PREDICTION OF ARM 3D-TRAJECTORY FROM HUMAN ELECTROCORTICOGRAMS

Yasuhiko Nakanishi, Takafumi Yanagisawa, Duk Shin, Chao Chen, Hiroyuki Kambara, Natsue Yoshimura, Masayuki Hirata, Toshiki Yoshimine, Yasuharu Koike

Abstract

Electrocorticogram (ECoG), which is less invasive than intracotical microwire, has higher spatiotemporal resolution than EEG. Although a number of studies have predicted three dimensional (3D) trajectories of monkeys’ arm with ECoG signals, those of human arms are not available yet as far as we know. In this study, we applied ECoG to a patient suffering from thalamic hemorrhage and predicted his 3D arm trajectories (4 joint angles and 6 coordinates at arm joints). He performed tasks of rotating three objects clockwise on a table. As results of leave-one-out cross-validation (LOO-CV), average Pearson’s correlation coefficients (CC) and normalized root-mean-square errors (nRMSE) were 0.44 ~ 0.73 and 0.18 ~ 0.42, respectively. We expect that our proposed method can contribute to further research in neuro-prosthesis and neuro-rehabilitation with ECoG signals.

Keywords

brain machine interface, electrocorticogram, human sensorimotor cortex

Download

IMEKO-TC18-2013-014.pdf.pdf

DOI

-

IMEKO TC

TC18 - Measurement of Human Functions

Back to the proceedings