SenSE: Multimodal Biometric Sensor for Optimal Regulation of Circadian Rhythm and Neurocognitive Performance

昼夜节律是由一个内部的生物钟来调节的，这个生物钟用每天的光和暗模式来控制我们的生物过程。它调节睡眠，新陈代谢，激素分泌和影响警觉性和工作效率的神经行为过程。昼夜节律的破坏对健康有负面影响。现代生活方式对维持健康的昼夜节律调节提出了挑战，例如在夜间暴露于强光下，以及最近的在家工作情况，模糊了工作和个人时间之间的界限。这个问题可以通过智能和互联的建筑环境来解决，这些环境可以促进个人的昼夜健康和工作效率。该项目的目标是获得可靠的数学模型，捕捉个人昼夜节律的动态和估计昼夜节律和相关神经行为过程的状态的能力。该项目将构建可穿戴硬件和软件，以从嘈杂的生物特征信号中提取有用的信息，这些信息可用于构建上述数学模型和状态估计。该软件将被部署为智能手机应用程序，该应用程序将使用这些信息，与人类用户进行交互，并为睡眠，照明和任务安排提供最佳建议，以保持健康的昼夜节律并优化工作效率。在教育方面，该项目将支持对研究生和本科生进行多学科研究的培训，将新的教学材料纳入工程设计课程，并开展外联活动，以提高目前在STEM领域代表性不足的学生群体对STEM的兴趣。 评估昼夜节律系统状态的临床标准是通过测量生物标志物，例如参与昼夜节律调节的激素浓度。这样的程序对于闭环反馈系统中的在线使用是不切实际的。 现成的可穿戴设备只能部分满足在线个性化生物特征测量的需求，因为它们只能测量有限的一组信号，并且不包括蓝光暴露的关键测量。在该项目中，研究人员将开发可穿戴传感器设备，这些设备（1）具有能量收集能力，（2）测量间接的昼夜节律相位标记，如活动记录仪，体温，心率/脉搏率变化，血压，ECG和EEG，以及（3）可以解析受试者暴露于蓝光的光谱内容。研究人员将开发信号处理算法，用于本项目中开发的各种传感模式的嘈杂异构生物特征数据。处理后的信号将用于昼夜节律系统和相关神经认知过程的模型识别和状态估计。状态估计算法将使用来自控制理论和机器学习的工具，并将无模型和基于模型的方法联合收割机结合起来，以实现对噪声和数据丢失的鲁棒性。该项目中开发的硬件和软件的有效性将在受控的实验室实验中进行评估，并评估健康人类受试者在昏暗光线下褪黑激素的发病情况。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

Circadian rhythms are regulated by an internal biological clock that synchronizes our biological processes with the daily light and dark pattern. It regulates sleep, metabolism, hormone secretion, and neurobehavioral processes that impact alertness and work productivity. Disruption of circadian rhythms has negative impacts on health. Modern lifestyle poses challenges in maintaining healthy circadian regulation, such as exposure to bright light during nighttime and, more recently, the working-from-home situation that blurs the boundary between work and personal time. This problem can be addressed by involving smart and connected built environments that promote individual circadian health and work productivity. The goal of this project is to obtain reliable mathematical models that capture the dynamics of individual circadian rhythms and the ability to estimate the state of circadian rhythms and related neurobehavioral processes. This project will build wearable hardware and software to extract useful information from noisy biometric signals that can be used in building the above-mentioned mathematical model and state estimation. The software will be deployed as a smartphone app that will use this information, interface with the human users, and provide optimal recommendations for sleep, lighting, and task schedules to maintain healthy circadian rhythms and optimize work productivity. On the educational front, the project will support the training of graduate and undergraduate students in multidisciplinary research, integration of new pedagogical material into the engineering design curriculum, and outreach activities to raise interest in STEM among student populations that are currently underrepresented in STEM fields. The clinical standard for assessing the state of the circadian system is by measuring biomarkers such as the concentration of hormones that participate in circadian rhythm regulation. Such procedures are impractical for online use in a closed-loop feedback system. Off-the-shelf wearable devices can only partially fill the need for online personalized biometric measurements because they only measure a limited set of signals and exclude the critical measurement on blue light exposure. In this project, the investigators will develop wearable sensor devices that (1) have energy harvesting capability, (2) measure indirect circadian phase markers such as actigraphy, body temperature, heart/pulse rate variation, blood pressure, ECG, and EEG, and (3) can resolve the spectral content of blue light exposure to the subject. The investigators will develop signal processing algorithms for noisy heterogeneous biometric data from various sensing modalities developed in this project. The processed signals will be used in model identification and state estimation of the circadian system and related neurocognitive processes. The state estimation algorithms will use tools from control theory and machine learning and combine model-free and model-based approaches to achieve robustness to noise and data dropouts. The validity of the hardware and software developed in this project will be evaluated in controlled in-lab experiments with the assessment of dim light melatonin onset in healthy human subjects.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.