Brain Machine Interface for forces and motion based on musculo-skeletal model

• ¹ Tokyo Institute of Technology, Japan

Non-invasive measurement method, such as EEG, fMRI or NIRS, has been used for Brain Computer Interface. EEG has nice temporal resolution, and it is used for BCI, such as amplitudes of different frequency bands; imagining movement of different parts of the body; slow cortical potentials and gamma band rhythms. Recently, Electrocorticography (ECoG) is an alternative approach to less invasive BMIs. Since ECoG records directly from neuronal activities on the cortical surface, ECoG has higher spatio- temporal resolution with better signal-to-noise ratio than scalp EEG. ECoG has also shown potential as a stable long-term recording method. Several studies using ECoG have already succeeded in the classification of movement direction, grasp type, and prediction of hand trajectory. Despite these successes, however, there still remains considerable work for the realization of ECoG- based prosthesis. The human neuromuscular system naturally modulates mechanical stiffness and vis- cosity to achieve proper interaction with the environment. Current rehabilitation robots can perform sophisticated operations including stiffness control. Our previous works suggested that the angle, torque, and stiffness of joints can be predicted from muscle activity. Therefore, decoding muscle activity is an important component for realizing BMI systems capable of controlling interaction force or stiffness. In this talk, we introduce a reconstruction of muscle activity time series using a simple mathematical musculoskeletal model (1,2). Then we introduce BMI using this musculoskeletal model from EEG, ECoG recordings (3-6). Reference (1) Hiroyuki Kambara, Duk Shin, and Yasuharu Koike. A computational model for optimal muscle activity considering muscle viscoelasticity in wrist movements. Journal of neurophysiology, 109:2145–60, 2013. (2) Duk Shin, Jaehyo Kim, and Yasuharu Koike. A myokinetic arm model for estimating joint torque and stiffness from EMG signals during maintained posture. Journal of neurophysiology, 101:387–401, 2009. (3) N Yoshimura, C S Dasalla, T Hanakawa, M A Sato, and Y Koike. Reconstruction of flexor and extensor muscle activities from electroencephalography cortical currents. Neuroimage, 59(2):1324–1337, 2012. (4) Duk Shin, Hidenori Watanabe, Hiroyuki Kambara, Atsushi Nambu, Tadashi Isa, Yukio Nishimura, and Yasuharu Koike. Prediction of Muscle Activities from Electrocorticograms in Primary Motor Cortex of Primates. PLoS ONE, 7, 2012. (5) Yasuhiko Nakanishi, Takufumi Yanagisawa, Duk Shin, Ryohei Fukuma, Chao Chen, Hiroyuki Kambara, Nat- sue Yoshimura, Masayuki Hirata, Toshiki Yoshimine, and Yasuharu Koike. Prediction of Three-Dimensional Arm Trajectories Based on ECoG Signals Recorded from Human Sensorimotor Cortex. PLoS ONE, 8, 2013. (6) Chao Chen, Duk Shin, Hidenori Watanabe, Yasuhiko Nakanishi, Hiroyuki Kambara, Natsue Yoshimura, Atsushi Nambu, Tadashi Isa, Yukio Nishimura, and Yasuharu Koike. Decoding grasp force profile from electrocorticography signals in non-human primate sensorimotor cortex. Neuroscience research, 83:1–7, 2014.