EEE482 Computational Neuroscience

The assignment and project implementations for EEE482 course, Bilkent University.

Coverage of the repository

Homework 1

• Solving a system of linear equations.
• Inspecting reverse statistical inference, a case study of Broca's are and engaging in language related cognitive activities.
• Implemented in Python (Jupyter Notebook)

Homework 2

• Calculation and Inference of Spike Triggered Average (STA) Images.
• Simulation of Lateral Geniculate Nuclei (LGN) and V1 Simple Cell Neuron receptive fields using Gaussian and Gabor filters.
• Implemented in Python (Jupyter Notebook)

Homework 3

• Building Ridge and Linear Regression models to predict responses of a BOLD (Blood Oxygen Level Dependent) neural population.
• Using Hypothesis Testing and Confidence Intervals to determine the relationship between neural populations and their corresponding responses.
• Implemented in Python (Jupyter Notebook)

Homework 4

• Applying Principal Component Analysis (PCA), Independent Component Analysis (ICA) and Non-negative Matric Factorization (NNMF) to represent face images and comparison of reconstruction quality.
• Applying probabilistic and non-probabilistic population decoding techniques (Winner-Takes-All (WTA), Maximum A Posteriori (MAP) and Maximum Likelihood (MLE)) given Gaussian-shaped tuning curves.
• Implemented in Python (Jupyter Notebook)

Project:

• @emredonmez98 @alpacino98
• Using Haxby dataset for MVPA (Multi-Voxel Pattern Analysis) to classify objects given fMRI data of subjects.
• Used techniques are
  • kNN
  • SVM
  • Naive Bayes
  • MLP
  • Logistic Regression
  • Random Forest and
  • Adaboost
  • Also worked on CNNs and 3D CNNs, but the results couldn't be presented due to the lacking computational resources.