Abstract:
In this paper, we introduce PEER (Predictive Eye Estimation Regression) - a simple, image-based approach to eye tracking that can occur simultaneously with fMRI experiments. We believe this to be the first report of eye tracking based on MR images. Multivariate SVM regression was used to model the relationship between acquired EPI images and fixation position. This SVM model can be used to estimate fixation in fMRI runs with matched sequence parameters. Successful eye tracking during extended periods of fixation and eye movement were demonstrated on a TR-by-TR basis. Notably, PEER does not alter fMRI results and can be applied at any fMRI site.
Reference:
abstract S. LaConte, S. Peltier, K. Heberlein, X. Hu. Predictive Eye Estimation Regression (PEER) for Simultaneous Eye Tracking and fMRI. In Proceedings 14th Scientific Meeting, International Society for Magnetic Resonance in Medicine, Seattle, page 2808, 2006. [bibtex]
Bibtex Entry:
@inproceedings{Seattle2808,
Author = {LaConte, S. and Peltier, S. and Heberlein, K. and Hu, X.},
Title ={Predictive Eye Estimation Regression ({P}{E}{E}{R}) for Simultaneous Eye Tracking and f{M}{R}{I}},
BookTitle = {Proceedings 14th Scientific Meeting, International Society for Magnetic Resonance in Medicine, Seattle},
Pages = {2808},
Abstract = {In this paper, we introduce PEER (Predictive Eye Estimation Regression) - a simple, image-based approach to eye tracking that can occur simultaneously with fMRI experiments. We believe this to be the first report of eye tracking based on MR images. Multivariate SVM regression was used to model the relationship between acquired EPI images and fixation position. This SVM model can be used to estimate fixation in fMRI runs with matched sequence parameters. Successful eye tracking during extended periods of fixation and eye movement were demonstrated on a TR-by-TR basis. Notably, PEER does not alter fMRI results and can be applied at any fMRI site.},
Keywords = {Seattle2808},
Year = {2006} }