Open AccessDecoding of Visual Attention from LFP Signals of Macaque MT
Author(s) -
Moein Esghaei,
Mohammad Reza Daliri
Publication year - 2014
Publication title -
plos one
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.99
H-Index - 332
ISSN - 1932-6203
DOI - 10.1371/journal.pone.0100381
Subject(s) - local field potential , macaque , decoding methods , focus (optics) , support vector machine , computer science , brain–computer interface , artificial intelligence , classifier (uml) , pattern recognition (psychology) , neuroscience , speech recognition , physics , electroencephalography , biology , telecommunications , optics
The local field potential (LFP) has recently been widely used in brain computer interfaces (BCI). Here we used power of LFP recorded from area MT of a macaque monkey to decode where the animal covertly attended. Support vector machines (SVM) were used to learn the pattern of power at different frequencies for attention to two possible positions. We found that LFP power at both low (<9 Hz) and high (31–120 Hz) frequencies contains sufficient information to decode the focus of attention. Highest decoding performance was found for gamma frequencies (31–120 Hz) and reached 82%. In contrast low frequencies (<9 Hz) could help the classifier reach a higher decoding performance with a smaller amount of training data. Consequently, we suggest that low frequency LFP can provide fast but coarse information regarding the focus of attention, while higher frequencies of the LFP deliver more accurate but less timely information about the focus of attention.