Work Description

Title: Example Data for "A Low-Power Band of Neuronal Spiking Activity Dominated by Local Single Units Improves the Performance of Brain-Machine Interfaces" Open Access Deposited

h
Attribute Value
Methodology
  • This dataset includes example simulated and in vivo data, some of which is synchronized with behavior. The simulated data in MultiUnitSim.mat includes one sorted unit from one nonhuman primate that is replayed spontaneously for 5 seconds to create realistic neural signals, with some signal-processed versions included. The in vivo data in MultiUnitInVivo.mat includes sorted unit spiking data as well as some signal processed versions of it. The OnlineTrainingData.mat includes behavioral, spiking, and spiking band power measurements all synchronized in time, captured from the same nonhuman primate performing a finger task. More details about the methods and the files can be found in Nason et al., 2020, Nature Biomedical Engineering and in the attached README.
Description
  • This data is a subset of the data used to generate figures similar to figures 1, 2, 3, and 4 in Nason et al., 2020, Nature Biomedical Engineering. The purpose of the study was to demonstrate the benefits of using spiking band power, a low-power but single unit specific recording signal, for brain-machine interfaces with nonhuman primates with the potential to impact low-power brain-machine interfaces with humans. All of the data is contained in .mat files, which can be commonly opened by Matlab and the Python scipy library.
Creator
Depositor
  • samnason@umich.edu
Contact information
Discipline
Funding agency
  • National Institutes of Health (NIH)
Keyword
Date coverage
  • 2018-03
Citations to related material
  • Nason, S.R., Vaskov, A.K., Willsey, M.S., Welle, E.J., An, H., Vu, P.P., Bullard, A.J., Nu, C.S., Kao, J.C., Shenoy, K.V., Jang, T., Kim, H.-S., Blaauw, D., Patil, P.G., and Chestek, C.A. (2020). A low-power band of neuronal spiking activity dominated by local single units improves the performance of brain–machine interfaces. Nat. Biomed. Eng. 4, 973–983. https://doi.org/10.1038/s41551-020-0591-0
Resource type
Last modified
  • 11/17/2022
Published
  • 07/22/2021
Language
DOI
  • https://doi.org/10.7302/wwya-5q86
License
To Cite this Work:
Nason, S. R., Vaskov, A. K., Willsey, M. S., Welle, E. J., An, H., Vu, P. P., Bullard, A. J., Nu, C. S., Kao, J. C., Shenoy, K. V., Jang, T., Kim, H., Blaauw, D., Patil, P. G., Chestek, C. A. (2021). Example Data for "A Low-Power Band of Neuronal Spiking Activity Dominated by Local Single Units Improves the Performance of Brain-Machine Interfaces" [Data set], University of Michigan - Deep Blue Data. https://doi.org/10.7302/wwya-5q86

Relationships

This work is not a member of any user collections.

Files (Count: 4; Size: 74.4 MB)

************************************************************************************************************************************
Example datasets from Nason et al., 2020, Nature Biomedical Engineering

"A low-power band of neuronal spiking activity dominated by local single units improves the performance of brain–machine interfaces"

Compiled by Samuel Nason for the Chestek Lab, March 10, 2021
Please contact samnason@umich.edu with any questions.

To cite this data set:
Nason, S., Vaskov, A., Willsey, M., Welle, E., An, H., Vu, P., Bullard, A., Nu, C., Kao, J., Shenoy, K., Jang, T., Kim, H., Blaauw, D., Patil, P., Chestek, C. Example Data for "A Low-Power Band of Neuronal Spiking Activity Dominated by Local Single Units Improves the Performance of Brain-Machine Interfaces" [Data set], University of Michigan - Deep Blue Data. https://doi.org/10.7302/wwya-5q86

Version 1: Initial upload
************************************************************************************************************************************

