Robust and accurate decoding of hand kinematics from entire spiking activity using deep learning

Increasing Robustness of Intracortical Brain-Computer Interfaces for Recording Condition Changes via Data Augmentation

BACKGROUND AND OBJECTIVE

Intracortical brain-computer interfaces (iBCIs) aim to help paralyzed individuals restore their motor functions by decoding neural activity into intended movement. However, changes in neural recording conditions hinder the decoding performance of iBCIs, mainly because the neural-to-kinematic mappings shift. Conventional approaches involve either training the neural decoders using large datasets before deploying the iBCI or conducting frequent calibrations during its operation. However, collecting data for extended periods can cause user fatigue, negatively impacting the quality and consistency of neural signals. Furthermore, frequent calibration imposes a substantial computational load.

METHODS

This study proposes a novel approach to increase iBCIs' robustness against changing recording conditions. The approach uses three neural augmentation operators to generate augmented neural activity that mimics common recording conditions. Then, contrastive learning is used to learn latent factors by maximizing the similarity between the augmented neural activities. The learned factors are expected to remain stable despite varying recording conditions and maintain a consistent correlation with the intended movement.

RESULTS

Experimental results demonstrate that the proposed iBCI outperformed the state-of-the-art iBCIs and was robust to changing recording conditions across days for long-term use on one publicly available nonhuman primate dataset. It achieved satisfactory offline decoding performance, even when a large training dataset was unavailable.

CONCLUSIONS

This study paves the way for reducing the need for frequent calibration of iBCIs and collecting a large amount of annotated training data. Potential future works aim to improve offline decoding performance with an ultra-small training dataset and improve the iBCIs' robustness to severely disabled electrodes.

Neural Data Transformer 2: Multi-context Pretraining for Neural Spiking Activity

The neural population spiking activity recorded by intracortical brain-computer interfaces (iBCIs) contain rich structure. Current models of such spiking activity are largely prepared for individual experimental contexts, restricting data volume to that collectable within a single session and limiting the effectiveness of deep neural networks (DNNs). The purported challenge in aggregating neural spiking data is the pervasiveness of context-dependent shifts in the neural data distributions. However, large scale unsupervised pretraining by nature spans heterogeneous data, and has proven to be a fundamental recipe for successful representation learning across deep learning. We thus develop Neural Data Transformer 2 (NDT2), a spatiotemporal Transformer for neural spiking activity, and demonstrate that pretraining can leverage motor BCI datasets that span sessions, subjects, and experimental tasks. NDT2 enables rapid adaptation to novel contexts in downstream decoding tasks and opens the path to deployment of pretrained DNNs for iBCI control. Code: https://github.com/joel99/context_general_bci.

Calibration-Free and Hardware-Efficient Neural Spike Detection for Brain Machine Interfaces

Recent translational efforts in brain-machine interfaces (BMI) are demonstrating the potential to help people with neurological disorders. The current trend in BMI technology is to increase the number of recording channels to the thousands, resulting in the generation of vast amounts of raw data. This in turn places high bandwidth requirements for data transmission, which increases power consumption and thermal dissipation of implanted systems. On-implant compression and/or feature extraction are therefore becoming essential to limiting this increase in bandwidth, but add further power constraints - the power required for data reduction must remain less than the power saved through bandwidth reduction. Spike detection is a common feature extraction technique used for intracortical BMIs. In this article, we develop a novel firing-rate-based spike detection algorithm that requires no external training and is hardware efficient and therefore ideally suited for real-time applications. Key performance and implementation metrics such as detection accuracy, adaptability in chronic deployment, power consumption, area utilization, and channel scalability are benchmarked against existing methods using various datasets. The algorithm is first validated using a reconfigurable hardware (FPGA) platform and then ported to a digital ASIC implementation in both 65 nm and 0.18 μm CMOS technologies. The 128-channel ASIC design implemented in a 65 nm CMOS technology occupies 0.096 mm2 silicon area and consumes 4.86 μW from a 1.2 V power supply. The adaptive algorithm achieves a 96% spike detection accuracy on a commonly used synthetic dataset, without the need for any prior training.

