CAREER: Robust Identification and Multi-Objective Control Methods for Neuronal Networks Under Uncertainty

职业：不确定性下神经网络的鲁棒识别和多目标控制方法

基本信息

批准号：
1845348
负责人：
Sabato Santaniello
金额：
$ 50万
依托单位：
University of Connecticut
依托单位国家：
美国
项目类别：
Continuing Grant
财政年份：
2019
资助国家：
美国
起止时间：
2019-03-01 至 2025-02-28
项目状态：
未结题

来源：
https://www.nsf.gov/awardsearch/showAward?AWD_ID=1845348&HistoricalAwards=false
关键词：
CAREER Robust Identification Multi Objective

项目摘要

Overview: The control of biological neural networks underpins the development of minimally-invasive brain therapies as well as adaptive learning for cyber-physical systems, but it remains challenging because of the irregular dynamics involved. These systems also have sparse and weak connections and a range of dynamics that cannot be fully probed. There is an urgent need to determine the impact of unmodeled dynamics on the controllability of these networks and develop robust controls accordingly, otherwise controllers will remain underperforming, fragile, and hard to calibrate. This is the case for deep brain stimulation (DBS), which follows a conservative "one-size-fits-all" paradigm and remains underutilized despite having the potential to treat millions of people worldwide. The objective of this CAREER program is to develop identification methods that estimate the impact of unmodeled dynamics on neuronal circuits and a robust control framework for these circuits. Brain circuits targeted by Parkinson's disease and DBS will be considered to maximize the impact of the research. The work will be paired with educational plans that address current limitations in the training of neural engineers and broaden the presence of first-generation college students in STEM.Intellectual Merits: This research will fill critical gaps in the knowledge base that provides linkage between global dynamics of a neural network and dynamics of individual neurons under control. It will also contribute a robust control framework for neural populations and brain circuits. Educational activities will fill critical gaps in the training of neural engineers by integrating modeling and control in the design process of neural prostheses. Applied to DBS, this research will help personalize DBS to PD populations who are currently excluded from this treatment, thus enabling new options for chronically ill patients.Broader Impacts: The research will benefit the well-being of Parkinson's disease patients, including patients who are now excluded from DBS. The training of neural engineers will also be improved, thus helping the formation of a better-trained and more globally-competitive workforce. The integration of research and outreach will finally create a pipeline to attract high-school students towards STEM fields and facilitate the learning of engineering principles at the pre-college level.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.

概述：生物神经网络的控制基于最小侵入性脑疗法的发展以及对网络物理系统的自适应学习，但由于涉及不规则的动态，它仍然具有挑战性。这些系统还具有稀疏和弱的连接以及无法完全探测的一系列动态。迫切需要确定未建模动力学对这些网络的可控性的影响，并相应地开发出强大的控件，否则控制器的表现不佳，脆弱且难以校准。深度大脑刺激（DBS）就是这种情况，它遵循保守的“千篇一律”范式，尽管有可能在全球范围内对待数百万人，但仍未得到充分利用。该职业计划的目的是开发识别方法，以估算未模拟动态对神经元电路的影响以及这些电路的强大控制框架。帕金森氏病和DBS靶向的脑电路将被考虑最大化研究的影响。这项工作将与教育计划配对，该计划应对培训神经工程师的当前局限神经网络和正在控制的个别神经元的动态。它还将为神经种群和脑电路提供强大的控制框架。教育活动将通过在神经假体的设计过程中整合建模和控制来填补神经工程师培训的关键空白。该研究适用于DBS，将有助于对目前被排除在此治疗中排除的PD种群中个性化DBS，从而为长期患者提供了新的选择。BRODER的影响：这项研究将使帕金森氏病患者的福祉受益现在排除在DBS之外。神经工程师的培训也将得到改善，从而有助于形成一个训练有素，全球更具竞争力的劳动力。研究和外展的整合最终将建立一条管道，以吸引高中生进入STEM领域，并促进大学前级别的工程原则学习。该奖项反映了NSF的法定任务，并被认为是值得通过使用评估的支持。基金会的智力优点和更广泛的影响审查标准。