Binocular Rivalry, EEG, and Machine Learning

Research areas

Description

Binocular rivalry occurs when different images are presented simultaneously to the two eyes, but the observer is only able to see one image at a time (see Figure 1). The image seen fluctuates between the two images, in a roughly regular pattern, and the observer has very little influence on this. Interestingly, the rate at which this happens varies considerably between individuals, and can be influenced by physiological states. We're interested in what is happening in the brain during this process, and whether rivalry state can be predicted by brain data. 

We have already begun collecting data for this project, and we are using EEG (electroencephalogram) to measure brain activity during binocular rivalry, and also during a resting state with eyes open and closed. We measure activity during two separate conditions: one where the participant reports their percept (which image they are seeing over time), and another where they don't report the percept but instead try to detect a faint probe stimulus. This should be harder to see when it's in the image that is currently being suppressed, but as we're not asking the observer to report what image is being suppressed here, we try to reconstruct this from the data. 

 
Figure 1. A schematic illustration of binocular rivalry. Participants are presented with two conflicting images, and the image perceived alters across time.
 
 
 
 
 
 
 

Goals

The question for project would be: can we use machine learning to train an observer on the EEG data where the subject reports rivalry state, and test on the EEG data where the percept is not reported? This should predict the likelihood of the probe stimulus being seen, and could potentially provide a new, objective method for measuring binocular rivalry without the need for the observer to report the subjective percept. 

Requirements

This project will appeal to students with excellent skills in experimentation, programming, and teamwork. The preference is for students who have finished/are taking the units of Artificial Intelligence, Document Analysis, and/or Machine Learning in the ANU or similar.

Background Literature

Alais, D., Keetels, M, & Freeman, A.W. (2014). Measuring perception without introspection. Journal of Vision, 14(11). http://jov.arvojournals.org/article.aspx?articleid...

 
Alais, D., Cass, J., O’Shea, R. P., & Blake, R. (2010). Visual sensitivity underlying changes in visual consciousness. Current Biology, 20(15), 1362-7. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2918735/
 
Mousa, F. A., El-Khoribi, R. A., & Shoman, M. E. (2015). EEG classification based on machine learning techniques. International Journal of Computer Applications, 128(4), 22–7.  http://www.ijcaonline.org/research/volume128/numbe...
 

Gain

The student will develop an understanding of the applications of machine learning to human neuroscience data, a rapidly growing field with many opportunities for new developments and discoveries. There is also potential for inclusion on published papers and conference travel to present them if the student's contribution is sufficient. 

Updated:  1 June 2019/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing