Sensitivity and specificity of upper extremity movements decoded from electrocorticogram

An H. Do, Po T. Wang, Christine E. King, Andrew Schombs, Jack J. Lin, Mona Sazgar, Frank P. K. Hsu, Susan J. Shaw, David E. Millett, Charles Y. Liu, Agnieszka A. Szymanska, Luis A. Chui (+1 others)
2013 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)  
Electrocorticogram (ECoG)-based brain computer interfaces (BCI) can potentially be used for control of arm prostheses. Restoring independent function to BCI users with such a system will likely require control of many degrees-offreedom (DOF). However, our ability to decode many-DOF arm movements from ECoG signals has not been thoroughly tested. To this end, we conducted a comprehensive study of the ECoG signals underlying 6 elementary upper extremity movements. Two subjects undergoing ECoG
more » ... ndergoing ECoG electrode grid implantation for epilepsy surgery evaluation participated in the study. For each task, their data were analyzed to design a decoding model to classify ECoG as idling or movement. The decoding models were found to be highly sensitive in detecting movement, but not specific in distinguishing between different movement types. Since sensitivity and specificity must be traded-off, these results imply that conventional ECoG grids may not provide sufficient resolution for decoding many-DOF upper extremity movements.
doi:10.1109/embc.2013.6610824 pmid:24111011 dblp:conf/embc/DoWKSLSHSMLSCN13 fatcat:bibobwppgbcslkfdb7uwualno4