0. Schedule

---

Meeting 10/19

Talk by Andrew Haun, from the Rokers vision lab

• Motion and depth perception fMRI data
• Relevant works investigating how motion and depth are represented in the brain
• Future directions that we can explore

I am working on some decoding analysis at this repo: motionDecoding_fMRI

---

Meeting 10/26

Intro to MVPA by Qihong:

• Matlab & Github basics
• Linear systems
• Full rank systems
• "Inconsistent" rows
• Underdetermined systems in the context of fMRI data
• Classification
• Least square classification
• Logistic regression
• Resources
• Linear algebra cheat sheet, by Professor Laurent Lessard.

---

Meeting 11/2

Intro to MVPA by Qihong:

• Regularization
• Motivation for sparsity
• Compare Lasso vs. Ridge penalty, varying true sparsity
• Introduce the motion fMRI data

---

Meeting 11/9

Intro to MVPA by Qihong:

• Elastic net with Glmnet
• Experimental Methodology
• Underfitting & Overfitting - bias & variance
• Train & test & validation, K-Folds cross validation
• Matlab practice

---

Meeting 11/16

Intro to MVPA by Qihong:

• Practice with the motion fMRI data
• Decode 2d motion: whole brain model vs. V1 model

---

Meeting 11/23 - Cancelled because of Thanksgiving

---

Meeting 11/30

Feature selection by Qihong:

• Iterative Lasso

---

Meeting 12/7

• Methodological reflection by Tim and Qihong
• What's next?!

---

Meeting 12/14 - Cancelled because of final exams

---

Meeting 12/20

Another Logistic Lasso practice by Qihong