We present a new deep multi-state Dynamic Recurrent Neural Network (DRNN) architecture for Brain Machine Interface (BMI) applications. Our DRNN is used to predict Cartesian representation of a computer cursor movement kinematics from open-loop neural data recorded from the posterior parietal cortex (PPC) of a human subject in a BMI system. We design the algorithm to achieve a reasonable trade-off between performance and robustness, and we constrain memory usage in favor of future hardware implementation. We feed the predictions of the network back to the input to improve prediction performance and robustness. We apply a scheduled sampling approach to the model in order to solve a statistical distribution mismatch between the ground truth and predictions. Additionally, we configure a small DRNN to operate with a short history of input, reducing the required buffering of input data and number of memory accesses. This configuration lowers the expected power consumption in a neural network accelerator. Operating on wavelet-based neural features, we show that the average performance of DRNN surpasses other state-of-the-art methods in the literature on both single-and multi-day data recorded over 43 days. Results show that multi-state DRNN has the potential to model the nonlinear relationships between the neural data and kinematics for robust BMIs.

Despite extensive theoretical work on biologically plausible learning rules, clear evidence about whether and how such rules are implemented in the brain has been difficult to obtain. We consider biologically plausible supervised-and reinforcement-learning rules and ask whether changes in network activity during learning can be used to determine which learning rule is being used. Supervised learning requires a credit-assignment model estimating the mapping from neural activity to behavior, and, in a biological organism, this model will inevitably be an imperfect approximation of the ideal mapping, leading to a bias in the direction of the weight updates relative to the true gradient. Reinforcement learning, on the other hand, requires no credit-assignment model and tends to make weight updates following the true gradient direction. We derive a metric to distinguish between learning rules by observing changes in the network activity during learning, given that the mapping from brain to behavior is known by the experimenter. Because brain-machine interface (BMI) experiments allow for precise knowledge of this mapping, we model a cursor-control BMI task using recurrent neural networks, showing that learning rules can be distinguished in simulated experiments using only observations that a neuroscience experimenter would plausibly have access to.

In this paper, the authors present a multi-state Dynamic Recurrent Neural Network architecture and training framework for Brain Machine Interface (BMI), including incorporating scheduled sampling and testing diverse neural features as input. The authors robustly analyze this model in comparison to other prior modeling frameworks on human posterior parietal cortical activity (PPC). This paper is of an impressive quality, containing rigorous and methodical analyses showing clear and significant improvements of their model. The authors compare to twelve baseline models and investigate many aspects of the modeling framework, including single-day vs multi-day performance, generalization of single-day training to other days, the reliance on amount of training data, the optimal preprocessing of neural feature inputs, and generalization of the models over time with different styles of retraining. The paper was very well-written, with most choices and details clearly explained.

Despite extensive theoretical work on biologically plausible learning rules, clear evidence about whether and how such rules are implemented in the brain has been difficult to obtain. We consider biologically plausible supervised- and reinforcement-learning rules and ask whether changes in network activity during learning can be used to determine which learning rule is being used. Supervised learning requires a credit-assignment model estimating the mapping from neural activity to behavior, and, in a biological organism, this model will inevitably be an imperfect approximation of the ideal mapping, leading to a bias in the direction of the weight updates relative to the true gradient. Reinforcement learning, on the other hand, requires no credit-assignment model and tends to make weight updates following the true gradient direction. We derive a metric to distinguish between learning rules by observing changes in the network activity during learning, given that the mapping from brain to behavior is known by the experimenter.

We present a new deep multi-state Dynamic Recurrent Neural Network (DRNN) architecture for Brain Machine Interface (BMI) applications. Our DRNN is used to predict Cartesian representation of a computer cursor movement kinematics from open-loop neural data recorded from the posterior parietal cortex (PPC) of a human subject in a BMI system. We design the algorithm to achieve a reasonable trade-off between performance and robustness, and we constrain memory usage in favor of future hardware implementation. We feed the predictions of the network back to the input to improve prediction performance and robustness. We apply a scheduled sampling approach to the model in order to solve a statistical distribution mismatch between the ground truth and predictions.

Soldiers controlling a robot dog with their mind as they patrol a dusty road and sweep an delipidated building may sound like science fiction, but it is the scene in a real world demonstration. The Australian Army has perfected mind-controlling abilities with eight sensors neatly packed inside a helmet that work in tandem with a Microsoft HoloLens. The innovation features an AI-decoder that translates a soldier's brain signals into explainable instructions that are sent to the robotic quadruped, allowing humans to stay focused on their surroundings. A new video shows military personal conducting a simulated patrol clearance using the robot dog, which was instructed to sweep a facility using what it read from a person's brain waves - and with 94 percent accuracy. The system was developed by the University of Technology Sydney that first unveiled the innovation last year, but recently published a new paper detailing the work. 'The user used our augmented brain–robot interface (aBRI) platform to control the robot systems,' reads the paper published by American Chemical Society on March 16.

We present a new deep multi-state Dynamic Recurrent Neural Network (DRNN) architecture for Brain Machine Interface (BMI) applications. Our DRNN is used to predict Cartesian representation of a computer cursor movement kinematics from open-loop neural data recorded from the posterior parietal cortex (PPC) of a human subject in a BMI system. We design the algorithm to achieve a reasonable trade-off between performance and robustness, and we constrain memory usage in favor of future hardware implementation. We feed the predictions of the network back to the input to improve prediction performance and robustness. We apply a scheduled sampling approach to the model in order to solve a statistical distribution mismatch between the ground truth and predictions.

A confluence of technological capabilities is creating an opportunity for machine learning and artificial intelligence (AI) to enable "smart" nanoengineered brain machine interfaces (BMI). The goal is for this new generation of technologies to be able to communicate with the brain in ways that support contextual learning and adaptation to changing functional requirements. This applies to both invasive technologies aimed at restoring neurological function, as in the case of neural prosthesis, as well as non-invasive technologies enabled by signals such as electroencephalograph (EEG). Advances in computation, hardware, and algorithms that learn and adapt in a contextually dependent way will be able to leverage the capabilities that nanoengineering offers the design and functionality of BMI. Eventually, these technologies will be able to carry out learning and adaptation in (near) real time, as external shifting demands from the environment and physiology require them.

Elon Musk announced late Tuesday night that the final goal of Neuralink, his brain-machine interface startup, is to allow humans to "achieve a symbiosis with artificial intelligence," and that by "merging with AI," humans will be able to keep up with AI. Musk plans to begin human trials on an early version of Neuralink intended to treat brain injuries next year. "Ultimately we can do a full brain machine interface," Musk said in an announcement that was widely livestreamed. "This is going to sound pretty weird. Ultimately we can achieve a symbiosis with artificial intelligence. This is not a mandatory thing, this is something you can choose to have if you want. This is going to be really important at a civilization-level scale. Even in a benign AI scenario, we will be left behind. With a high-bandwidth brain machine interface we can go along for the ride and have the option of merging with AI." Musk has become famous for his moonshot projects, his lofty promises, his quick temper on Twitter, and his various plans for society that don't include input from the rest of us.