Neural Decoding for Intracortical Brain-Computer Interfaces

Brain-computer interfaces have revolutionized the field of neuroscience by providing a solution for paralyzed patients to control external devices and improve the quality of daily life. To accurately and stably control effectors, it is important for decoders to recognize an individual's motor intention from neural activity either by noninvasive or intracortical neural recording. Intracortical recording is an invasive way of measuring neural electrical activity with high temporal and spatial resolution. Herein, we review recent developments in neural signal decoding methods for intracortical brain-computer interfaces. These methods have achieved good performance in analyzing neural activity and controlling robots and prostheses in nonhuman primates and humans. For more complex paradigms in motor rehabilitation or other clinical applications, there remains more space for further improvements of decoders.

The influence of non-stationarity of spike signals on decoding performance in intracortical brain-computer interface: a simulation study

Introduction

Intracortical Brain-Computer Interfaces (iBCI) establish a new pathway to restore motor functions in individuals with paralysis by interfacing directly with the brain to translate movement intention into action. However, the development of iBCI applications is hindered by the non-stationarity of neural signals induced by the recording degradation and neuronal property variance. Many iBCI decoders were developed to overcome this non-stationarity, but its effect on decoding performance remains largely unknown, posing a critical challenge for the practical application of iBCI.

Methods

To improve our understanding on the effect of non-stationarity, we conducted a 2D-cursor simulation study to examine the influence of various types of non-stationarities. Concentrating on spike signal changes in chronic intracortical recording, we used the following three metrics to simulate the non-stationarity: mean firing rate (MFR), number of isolated units (NIU), and neural preferred directions (PDs). MFR and NIU were decreased to simulate the recording degradation while PDs were changed to simulate the neuronal property variance. Performance evaluation based on simulation data was then conducted on three decoders and two different training schemes. Optimal Linear Estimation (OLE), Kalman Filter (KF), and Recurrent Neural Network (RNN) were implemented as decoders and trained using static and retrained schemes.

Results

In our evaluation, RNN decoder and retrained scheme showed consistent better performance under small recording degradation. However, the serious signal degradation would cause significant performance to drop eventually. On the other hand, RNN performs significantly better than the other two decoders in decoding simulated non-stationary spike signals, and the retrained scheme maintains the decoders' high performance when changes are limited to PDs.

Discussion

Our simulation work demonstrates the effects of neural signal non-stationarity on decoding performance and serves as a reference for selecting decoders and training schemes in chronic iBCI. Our result suggests that comparing to KF and OLE, RNN has better or equivalent performance using both training schemes. Performance of decoders under static scheme is influenced by recording degradation and neuronal property variation while decoders under retrained scheme are only influenced by the former one.

Algorithm and hardware considerations for real-time neural signal on-implant processing

Objective Various on-workstation neural-spike-based brain machine interface(BMI) systems have reached the point of in-human trials, but on-node and on-implant BMI systems are still under exploration. Such systems are constrained by the area and battery. Researchers should consider the algorithm complexity, available resources, power budgets, CMOS technologies, and the choice of platforms when designing BMI systems. However, the effect of these factors is currently still unclear. Approaches. Here we have proposed a novel real-time 128 channel spike detection algorithm and optimised it on Microcontroller(MCU) and Field Programmable Gate Array(FPGA) platforms towards consuming minimal power and memory/resources. It is presented as a use case to explore the different considerations in system design. Main results. The proposed spike detection algorithm achieved over 97% sensitivity and a smaller than 3% false detection rate. The MCU implementation occupies less than 3KB RAM and consumes 31.5μW/ch. The FPGA platform only occupies 299 logic cells and 3KB RAM for 128 channels and consumes 0.04μW/ch. Significance. On the spike detection algorithm front, we have eliminated the processing bottleneck by reducing the dynamic power consumption to lower than the hardware static power, without sacrificing detection performance. More importantly, we have explored the considerations in algorithm and hardware design with respect to scalability, portability, and costs. These findings can facilitate and guide the future development of real-time on-implant neural signal processing platforms.

