Data from: Using adversarial networks to extend brain computer interface decoding accuracy over time

Existing intracortical brain computer interfaces (iBCIs) transform neural activity into control signals capable of restoring movement to persons with paralysis. However, the accuracy of the "decoder" at the heart of the iBCI typically degrades over time due to turnover of recorded neurons. To compensate, decoders can be recalibrated, but this requires the user to spend extra time and effort to provide the necessary data, then learn the new dynamics. As the recorded neurons change, one can think of the underlying movement intent signal being expressed in changing coordinates. If a mapping can be computed between the different coordinate systems, it may be possible to stabilize the original decoder's mapping from brain to behavior without recalibration. We previously proposed a method based on Generalized Adversarial Networks (GANs), called "Adversarial Domain Adaptation Network" (ADAN), which aligns the distributions of latent signals within underlying low-dimensional neural manifolds. However, we tested ADAN on only a very limited dataset. Here we propose a method based on Cycle-Consistent Adversarial Networks (Cycle-GAN), which aligns the distributions of the full-dimensional neural recordings. We tested both Cycle-GAN and ADAN on data from multiple monkeys and behaviors and compared them to a third, quite different method based on Procrustes alignment of axes provided by factor analysis. All three methods are unsupervised and require little data, making them practical in real life. Overall, Cycle-GAN had the best performance and was easier to train and more robust than ADAN, making it ideal for stabilizing iBCI systems over time.This data set includes behavioral recordings and extracellular neural recordings from the primary motor cortex of a rhesus macaque during 20 separate isometric wrist experiments over 95 days. Stephanie N. Thacker collected the data in the laboratory of Lee Miller in 2015 when she was a Ph.D student in Northwestern University. Xuan Ma and Fabio Rizzoglio processed the data for use in [1], which utilized adversarial neural networks to "align" the neural signals over a long period of time to extend the accuracy of an iBCI decoder calibrated only on the very first day (day-0). Please cite this article if you are using this dataset: [1] Ma X, Rizzoglio F, Perreault EJ, Miller LE, Kennedy A. Using adversarial networks to extend brain computer interface decoding accuracy over time. eLife 2023;12:e84296. DOI: https://doi.org/10.7554/eLife.84296 There are in total 20 sessions recorded over 95 days, named as 'Jango_YYYYMMDD_XXX.mat'. 'Jango' is the name of the monkey. 'YYYYMMDD' presents the date when the recording session was made. XXX' is the number of the recording session on a day, but since there was only one session on each day 'XXX' is always '001'. Different types of data on each day are organized in a data structure named 'xds' (cross-platform data structure) and saved in MATLAB '.mat' format files. These files could be directly opened in MATLAB, and you can develop your own MATLAB codes to analyze them. We also provide Python codes ('xds', https://github.com/limblab/xds) to load these '.mat' files in Python ('h5py' is needed). A tutorial for 'xds' can be found here. You could develop your own Python codes for data analysis, but there is no need to write codes from scratch using 'h5py' or 'scipy' to load the data files by yourself, since we have already done it for you.Funding provided by: National Institutes of HealthCrossref Funder Registry ID: http://dx.doi.org/10.13039/100000002Award Number: R01 NS053603Funding provided by: National Institutes of HealthCrossref Funder Registry ID: http://dx.doi.org/10.13039/100000002Award Number: R01 NS074044Electrophysiology: We implanted a 96-channel Utah electrode array (Blackrock Neurotech, Inc.) in the hand representation area of the right primary motor cortex (M1) of a monkey (monkey J), contralateral to the arm being used for the task. The implant site was pre-planned and finally determined during the surgery with reference to the sulcal patterns and the muscle contractions evoked by intraoperative surface cortical stimulation. We also implanted intramuscular leads in forearm and hand muscles of the arm used for the task in a separate procedure. Electrode locations were verified during surgery by stimulating each lead. Behavioral task: Monkeys J was trained to perform an isometric wrist task, which required him to control the cursor on a screen by exerting forces on a small box placed around the left hand. The box was padded to comfortably constrain the monkey's hand and minimize the movement within the box, and the forces were measured by a 6 DOF load cell (JR3 Inc., CA) aligned to the wrist joint. During the task, flexion/extension force moved the cursor right and left respectively, while force along the radial/ulnar deviation axis moved the cursor up and down. Each trial started with the appearance of a center target requiring the monkey to hold for a random time (0.2 – 1.0 s), after which one of eight possible outer targets selected in a block-randomized fashion appeared, accompanied with an auditory go cue. The monkey was allowed to move the cursor to the target within 2.0 s and hold for 0.8 s to receive a liquid reward. All surgical and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Northwestern University under protocol #IS00000367, and are consistent with the Guide for the Care and Use of Laboratory Animals. Data collection and preprocessing: M1 neural activity was recorded during task performance using a Cerebus system (Blackrock Neurotech, Inc.). The signals on each channel were digitalized, bandpass filtered (250 ~ 5000 Hz) and converted to spike times based on threshold crossings. The threshold was set with respect to the root-mean square (RMS) activity on each channel and kept consistent across different recording sessions (-5.5 x RMS). The time stamp and a 1.6 ms snippet of each spike surrounding the time of threshold crossing were recorded. For all analyses in this study, we used multiunit threshold crossings on each channel instead of discriminating well isolated single units. We applied a Gaussian kernel (S.D. = 100 ms) to the spike counts in 50 ms, non-overlapping bins to obtain a smoothed estimate of firing rate as function of time for each channel. The EMG signals were differentially amplified, band-pass filtered (4-pole, 50 ~ 500 Hz) and sampled at 2000 Hz. The EMGs were subsequently digitally rectified and low-pass filtered (4-pole, 10 Hz, Butterworth) and subsampled to 20 Hz. EMG channels with substantial noise were not included in the analyses