Related Experiment Video
Updated: Jul 10, 2026

10:41
A Wireless, Bidirectional Interface for In Vivo Recording and Stimulation of Neural Activity in Freely Behaving Rats
Published on: November 7, 2017
HermesB: a continuous neural recording system for freely behaving primates.
Gopal Santhanam1, Michael D Linderman, Vikash Gilja
1Department of Electrical Engineering, Stanford University, Stanford, CA 94305, USA. gopals@stanford.edu
IEEE Transactions on Bio-Medical Engineering
|November 21, 2007
Summary
A new wearable neural recording system, HermesB, captures brain activity and movement in freely behaving primates. This technology reveals signal variations during natural behaviors, essential for advancing neural prosthetics.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Implantable Devices
Background:
- Chronically implanted electrode arrays are crucial for electrophysiology and neural prosthetics.
- Current experimental methods often require subject restraint, limiting data collection duration and scope.
- Developing systems for freely behaving subjects is key to advancing neural prosthetics.
Purpose of the Study:
- To introduce HermesB, a dual-channel, battery-powered neural recording system with an integrated accelerometer.
- To enable long-duration, continuous neural and acceleration data acquisition in freely behaving primates.
- To investigate neural signal variability and behavioral contexts in unrestrained subjects.
Main Methods:
- Developed HermesB: a self-contained, programmable system recording broadband neural data (30 kS/s) and acceleration.
- Utilized a removable compact flash card for data storage up to 48 hours.
- Collected long-duration data from an adult macaque monkey.
Main Results:
- Observed significant variations in action potential shape and noise over time (up to 24 hours).
- Action potential voltage varied by up to 30%, with step changes up to 25% correlated with head acceleration.
- Reduced 5-25 Hz local field potential (LFP) power and increased firing rate variability during active periods.
- Successfully classified 93% of recording blocks as active/inactive using LFP power thresholds.
Conclusions:
- Neural signals are not stable over long durations and require adaptive processing for spike sorting.
- HermesB provides insights into neural dynamics during natural behaviors, inaccessible with traditional methods.
- The system is valuable for advancing neural prosthetics and electrophysiological research by capturing behavioral context.

