Collaborative Research: Integrated memristor neural networks for in-situ analysis of intracellular neuronal recordings

合作研究：用于细胞内神经元记录原位分析的集成忆阻器神经网络

• 批准号: 1915550
• 负责人: Wei Lu
• 金额: $ 24.1万
• 依托单位: Regents of the University of Michigan - Ann Arbor
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-06-01 至 2023-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1915550&HistoricalAwards=false
• 关键词: Collaborative; Research; Integrated; memristor; neural

The ability to detect and analyze individual neuron spiking patterns over large areas will have profound impacts on neuroscience and medicine, as it allows the functions of the brain to be directly mapped to underlying neuron activities and enables precise brain disease detection and drug developments. However, although recent advances in neuroprobes such as nanoelectrode arrays make it possible to pick up signals from single neurons, scaling the systems to thousands and possibly millions of neuron sites is not practical due the enormous resources and time required to digitize, store and analyze the vast amount of data. This proposal aims to precisely address these challenges by integrating an artificial neural network on the nanoelectrode array, such that cell signals picked up by the electrodes are directly processed by the artificial neural network, and only the processed, "useful" data need to be amplified and transmitted, allowing real-time analysis with very low power. Undergraduate and graduate students will be trained to obtain state-of-the-art nanotechnology and neuroengineering techniques. Knowledge and techniques developed during research will be incorporated into course materials and other types of publications to allow cross pollination for students among different disciplines and broad dissemination to the general public. The proposed new neural recording system will offer unparalleled spatial resolution and processing capabilities, where high density and highly sensitive nanoelectrode arrays are integrated with a memristor-based artificial neural network that allows real time signal processing. By optimizing and utilizing internal dynamic ionic processes in the memristors, the artificial network can be directly driven by the spike trains from biological neurons without amplification or other pre-processing, where responses from the memristor network can be used to analyze temporal patterns in the neuron spikes and to connect detected neuronal activity with functions of the biological networks. The tightly coupled memristor network with the biological network can further allow functions in the biological system to be directly mapped on the electrical system, and potentially lead to future neural prosthesis and augmentation applications. New materials, devices, and networks will be developed, along with new recording and computing strategies that broaden the impact of the proposed project.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

检测和分析大面积单个神经元放电模式的能力将对神经科学和医学产生深远影响，因为它允许大脑功能直接映射到潜在的神经元活动，并使精确的脑部疾病检测和药物开发成为可能。然而，尽管纳米电极阵列等神经探针的最新进展使从单个神经元获取信号成为可能，但将系统扩展到数千个甚至可能数百万个神经元站点是不切实际的，因为数字化、存储和分析大量数据需要巨大的资源和时间。这一提议旨在通过在纳米电极阵列上集成人工神经网络来精确地应对这些挑战，使电极拾取的细胞信号直接由人工神经网络处理，只需放大和传输处理后的“有用”数据，从而能够以非常低的功率进行实时分析。本科生和研究生将接受培训，以获得最先进的纳米技术和神经工程技术。在研究过程中开发的知识和技术将被纳入课程材料和其他类型的出版物，使不同学科的学生能够相互授粉，并向公众广泛传播。拟议的新神经记录系统将提供无与伦比的空间分辨率和处理能力，其中高密度和高灵敏度的纳米电极阵列与基于忆阻器的人工神经网络集成在一起，允许实时信号处理。通过优化和利用忆阻器中的内部动态离子过程，人工网络可以直接由生物神经元的棘波序列驱动，而不需要放大或其他预处理，其中来自忆阻器网络的响应可以用来分析神经元棘波的时间模式，并将检测到的神经元活动与生物网络的功能联系起来。与生物网络紧密耦合的忆阻器网络可以进一步允许生物系统中的功能直接映射到电子系统上，并可能导致未来的神经假体和增强应用。将开发新的材料、设备和网络，以及扩大拟议项目影响的新记录和计算策略。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Memristor networks for real-time neural activity analysis

• DOI: 10.1038/s41467-020-16261-1
• 发表时间: 2020-05
• 期刊: Nature Communications
• 影响因子: 16.6
• 作者: Xiaojian Zhu;Qiwen Wang;Wei D. Lu

Spatiotemporal Spike Pattern Detection with Second-order Memristive Synapses

使用二阶忆阻突触进行时空尖峰模式检测

• DOI: 10.1109/iscas48785.2022.9937414
• 发表时间: 2022
• 期刊: IEEE International Symposium on Circuits and Systems (ISCAS
• 作者: Wu, Yuting;Yoo, Sangmin;Meng, Fan-Hsuan;Lu, Wei D.
• 通讯作者: Lu, Wei D.

Hierarchical architectures in reservoir computing systems

• DOI: 10.1088/2634-4386/ac1b75
• 发表时间: 2021-05
• 期刊: Neuromorphic Computing and Engineering
• 作者: John Moon;Yuting Wu;Wei D. Lu

Neural Functional Connectivity Reconstruction with Second‐Order Memristor Network

• DOI: 10.1002/aisy.202000276
• 发表时间: 2021-05
• 期刊: Advanced Intelligent Systems
• 影响因子: 7.4
• 作者: Yuting Wu;John Moon;Xiaojian Zhu;W. Lu

Neural connectivity inference with spike-timing dependent plasticity network

使用尖峰时序相关可塑性网络进行神经连接推理

• DOI: 10.1007/s11432-021-3217-0
• 发表时间: 2021
• 期刊: Science China Information Sciences
• 作者: Moon, John;Wu, Yuting;Zhu, Xiaojian;Lu, Wei D.
• 通讯作者: Lu, Wei D.