Model-based training for BCI rehabilitation

描述(由申请人提供)：本申请涉及广泛的挑战领域：使能技术，挑战主题=06-HD-101，假肢的改进接口，以影响康复结果。在解码运动意图和记录神经活动模式群体的技术方面的最新进展，创造了神经科学中最令人兴奋的研究领域之一，因为这项工作有望为瘫痪的人恢复丢失的运动输出。大脑控制界面的成功关键取决于用户以特定方式调节神经活动的能力。到目前为止，受试者接受的培训方法都是特别的。然而，随着更复杂的设备的开发，同时改进学习技术的需求变得至关重要。基于我们在最先进的脑驱动假体方面的成功，我们提出了一种新颖的、模型驱动的训练方法。有了这些界面，所有的行为都是由神经活动直接驱动的，我们有一个无与伦比的机会来操纵任务难度和监控表现。这将允许我们对学习过程进行建模，使用一种新的操作员-计算机共享控制算法，并定义设置点来驱动学习，以最大限度地减少掌握复杂的脑控假肢设备所需的训练量。这些方法很可能推广到瘫痪患者的运动康复。大脑控制的假肢在记录的神经活动和外部设备的运动之间提供了直接的联系。受试者必须学会调节他们的神经活动模式，才能成功地控制这些设备。这项建议概述了一种新的方法，以促进一种学习方法，导致控制复杂的，大脑控制的接口，可以推广到广泛的治疗康复。

DESCRIPTION (provided by applicant): This application addresses broad Challenge Area: Enabling Technologies, Challenge Topic =06-HD-101, Improved Interfaces for Prostheses to Impact Rehabilitation Outcomes. Recent advances in decoding motor intentions and the technology to record populations of neural activity patterns, has created one of the most exciting research areas in neuroscience because this work holds the promise of restoring lost motor output to those who are paralyzed. The success of brain-controlled interfaces depends critically on the user's ability to modulate neural activity in a specific way. To date, subjects are trained with methods that are ad hoc. However, as more sophisticated devices are developed, the need for concurrent improvements in learning techniques is becoming critical. Based on our success with state-of-the-art brain-driven prosthetics, we are proposing a novel, model-driven training approach. With these interfaces, all behavior is driven directly by neural activity and we have an unparalleled opportunity to manipulate task difficulty and monitor performance. This will allow us to model the learning process, use a novel operator-computer shared-control algorithm and define set points to drive learning in a way that minimizes the amount of training needed to master control of complex, brain-controlled prosthetic devices. These methods are likely to generalize to motor rehabilitation of paralyzed individuals. Brain-controlled prosthetics provide a direct link between recorded neural activity and the movement of external devices. Subjects must learn to modulate their neural activity patterns to control these devices successfully. This proposal outlines a novel method to facilitate a learning method leading to the control of complex, brain-controlled interfaces that may be generalized to a wide range of therapeutic rehabilitation.