CIF: Small: Advanced Ion Channel Models for Neurological Signal Processing -- Theory and Application to Brain-Computer Interfacing

研究人员正在研究大脑内神经元产生的电子噪声信号的科学理论和工程后果。他们正在对这种神经元噪声进行更深入的了解，主要是为了改善脑机接口（bci），从而提高各种瘫痪性残疾患者的生活质量，例如Lou Gehrig？年代的疾病。通过了解背景噪声的统计特征，他们能够减少其对瘫痪患者大脑产生的意识信号的干扰。这为轮椅、定制的互联网浏览器和家庭环境创造了一个更可靠的控制信号。此外，研究人员正在应用他们先进的神经元模型来深入了解未患病的脑组织及其产生和传递信息的过程。这有望对理解自然和合成神经网络的功能产生重大影响。具体来说，研究人员开发了一种新的神经元离子通道数学和随机模型，考虑了量子力学和热力学因素。在热平衡中，熵最大化原理应用于这些量子离子通道的种群，解释了所谓的1/f噪声在神经和脑电图（EEG）记录中无处不在的存在。从这些模型中得到的参数与一种新的信号处理范式相结合，称为倍频平均频谱整流，可以显着降低1/f噪声对脑电图信号的干扰影响。因此，一种称为稳态视觉诱发电位（SSVEP）脑接口的实验类型的刺激频率可以增加到30Hz以上的高伽马波段。这极大地减少了这些SSVEP脑机接口的负面副作用，并使它们首次实际应用于瘫痪患者，以及飞行员和外科医生的平视显示器，以及高性能游戏控制。

The investigators are studying the scientific theory and engineering consequences of the electrical noise signals generated by neurons inside the brain. They are developing a deeper understanding of this neuronal noise primarily to improve brain-computer interfaces (BCIs), which enhance the quality of life of patients with various paralyzing disabilities such as Lou Gehrig?s disease. By understanding the statistical characteristics of the background noise, they are able to reduce its interference with the conscious signal generated by the brain of the paralyzed patient. This creates a much more reliable control signal for wheelchairs, customized internet browsers, and the home environment. In addition, the investigators are applying their advanced neuron models to gain insight into non-diseased brain tissue and the processes by which it creates and transmits information. This promises to have significant implications for understanding the functioning of both natural and synthetic neural networks.Specifically, the investigators have developed a novel mathematical and stochastic model for neuronal ion channels that takes into account quantum mechanical and thermodynamic considerations. The principle of entropy maximization applied to a population of these quantum ion channels in thermal equilibrium explains the ubiquitous presence of the so-called 1/f-noise in neural and electroencephalographic (EEG) recordings. The parameters derived from these models are combined with a new signal processing paradigm called octave-averaged spectral rectification to dramatically reduce the interfering effect of 1/f-noise on EEG signals. As a result, the stimulus frequencies of an experimental type of brain interface called steady-state visual evoked potential (SSVEP) BCIs can be increased into the high-gamma band above 30Hz. This dramatically reduces the negative side effects of these SSVEP BCIs and makes them of practical use for the first time to paralyzed patients as well as heads-up displays for pilots and surgeons, and high-performance game control.