Deep learning for neuroimaging - Exercise

Description:

• This exercise is a copy from Brainkhack School, but have small adaptation and change the recipient.
• This exercise seems unrelated to the video clip in this module, but idea of deep learning application is the same. We would learn how to apply basic machine learning and deep learning decoding on fMRI data.
• You will follow the tutorial in MAIN educational workshop and complete the exercises in Brain decoding with SVM and Brain decoding with MLP as Jupyter notebook files.
• Follow the steps in submission and hand in your assignment to TA.

Instructions:

• Let’s have a look at application on functional data with the MAIN educational workshop on brain encoding and decoding .
• Please follow the introduction , set-up your environment and clone the material from GitHub.
• Throughout this tutorial you will be using the Haxby data set. Please read through and understand how to access it here and go through the original support-vector machine analysis of the study and complete the exercises. Write your answer in the Jupyter notebook file.
• After understanding the workflow of functional data, please go through the Multi-Layer Perceptron and complete the relevant exercise Jupyter notebook file.
• (Optional) You can learn about the Graph Convolution Network and how to work with timeseries data!

Submission:

• You should write your answers and execute the outcomes as a Jupyter notebook file or on Colab. You can have separate files or one file for exercises in SVM and MLP.
• If you would like to submit a Colab link, make sure the Colab is viewable for anyone who have the link.
• Mail your files or Colab link to brainhackschooltaiwan@gmail.com with the subject title [BHSTW] <Your_Student_ID> Deep learning for neuroimaging (e.g., [BHSTW] B05202021 Deep learning for neuroimaging) .
• Attempts: no limits
• Points: 0 (fail) / 1 (pass)