BIOENG-310 — Week 7 Graded Exercise

Aligning Computational Models to Inferior Temporal (IT) Cortex Activity

This repository contains a Jupyter notebook developed for
BIOENG-310: Neuroscience Foundations for Engineers at EPFL.

The exercise focuses on analyzing neural population recordings from the inferior temporal (IT) cortex of macaque monkeys and aligning computational vision models to neural activity using Representational Dissimilarity Matrices (RDMs).

---

Overview

The notebook addresses the following components:

• Data visualization

  • Visualization of stimulus images
  • Visualization of stimulus-averaged IT response time series
  • Visualization of time-averaged neural responses (70–170 ms post-stimulus)

• Neural decoding

  • Decoding object category from IT population activity using logistic regression
  • Comparison of decoding performance against theoretical chance level

• Model–brain alignment (RDM)

  • Alignment of Gabor filter representations to IT activity
  • Alignment of AlexNet convolutional layer activations to IT activity

---

Dependencies

• Python
• Jupyter
• NumPy
• xarray
• Matplotlib
• OpenCV (opencv-python)
• scikit-learn
• brainscore-vision
• brainio
• tqdm