Data for: Brain control of bimanual movement enabled by recurrent neural networks
Data files
Dec 28, 2023 version files 138.36 MB
-
bimanualData.zip
138.35 MB
-
README.md
2.16 KB
Jan 02, 2024 version files 138.36 MB
-
bimanualData.zip
138.35 MB
-
README.md
2.16 KB
Jan 18, 2024 version files 138.36 MB
-
bimanualData.zip
138.35 MB
-
README.md
2.17 KB
Abstract
Brain-computer interfaces have so far focused largely on enabling the control of a single effector, for example a single computer cursor or robotic arm. Restoring multi-effector motion could unlock greater functionality for people with paralysis (e.g., bimanual movement). However, it may prove challenging to decode the simultaneous motion of multiple effectors, as we recently found that a compositional neural code links movements across all limbs and that neural tuning changes nonlinearly during dual-effector motion. In this study, we demonstrate the feasibility of high-quality bimanual control of two cursors via neural network (NN) decoders.
This dataset represents all neural activity recorded during these experiments. This includes the neural activity corresponding to unimanual and bimanual hand movements during (1) instructed delay experiments and (2) real-time BCI control of two cursors.
Code associated with the data can be found here: https://github.com/d-r-deo/bimanualBCI
The journal article can be found here: https://doi.org/10.1038/s41598-024-51617-3
README: Data from: Brain control of bimanual movement enabled by recurrent neural networks
https://doi.org/10.5061/dryad.sn02v6xbb
Brain-computer interfaces have so far focused largely on enabling the control of a single effector, for example a single computer cursor or robotic arm. Restoring multi-effector motion could unlock greater functionality for people with paralysis (e.g., bimanual movement). However, it may prove challenging to decode the simultaneous motion of multiple effectors, as we recently found that a compositional neural code links movements across all limbs and that neural tuning changes nonlinearly during dual-effector motion. In this study, we demonstrate the feasibility of high-quality bimanual control of two cursors via neural network (NN) decoders. Neural activity was recorded with microelectrode arrays, and neural features are provided in the form of binned threshold crossings (20 ms bins).
This dataset represents all of the neural activity recorded during these experiments. This includes the neural activity corresponding to unimanual and bimanual hand movements during (1) instructed delay experiments and (2) real-time BCI control of two cursors.
Description of the data and file structure
There is one .zip file that contains the data for each of the 13 sessions included in the study.
Each session's data is provided as its own .mat file and may contain the following:
(1) data from the instructed delay tasks that assessed tuning to attempted unimanual and bimanual hand movements (these data are referred to as 'open-loop' or OL)
(2) data from the real-time BCI control of two cursors (these data are referred to as 'closed-loop' or CL, and occasionally as 'closed-loop recalibration' or CLR)
All data consists of .mat files that are intended to be loaded with MATLAB or Python (scipy.io.loadmat). There is one readme file that describes the variables and experiments in detail.
Sharing/Access information
The data currently reside only on Dryad, and data were not derived from any other sources.
Code/Software
Code associated with the data can be found here: https://github.com/d-r-deo/bimanualBCI
Methods
Neural signals were recorded from two 96-channel Utah microelectrode arrays using the NeuroPortTM system from Blackrock Microsystems. First, neural signals were analog filtered from 0.3 to 7.5 kHz and subsequently digitized at 30kHz with 250 nV resolution. Next, common mode noise reduction was accomplished via a common average reference filter which subtracted the average signal across the array from every electrode. Finally, a digital high-pass filter at 250 Hz was applied to each electrode prior to spike detection.
Spike threshold crossing detection was implemented using a -3.5 x RMS threshold applied to each electrode, where RMS is the electrode-specific root mean square of the time series voltage recorded on that electrode.
Neural data was recorded from participant T5 in 3-5 hour “sessions”, with breaks, on scheduled days. T5 either performed attempted movements of one or both hands as governed by an instructed delay task, or performed real-time brain-computer interface control of two cursors.
This dataset is deidentified and contains binned spiking data (20ms bins) and task information, such as cursor positions, target positions, and trial timing.