Researchers at Ecole Polytechnique Fédérale de Lausanne in Switzerland have designed an advanced neural chip that can detect and suppress symptoms from a variety of neurological disorders, including Parkinson’s and epilepsy. The closed-loop neuromodulation system, which the researchers have called NeuralTree, includes soft implantable electrodes, a processor for machine learning, and a 256 channel sensing array. The device is also energy efficient, helping to extend battery life. The technology can spot the signs of upcoming tremors or seizures, for example, and initiate neurostimulation to reduce or avoid the symptoms.
Neuromodulation offers enormous hope for those with neurological disorders, and the technology is developing apace. The concept of an implanted chip that can reduce or negate neurological symptoms before they arise is like something from a science fiction novel, but here we are. This latest neural chip is unique in that it can address the symptoms of several neurological disorders and boasts some advanced capabilities.
“NeuralTree benefits from the accuracy of a neural network and the hardware efficiency of a decision tree algorithm,” said Mahsa Shoaran, one of the lead developers of the new device. “It’s the first time we’ve been able to integrate such a complex, yet energy-efficient neural interface for binary classification tasks, such as seizure or tremor detection, as well as multi-class tasks such as finger movement classification for neuroprosthetic applications.”
The chip monitors brain waves, looking for signs of an upcoming neurological event, such as a seizure. Once it has identified such a neurological biomarker, a neurostimulator in the chip sends an electrical pulse through the implanted electrodes to block the aberrant activity. The chip includes 256 input channels, which is significantly more than the 32 that previous similar devices permitted.
The chip is also tiny at 3.48 mm2 and its algorithm does not prioritize power-intensive processes, helping to save battery life. Previous devices have typically focused on treating epileptic seizures, but the researchers behind NeuralTree have also trained the machine learning algorithms to recognize neural signals that herald an impending tremor episode in Parkinson’s patients.
“To the best of our knowledge, this is the first demonstration of Parkinsonian tremor detection with an on-chip classifier,” said Shoaran. “Eventually, we can use neural interfaces for many different disorders, and we need algorithmic ideas and advances in chip design to make this happen. This work is very interdisciplinary, and so it also requires collaborating with labs like the Laboratory for Soft Bioelectronic Interfaces, which can develop state-of-the-art neural electrodes, or labs with access to high-quality patient data.”
Study in journal IEEE Journal of Solid-State Circuits: NeuralTree: A 256-Channel 0.227-μJ/Class Versatile Neural Activity Classification and Closed-Loop Neuromodulation SoC
Via: EPFL