The significance of neural inter-frequency power correlations

It is of great interest in neuroscience to determine what frequency bands in the brain have covarying power. This would help us robustly identify the frequency signatures of neural processes. However to date, to the best of the author's knowledge, a comprehensive statistical approach to this question that accounts for intra-frequency autocorrelation, frequency-domain oversampling, and multiple testing under dependency has not been undertaken. As such, this work presents a novel statistical significance test for correlated power across frequency bands for a broad class of non-stationary time series. It is validated on synthetic data. It is then used to test all of the inter-frequency power correlations between 0.2 and 8500 Hz in continuous intracortical extracellular neural recordings in Macaque M1, using a very large, publicly available dataset. The recordings were Current Source Density referenced and were recorded with a Utah array. The results support previous results in the literature that show that neural processes in M1 have power signatures across a very broad range of frequency bands. In particular, the power in LFP frequency bands as low as 20 Hz was found to almost always be statistically significantly correlated to the power in kHz frequency ranges. It is proposed that this test can also be used to discover the superimposed frequency domain signatures of all the neural processes in a neural signal, allowing us to identify every interesting neural frequency band.

Implantable brain machine interfaces: first-in-human studies, technology challenges and trends

Implantable brain machine interfaces (BMIs) are now on a trajectory to go mainstream, wherein what was once considered last resort will progressively become elective at earlier stages in disease treatment. First-in-human successes have demonstrated the ability to decode highly dexterous motor skills such as handwriting, and speech from human cortical activity. These have been used for cursor and prosthesis control, direct-to-text communication and speech synthesis. Along with these breakthrough studies, technology advancements have enabled the observation of more channels of neural activity through new concepts for centralised/distributed implant architectures. This is complemented by research in flexible substrates, packaging, surgical workflows and data processing. New regulatory guidance and funding has galvanised the field. This culmination of resource, efforts and capability is now attracting significant investment for BMI commercialisation. This paper reviews recent developments and describes the paradigm shift in BMI development that is leading to new innovations, insights and BMI translation.

Selection of Essential Neural Activity Timesteps for Intracortical Brain-Computer Interface Based on Recurrent Neural Network

Intracortical brain–computer interfaces (iBCIs) translate neural activity into control commands, thereby allowing paralyzed persons to control devices via their brain signals. Recurrent neural networks (RNNs) are widely used as neural decoders because they can learn neural response dynamics from continuous neural activity. Nevertheless, excessively long or short input neural activity for an RNN may decrease its decoding performance. Based on the temporal attention module exploiting relations in features over time, we propose a temporal attention-aware timestep selection (TTS) method that improves the interpretability of the salience of each timestep in an input neural activity. Furthermore, TTS determines the appropriate input neural activity length for accurate neural decoding. Experimental results show that the proposed TTS efficiently selects 28 essential timesteps for RNN-based neural decoders, outperforming state-of-the-art neural decoders on two nonhuman primate datasets (R2=0.76±0.05 for monkey Indy and CC=0.91±0.01 for monkey N). In addition, it reduces the computation time for offline training (reducing 5–12%) and online prediction (reducing 16–18%). When visualizing the attention mechanism in TTS, the preparatory neural activity is consecutively highlighted during arm movement, and the most recent neural activity is highlighted during the resting state in nonhuman primates. Selecting only a few essential timesteps for an RNN-based neural decoder provides sufficient decoding performance and requires only a short computation time.

Inferring entire spiking activity from local field potentials

Extracellular recordings are typically analysed by separating them into two distinct signals: local field potentials (LFPs) and spikes. Previous studies have shown that spikes, in the form of single-unit activity (SUA) or multiunit activity (MUA), can be inferred solely from LFPs with moderately good accuracy. SUA and MUA are typically extracted via threshold-based technique which may not be reliable when the recordings exhibit a low signal-to-noise ratio (SNR). Another type of spiking activity, referred to as entire spiking activity (ESA), can be extracted by a threshold-less, fast, and automated technique and has led to better performance in several tasks. However, its relationship with the LFPs has not been investigated. In this study, we aim to address this issue by inferring ESA from LFPs intracortically recorded from the motor cortex area of three monkeys performing different tasks. Results from long-term recording sessions and across subjects revealed that ESA can be inferred from LFPs with good accuracy. On average, the inference performance of ESA was consistently and significantly higher than those of SUA and MUA. In addition, local motor potential (LMP) was found to be the most predictive feature. The overall results indicate that LFPs contain substantial information about spiking activity, particularly ESA. This could be useful for understanding LFP-spike relationship and for the development of LFP-based BMIs.