Our neuroscience lab is searching for a Computer Science or Engineering student with experience in computer vision and machine learning to tackle a difficult project of detecting and logging sleeping patterns of participants in fMRI studies. Using a camera directed at the participant’s eye during fMRI scans, we have recorded hours of footage to determine when participants fall asleep. Our goal is to establish one or more machine learning (ML) models to automate the process of detecting when participants’ eyes are closed and for how long. It is important to determine when participants fall asleep to remove those portions of the scan data from the neuroimaging analyses.

We would ideally prefer the use of either MATLAB’s Computer Vision Toolbox or Python libraries, such as OpenCV or Tensorflow/Keras, since these are the languages used by the lab. However, we are open to the use of other options if necessary. Two major obstacles that make this a particularly difficult problem:

1. Significant, but patternistic interference in the video due to the magnetic fields and radiofrequency waves of the MRI machine during active scanning

2. Inconsistent camera position and lighting due to the camera and IR illuminator needing to be manually placed for each scan

If you are a CS or engineering student with computer vision / ML experience and are interested in tackling this problem for a summer internship, please contact us using the information below.

---

Dr. Keith Sudheimer
Clinical Assistant Professor
(650) 701-3674
ksudheim@stanford.edu

Dalton Duvio
Clinical Research Coordinator
(650) 498-5904
dlduvio@stanford.edu