Information capacity of the corticospinal tract recordings as a neural interface

Research output: Contribution to journalArticlepeer-review

3 Scopus citations


Recording the motor output of the central nervous system from the cervical spinal cord was investigated as a method of generating voluntary command signals, potentially to be used in quadriplegic individuals. Corticospinal volleys evoked by motor cortex stimulation were recorded from the spinal cord surface with multicontact electrodes in anesthetized cats. The multicontact recordings were analyzed for their information-carrying capacity as a neural interface. Neural signals resulting from the stimulation of various points in the motor cortex were considered as symbols of an alphabet that were sent through a discrete information channel. The information capacity of this channel at the thermal noise level of the electrode contacts was calculated. The maximum information rate was 1.57 bits in a trial for a 4-symbol alphabet. The background noise that reduces the information rate to 50% of its maximum theoretical value was defined as the half-bitrate-noise-tolerance (HBR-NoiseTol) and used as a measure of symbol distinguishability. The HBR-NoiseTol for all trials on average was 24 ± 12%, 18 ± 10%, and 15 ± 9% for interfaces with 2-, 3-, and 4-symbol alphabets (n = 11 trials). The average peak-to-peak amplitude of the neural volleys was 13.5 ± 6.7 μV (n = 11). These results suggest that the corticospinal signals can be recorded with spatial selectivity from the spinal cord surface and thus warrant further investigation of their potential use for a spinal cord-computer interface.

Original languageEnglish (US)
Pages (from-to)823-830
Number of pages8
JournalAnnals of Biomedical Engineering
Issue number6
StatePublished - Jun 2004
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering


  • Brain-computer interface
  • Spinal cord injury
  • Spinal cord recording
  • Voluntary command generation


Dive into the research topics of 'Information capacity of the corticospinal tract recordings as a neural interface'. Together they form a unique fingerprint.

Cite this