Real-time supervised learning of fMRI-based cognitive states

by Cyrus Eierud (Virginia Tech and Baylor College of Medicine)

Friday 9 Sept 2011, 10:00 → 11:00 Europe/Stockholm

RB35

Since the soft margin support vector machine (SVM) was proposed by Vapnik and Cortes (1995) SVM classification has been applied to finance, genome and many other areas. In 2001 Haxby et al. differentiated brain fMRI patterns between object classes (i.e. subject seeing houses and faces). In 2005 LaConte et al. presented a SVM solution, which classified fMRI mind states. In 2007 Laconte et al. published a real-time solution where subjects could control a cursor with the mind using SVM and fMRI.