Bridging Complexity in Neural Dynamics

This repository contains resources and instructions to be used for the tutorial of the CNSP 2025 workshop presented by H. Weissbart (on the 2nd of September at 6.30pm IST).

⚠️ Tutorial files TBC ... in construction

Objectives

This tutorial presents advanced methodologies to extend Temporal Response Function (TRF) analysis beyond its usual linear framework, deepening our understanding of complex neural interactions. Traditionally, TRF derives “evoked potential”-like responses from continuous recordings with arbitrary stimulus signals, accommodating diverse event and predictor structures. We briefly review fundamental usage, including model equation solutions and regularisation strategies—such as incorporating prior spectral information in singular value decomposition-based TRF solutions. Building on this, we propose using non-linear mappings to synthesise more complex signals, empowering linear convolutional models to extract precise temporal responses to sophisticated neural features, such as phase clustering and phase-amplitude coupling. We also introduce a technique that modulates samples using ongoing neural phase or other stimulus-bound data, weighting samples entering the TRF decomposition, and enabling hierarchical representations of stimulus dependencies.

Content

There are both data and notebooks accessible here, either directly or via simple terminal commands/programs. The repository is organised as follows:

• data: data used in the tutorial, further instructions are detailed in data/readme.md
• notebooks: the heart of the tutorial
• utils: a collection of modules and scripts that are sometimes used in the tutorial notebooks

Setup

Mamba/Conda documentation

For those new to package management in Python, Conda is an open-source platform that simplifies package installation and environment management. For a comprehensive beginner’s guide on how to get started with Conda, explore the official Conda documentation.

If you are looking for a faster alternative to Conda, Mamba is an excellent choice that provides quick environment management and package installations. You can learn how to efficiently use Mamba by checking out the Mamba documentation.

Install env or requirements

We will be using python >3.10*, to keep a modern (maintained) version. Note that the tutorial itself was written/tested using Python 3.12.10.

You can start with a fresh environment if you like, using either requirements.txt via pip or the environment.yaml via conda or mamba:

• Using pip:

  # If installing in a new environment:
  mamba create -n myenv python>=3.10 pip 
  mamba activate myenv
  # You can go straight to `pip install` if installing in an existing env:
  pip install -r requirements.txt

• Using mamba (or conda, simply replace mamba by conda in the command below):

  mamba create -n myenv -f environment.yaml