Prediction of arm movement trajectories from ECoG-recordings in humans

J Neurosci Methods. 2008 Jan 15;167(1):105-14. doi: 10.1016/j.jneumeth.2007.10.001. Epub 2007 Oct 10.


Electrocorticographic (ECoG) signals have been shown to contain reliable information about the direction of arm movements and can be used for on-line cursor control. These findings indicate that the ECoG is a potential basis for a brain-machine interface (BMI) for application in paralyzed patients. However, previous approaches to ECoG-BMIs were either based on classification of different movement patterns or on a voluntary modulation of spectral features. For a continuous multi-dimensional BMI control, the prediction of complete movement trajectories, as it has already been shown for spike data and local field potentials (LFPs), would be a desirable addition for the ECoG, too. Here, we examined ECoG signals from six subjects with subdurally implanted ECoG-electrodes during continuous two-dimensional arm movements between random target positions. Our results show that continuous trajectories of 2D hand position can be approximately predicted from the ECoG recorded from hand/arm motor cortex. This indicates that ECoG signals, related to body movements, can directly be transferred to equivalent controls of an external effector for continuous BMI control.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Adult
  • Arm / physiopathology*
  • Brain Mapping
  • Electric Stimulation / methods
  • Electroencephalography*
  • Epilepsy / physiopathology
  • Epilepsy / therapy
  • Humans
  • Motor Cortex / physiopathology*
  • Movement / physiology*
  • Numerical Analysis, Computer-Assisted
  • Predictive Value of Tests
  • Spectrum Analysis
  • User-Computer Interface*