Brain Encoding Models using Taskonomy

Background and Objectives

This project compares encoding models using Taskonomy. The assignment consists of two main problems, each with specific sub-tasks:

• Problem 1: RSA and stimulus dependence: Compute and analyze a Representational Similarity Analysis (RSA) matrix using Taskonomy networks and BOLD5000 stimuli data.
• Problem 2: Testing an asymmetric distance: Build encoding models for predicting fMRI responses, and analyzing symmetric and asymmetric relationships between models.

It is implemented as part of CMU 10-733 Course on Representation and Generation in Neuroscience and AI.

Repository Structure

This repository contains the following files and directories:

• notebook.ipynb: A comprehensive Jupyter notebook with all code.
• README.md: This file provides an overview of the project, its objectives, and the contents of the repository.

References

• https://www.cs.cmu.edu/~neural-taskonomy/
• http://taskonomy.stanford.edu/
• https://bold5000-dataset.github.io/website/download.html
• https://github.com/alexsax/midlevel-reps/blob/master/README.md#installing-visualpriors
• https://github.com/ariaaay/NeuralTaskonomy/blob/master/code/make_task_tree.py

---

Note: All data used in this project is sourced ethically, and the analysis adheres to the highest standards of research integrity and ethical guidelines.