analyzing how individual neurons encode behavioral and task variables using mutual information.
Module Components
Quick Links
- Main Analysis Pipelines
INTENSE Pipelines - High-level functions for significance testing
compute_cell_feat_significance()- Neuron-feature analysiscompute_feat_feat_significance()- Feature-feature dependenciescompute_cell_cell_significance()- Neuron-neuron connectivitycompute_embedding_selectivity()- Embedding dimension selectivity
- Statistical Tools
INTENSE Statistics - Distribution fitting, testing, and p-value computation
Multiple Comparison Correction - Multiple comparison corrections
- Visualization
INTENSE Visualization - Selectivity heatmaps, summaries, and neuron-feature pair plots
- Advanced Analysis
INTENSE Disentanglement - Mixed selectivity disentanglement and feature correlation analysis
- Delay Optimization
Delay Optimization - Temporal delay optimization between time series
- Input Validation
Input Validation - Time series and parameter validation
- FFT Infrastructure
FFT Dispatch and Caching - FFT type dispatch and MI caching
- Multiple Comparison Correction
Multiple Comparison Correction - P-value threshold calculation (Holm, FDR, Bonferroni)
- Core Implementation
INTENSE Core Implementation - Low-level MI computation functions