Next generation brain-machine interfaces controlled synergistically with artificial intelligence

项目摘要 在美国，超过500万人-近50人中有1人-患有某种形式的瘫痪，原因包括 中风、脊髓损伤、多发性硬化和ALS。失去行动能力的人也会失去一个 在他们的生活中有着深刻的控制、自由和独立感。但是瘫痪并不会夺走一个人的 意图或欲望移动;大脑仍然通过神经信号编码这些想法。脑机接口 （BMI）旨在通过将这些神经信号转化为行动来恢复与世界沟通的能力。 BMI将神经信号解码成屏幕上计算机光标或机器人手臂的运动， 用户可以自主地与世界互动。 虽然BMI已经存在了二十多年，但它们仍然处于可追溯到2004年的试点临床试验中， 尚未得到广泛使用。其主要原因是BMI的表现没有达到 克服成本和风险的性能。这对于需要神经外科手术的两种侵入性BMI都是如此， 以及非侵入性BMI，其可以在没有外科手术的情况下使用，但性能较低。 这项提案旨在通过广泛使用， 在未来五年内可获得的BMIs。为了实现这一突破性的目标，需要有一个范式 BMI基本运作方式的转变。革命性的下一代BMI产生重大影响， 旨在：（1）实现当前BMI系统无法实现的前所未有的性能 (e.g.,超过一个数量级的改进）和（2）最小化患者的成本，理想地是非侵入性的。恩- 为了从根本上超越性能与成本之间的权衡，需要全新的BMI体系结构， 以更低的风险和成本获得卓越的性能。 为了满足这一需求，我建议开发下一代BMI，其中用户和阿尔蒂智能（AI） 代理协同合作。我们称之为“AI-BMI”。AI代理预测用户的预期运动动作 并且协同地帮助完成它们。重要的是，人工智能代理有助于精确和详细地执行 用户的预期动作，增强性能。通过这样做，这种架构从根本上改变了 BMI的设计目标从精确运动（困难）的神经解码到行为和神经解码 用户意图的推断（更容易）。非侵入性的AI-BMI如果成功，将是变革性的， 下一代BMI技术，使瘫痪的人再次移动，但减轻了医疗风险 与神经外科相关，并降低系统成本。

PROJECT SUMMARY Over 5 million people in the USA – nearly 1 person in 50 – live with a form of paralysis due to causes including stroke, spinal cord injury, multiple sclerosis, and ALS. People who have lost the ability to move also lose a profound sense of control, freedom, and independence in their lives. But paralysis does not take away one's intent or desire to move; the brain still encodes these thoughts through neural signals. Brain-machine interfaces (BMIs) aim to restore the ability to communicate with the world by translating these neural signals into actions. BMIs decode neural signals into the movements of a computer cursor on a screen or a robotic arm, allowing the user to interact with the world autonomously. While BMIs have existed for over two decades, they have remained in pilot clinical trials dating back to 2004 and have not achieved widespread use. The key reason for this is that BMI performance has not achieved levels of performance that overcome their costs and risks. This is true for both invasive BMIs requiring neurosurgery, as well as for non-invasive BMIs, which can be used without surgical procedures but achieve low performance. This proposal aims to dramatically increase the quality of life for millions with paralysis by making widely accessi- ble BMIs available within the next five years. To achieve this groundbreaking goal, there needs to be a paradigm shift in the way BMIs fundamentally operate. Revolutionary next generation BMIs making significant impact must be designed to: (1) achieve categorically unprecedented performance not possible with current BMI systems (e.g., over an order of magnitude improvement) and (2) minimize cost to patients, ideally being non-invasive. En- tirely novel BMI architectures are needed to fundamentally transcend a performance vs cost trade-off, achieving excellent performance at lower risks and costs. To meet this need, I propose to develop a next-generation BMI where the user and an artificial intelligence (AI) agent synergistically cooperate. We term this an “AI-BMI.” The AI agent predicts the user's intended motor actions and synergistically helps to complete them. Critically, the AI agent aids the precise and detailed execution of the user's intended movements, augmenting performance. By doing so, this architecture fundamentally changes the design objectives of BMIs from neural decoding of precise movements (difficult) to behavioral and neural inference of the user's intent (easier). Non-invasive AI-BMIs, if successful, would be transformative, enabling next generation BMI technology that allows people with paralysis to once again move, but mitigates medical risks associated with neurosurgery and lowers system costs.