Introduction
Understanding the cognitive state of the brain has been an interesting problem attracting researchers from many domains. With the advent of Functional magnetic Resonance Imaging (FMRI), a 3 dimensional brain imaging method that describes the state of brain by observing the change in blood flow level there was a breakthrough in understanding the state of brain during decision making or any cognitive state. Use of Machine Learning algorithms for predicting the cognitive state became feasible through FMRI data of brain. This is further supported by the evidence that the change of blood level in brain is caused by the neurons. The fact that change in blood flow and neuron activation are correlated gives a strong motivation to use Machine Learning algorithms to detect the cognitive state of the brain. We exercised machine learning methods and used them to train on temporal fMRI patterns which support probabilistic predictions about the cognitive states of the human subject.
Problem Statement
he problem we are addressing is to train classifiers to predict whether the subject is perceiving a picture or text. The project will focus on applying machine learning methods to train classifiers on a sparse, noisy and high dimensional data. Appropriate feature selection, feature abstraction and classifier training methods have been applied.
Algorithms
- Support Vector Machines
- Logistic regression
- Gaussian Naive Bayes
- Random Forest
- K Nearest Neighbours.
Technologies and Tools
- Python
- Sklearn
Code and Report
The code for the project can be found here.
The report for further reading can be found here.
Acknowledgment
- I would like to thank Dr. Hanghang Tong for guiding us throughout the project.
- I would like to thank my team members Aadhavan Sadasivam, Avinash Patil, Manjusha Ravindranath and Satrajit Maitra for their contributions.