A wireless closed-loop sleep modulation system-on-chip

无线闭环睡眠调制片上系统

• 批准号: 10733872
• 负责人: Xilin Liu
• 金额: $ 46.24万
• 依托单位: UNIVERSITY OF PENNSYLVANIA
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-06-01 至 2027-05-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10733872
• 关键词: Acceleration; Acoustic Stimulation; Address; Affect; Agreement; Amplifiers; Animal Model; Animals; Auditory; Behavior; Biological Models; Bite; Brain; Canis familiaris; Charge; Classification; Communities; Data Analyses; Data Compression; Devices; Disease; Electric Stimulation; Electrophysiology (science); Eye; Family suidae; Felis catus; Health; Home; Housing; Human; Immune; Intervention; Investigation; Label; Learning Module; Light; Macaca; Machine Learning; Metabolic; Modeling; Modification; Monkeys; Mus; Muscle; Neurodegenerative Disorders; Neurons; Noise; Optics; Performance; Phase; Physiological; Physiology; Play; Polysomnography; Process; Qualifying; REM Sleep; Radio; Rattus; Research; Rodent; Role; Running; Scientist; Sensory; Signal Transduction; Silicon; Sleep; Sleep Architecture; Sleep Disorders; Sleep Stages; Specificity; System; Techniques; Technology; Testing; Time; Wakefulness; Weight; algorithm training; analog; cognitive benefits; data acquisition; data exchange; data reduction; design; digital; energy efficiency; experience; experimental study; free behavior; hypocretin; improved; in vivo; in vivo evaluation; industry partner; long term memory; microelectronics; microsystems; miniaturize; multimodality; neural; neural circuit; neurotransmission; new technology; non rapid eye movement; nonhuman primate; noradrenergic; novel therapeutic intervention; novel therapeutics; operation; optogenetics; prevent; rapid eye movement; sleep behavior; sleep regulation; wireless; wireless communication; wireless electronic

PROJECT SUMMARY Sleep plays a critical role in a vast array of physiological and pathophysiological processes, including executive brain functions, immune and metabolic functions, and neurodegenerative diseases. An improved understanding of sleep mechanisms is important to developing new treatment strategies. Sleep circuits can be manipulated with numerous techniques to infer causal relationships to physiology and behavior. Effects of these manipulations are generally sleep stage or brain state dependent. Thus, closed-loop interrogation provides a powerful paradigm, whereby a temporally precise manipulation (optogenetic, electrical, or sensory stimulation) is delivered in a brain-state dependent manner. This paradigm has been used successfully in mice and humans tethered to data acquisition, state classification, and stimulation hardware. However, tethering can adversely affect natural sleep behavior and prevents use in larger animal models. Here, we propose to develop a fully integrated, wireless, sleep modulation system-on-chip (SoC) with the capability to deliver state-dependent, temporally-precise optical, electrical, and auditory stimulation. In Aims 1 and 2, Co-PI Dr. Liu will iteratively develop an ultra-low-power SoC for polysomnography (PSG) signal acquisition and multi-modal closed-loop stimulation. The final SoC will have five modules: (1) a low-noise, high-precision PSG acquisition module, (2) a programmable mixed-signal data compression module for energy efficiency, (3) a low-power ultra-wideband wireless transmitter, (4) a machine learning-based sleep stage classification module, and (5) a fully programmable multi-modal stimulator. Full integration of the system will result in a small (2 cm3), light (2 g) battery-powered device able to operate continuously for over 10 h on a single charge. In Aim 3, running concurrently with the first two aims, Co-PI Dr. Richardson will validate performance of the SoC with sleep studies in both small (rats) and large (macaques) animal models. Interleaved recordings from the wireless SoC and a tethered commercial system will assess fidelity of compressed PSG acquisition and the impact of tethering on sleep architecture. Real-time 2-stage and all-stage classifiers will be compared to offline ground truth labels. Finally, in-phase closed-loop auditory stimulation will be used to enhance slow wave activity. Our team is uniquely qualified for this proposal since we have extensive experience developing the key modules in the proposed SoC and using wireless electronics for free behavior electrophysiological experiments in both animal models. Once developed, two industry partners, CMC Microsystems and Open Ephys, Inc., will facilitate distribution of the SoC to the worldwide community. The closed-loop SoC will enable unprecedented hypothesis- driven research on sleep-stage specific neural circuits and interventions in models spanning from mice to monkeys, thereby accelerating the mechanistic understanding of sleep and treatment of sleep disorders.

项目总结 睡眠在大量的生理和病理生理过程中起着关键作用，包括执行 脑功能、免疫和代谢功能以及神经退行性疾病。更好的理解 睡眠机制的研究对于开发新的治疗策略非常重要。休眠电路可以被操纵 有许多技术可以推断生理和行为之间的因果关系。这些措施的影响 操作通常取决于睡眠阶段或大脑状态。因此，闭环询问提供了一种 强大的范例，借此进行时间上精确的操作(光遗传、电刺激或感觉刺激) 是以依赖于大脑状态的方式传递的。这一范例已经成功地在老鼠和人类身上使用。 捆绑到数据采集、状态分类和刺激硬件。然而，系留可能会对 影响自然睡眠行为，并阻止在较大的动物模型中使用。在这里，我们建议开发一种完全 集成的无线睡眠调制片上系统(SoC)，能够提供依赖状态的 时间上精确的光、电和听觉刺激。在目标1和目标2中，共同派刘博士将迭代 一种用于多导睡眠图信号采集和多模式闭环的超低功耗SoC设计 刺激。最终的SoC将包括五个模块：(1)低噪声、高精度的PSG采集模块；(2) 可编程混合信号能效数据压缩模块，(3)低功耗超宽带 无线发射机，(4)基于机器学习的睡眠阶段分类模块，和(5)完整的 可编程多模式刺激器。该系统的完全集成将产生一个小(2厘米3)、轻(2克) 电池供电的装置，一次充电可连续工作10小时以上。在Aim 3中，跑步 在前两个目标的同时，联合PI理查森博士将通过睡眠研究来验证SoC的性能 在小(大鼠)和大(猕猴)动物模型中。来自无线SoC的交错录制和 系留商业系统将评估压缩PSG采集的保真度以及系留对 睡眠架构。实时两阶段和所有阶段的分类器将与离线地面事实标签进行比较。 最后，将使用同相闭环听觉刺激来增强慢波活动。我们的团队是 独一无二地符合这项建议，因为我们在开发 提出SOC并使用无线电子学进行两种动物的自由行为电生理实验 模特们。一旦开发完成，CMC MicroSystems和Open Ephys，Inc.这两个行业合作伙伴将促进 将SoC分发给世界各地的社区。闭环系统芯片将实现前所未有的假设- 从小鼠到小鼠的睡眠阶段特异性神经回路和干预的驱动研究 猴子，从而加速了对睡眠的机械性理解和睡眠障碍的治疗。

项目成果

Design of a Sleep Modulation System with FPGA-Accelerated Deep Learning for Closed-loop Stage-Specific In-Phase Auditory Stimulation.

采用 FPGA 加速深度学习的睡眠调制系统设计，用于闭环特定阶段同相听觉刺激。

• DOI: 10.1109/iscas46773.2023.10181356
• 发表时间: 2023
• 期刊: IEEE International Symposium on Circuits and Systems proceedings. IEEE International Symposium on Circuits and Systems
• 作者: Sun,Mingzhe;Zhou,Aaron;Yang,Naize;Xu,Yaqian;Hou,Yuhan;Richardson,AndrewG;Liu,Xilin
• 通讯作者: Liu,Xilin