1. MultiUnitSim.mat
Description: This is example data used to generate the simulation side of figure 3. Two arrangements were simulated, those in figure 3a and figure 3c. Five simulations are included for each arrangement, each having random noise and random spike timings. Modifications to this simulated data can generate components of figures 1 and 2 as well.
Fields:
- SamplingRate: Contains the sampling rate at which neural activity was simulated, in Hz.
- DownsampledRate: Contains the downsampled rate of the final comparison signal, in Hz.
- SpikingBand: The filter settings used to filter simulated recordings, in Hz.
- SimulationDuration: The amount of time simulated, in seconds.
- SmoothDuration: The time window used to zero-phase smooth at DownsampledRate, in seconds.
- SmoothType: The type of smoothing used.
- SmoothWindow: The smoothing values used.
- SNRs: The signal-to-noise ratios of the simulated neurons.
- FRs: The average firing rates of the simulated neurons, in Hz.
- Simulations.SpikeTimes: An indicator vector for each neuron's initiation of a spike, at SamplingRate.
- Simulations.IndividualRecordings: Simulated recordings for each neuron without noise, at SamplingRate. Amplitudes are scaled to match each neuron's SNR.
- Simulations.Noise: Simulated white noise from a standard normal distribution, at SamplingRate.
- Simulations.TrueFR: An indicator vector for each neuron's initiation of a spike, at DownsampledRate.
- Simulations.SBP: The raw spiking band power of the field of neurons. This was computed by:
1) Summing Simulations.IndividualRecordings along dimension 2.
2) Summing the result of step 1 with Simulations.Noise.
3) Filtering to SpikingBand with a 2nd order causal Butterworth filter.
4) Downsampling to DownsampledRate.
5) Taking the absolute value.
- Simulations.SmoothedTrueFR: Smoothed true firing rate of each simulated neuron computed by using SmoothWindow to zero-phase filter Simulations.TrueFR.
- Simulations.SmoothedSBP: Smoothed spiking band power, computed by using SmoothWindow to zero-phase filter Simulations.SBP.

2. MultiUnitInVivo.mat
Description: This is the data used to generate figure 3e. This recording was obtained from one of monkey W's channels.
Fields:
- SamplesPerSpike: The number of samples used to assume the duration of one spike.
- SamplingRate: Contains the sampling rate at which neural activity was simulated, in Hz.
- DownsampledRate: Contains the downsampled rate of the final comparison signal, in Hz.
- SpikingBand: The filter settings used to filter simulated recordings, in Hz.
- SmoothDuration: The time window used to zero-phase smooth at DownsampledRate, in seconds.
- SmoothType: The type of smoothing used.
- SmoothWindow: The smoothing values used.
- UnitMin: The minimum peak of the averaged waveform for each unit, in uV.
- UnitAmp: The peak-to-peak amplitude of the averaged waveform for each unit, in uV.
- Raw: 10 seconds of raw recording, hardware-filtered with 1st-order high-pass at 0.3Hz and 3rd-order low-pass at 7.5kHz, and captured at SamplingRate.
- SpikeTimes: Each element of the cell array contains the spike times corresponding to each unit, in samples.
- FR: An indicator of a spike for each unit captured at DownsampledRate.
- SBP: The raw spiking band power of the recording. This was computed by:
1) Filtering Raw to SpikingBand with a 2nd order causal Butterworth filter.
2) Downsampling to DownsampledRate.
3) Taking the absolute value.
- SmoothedFR: Smoothed true firing rate of each unit computed by using SmoothWindow to zero-phase filter FR.
- SmoothedSBP: Smoothed spiking band power, computed by using SmoothWindow to zero-phase filter SBP.

3. OnlineTrainingData.mat
Description: One of the sets of training data used to train online spiking band power and threshold crossing rate Kalman filters, contributing to figure 4, from monkey W. The monkey performed a center-out task moving all of his fingers together. Each element of the structure array contains the information for one trial. Behavior is normalized to the range of 0 (full extension) to 1 (full flexion). The required hold time for successful acquisition in all trials was 750ms. Each target occuppied 15% of the full range of motion. Each trial had an automatic timeout of 10s. All included trials were consecutive and successful.
Fields:
- TargetPos: The position of the center of the target in the range of motion.
- FingerAngle: The position of the fingers within the range of motion, sampled every 1ms.
- ExperimentTime: The time of the experiment, in ms.
- SpikingBandPower: The spiking band power for all 96 channels accumulated each ms, acquired via the following procedure:
1) Filtered to 300-1,000Hz using the Digital Filter Editor in the Central Software Suite (Blackrock Microsystems, LLC).
2) Sampled at 2 kilo-samples per second.
3) Transmitted from the Cerebus system to the xPC Target system (see methods in the main paper) at 2 kilo-samples per second.
4) All samples received by the xPC within 1ms are absolute-valued then summed within each channel, contained in the SpikingBandPower field.
- SampleWidth: The number of SpikingBandPower samples received by the xPC during each 1ms.
- Channel: An array of structures containing the threshold crossing times for each of the 96 channels. Threshold crossing times correspond to ExperimentTime.

Download All Files (To download individual files, select them in the “Files” panel above)

Best for data sets < 3 GB. Downloads all files plus metadata into a zip file.



Best for data sets > 3 GB. Globus is the platform Deep Blue Data uses to make large data sets available.   More about Globus

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.