The effects of blood vessels on electrocorticography

J Neural Eng. 2011 Aug;8(4):044002. doi: 10.1088/1741-2560/8/4/044002. Epub 2011 Jun 8.

Abstract

Electrocorticography, primarily used in a clinical context, is becoming increasingly important for fundamental neuroscientific research, as well as for brain-computer interfaces. Recordings from these implanted electrodes have a number of advantages over non-invasive recordings in terms of band width, spatial resolution, smaller vulnerability to artifacts and overall signal quality. However, an unresolved issue is that signals vary greatly across electrodes. Here, we examine the effect of blood vessels lying between an electrode and the cortex on signals recorded from subdural grid electrodes. Blood vessels of different sizes cover extensive parts of the cortex causing variations in the electrode-cortex connection across grids. The power spectral density of electrodes located on the cortex and electrodes located on blood vessels obtained from eight epilepsy patients is compared. We find that blood vessels affect the power spectral density of the recorded signal in a frequency-band-specific way, in that frequencies between 30 and 70 Hz are attenuated the most. Here, the signal is attenuated on average by 30-40% compared to electrodes directly on the cortex. For lower frequencies this attenuation effect is less pronounced. We conclude that blood vessels influence the signal properties in a non-uniform manner.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Analysis of Variance
  • Blood Vessels / anatomy & histology
  • Blood Vessels / physiology*
  • Cerebral Cortex / anatomy & histology
  • Cerebral Cortex / blood supply
  • Cerebral Cortex / physiology
  • Cerebrovascular Circulation / physiology*
  • Craniotomy
  • Data Interpretation, Statistical
  • Electrodes, Implanted
  • Electroencephalography / methods*
  • Electroencephalography / statistics & numerical data
  • Epilepsy / pathology
  • Epilepsy / surgery
  • Humans
  • User-Computer Interface