Understanding Behavioral Variability in Outcome After SCI

项目总结 现在有机会实现健康管理向个性化理疗的范式转变- 行为(生物识别)监控-使用可穿戴设备预测、预防和更好地管理疾病 技术，以及轮椅和家庭中嵌入的传感器技术。 我们的主要目标是解释在相同的生物特征变量中收集的组合变化 在自然行为的小鼠脊髓损伤的进展过程中无创。在控制良好的动物研究中，我们 建议应用机器学习算法来识别对疾病有预测作用的“数字生物签名” 出现和/或表达，因此在基于反馈的缓解中使用。为了实现这一目标，我们有 专门设计的专用仪表式鼠笼，带有市售传感器，可实现连续 长期无创的家庭笼子捕捉这些生物特征，以开发这样的数字原型 生物签名。 重点是了解睡眠障碍、神经性疼痛、 术后体温调节功能障碍、心肺功能障碍和自主神经危机(自主神经反射障碍) SCI。因此，基于家笼传感器的捕获包括所有运动事件、呼吸、心率、3状态睡眠、 皮肤温度热图和感官偏好测试。 我们的总体假设是，在脊髓损伤的发展过程中，联合连续捕获了几个变量 将识别与新出现的功能障碍有关的新的“数字生物签名”。更长期的目标是将 将数字生物签名捕获到基于实时反馈的方法中，以限制疾病表达。 两个SCI模型将被用来量化紧急功能障碍的变异性和时间相关性。 生物测量的改变：[1]T9-10挫伤脊髓损伤和[2]T2-3完全横断。对两个人都是 实验系列，变量将在位于以下位置的专用仪表式家笼中持续捕获 脊髓损伤或假手术前和术后10周的环境控制室。已被捕获 生物识别技术将被进一步归类为机器学习，其基础是从 将非侵入性生物测定数字生物签名与感觉和自主神经功能障碍联系起来的更多常规测试 用既定的措施治疗脊髓损伤后的生理行为障碍。 如果成功，实时捕获功能障碍的数字生物签名可能会对 脊髓损伤患者的个体化用药。这是因为获得的生物签名随后可能会提供 从嵌入式/可穿戴传感器获得的模拟捕获的生物特征的模板识别功能 在临床人群中。

PROJECT SUMMARY Opportunities now exist to implement a paradigm shift in health management towards individualized physio- behavioral (biometric) monitoring - to predict, to prevent, and to better manage disease using wearable technologies, as well as embedded sensor technologies within wheelchairs as well as within the home. Our broad objective is to interpret collected combinatorial changes in the same biometric variables captured noninvasively during the progression of SCI in naturally behaving mice. In well-controlled animal studies, we propose to apply machine learning algorithms to identify ‘digital biosignatures’ that are predictive to disease emergence and/or expression, and therefore of use in feedback-based mitigation. To achieve this, we have engineered specialty instrumented mouse home-cages with commercially available sensors that enable continuous long-term noninvasive home cage capture of these biometrics to prototype development of such digital biosignatures. Emphasis is on understanding temporal interrelations in the emergence of sleep disturbances, neuropathic pain, thermoregulatory dysfunction, cardiorespiratory dysfunction and autonomic crises (autonomic dysreflexia) after SCI. Accordingly, home cage sensor-based capture includes all motor events, respiration, heart rate, 3-state sleep, skin temperature thermography and sensory preference testing. Our overarching hypothesis is that combined continuous capture several variables during the progression of SCI will identify novel ‘digital biosignatures’ that link to emergent dysfunction. The longer-term goal is to incorporate capture of digital biosignatures into real-time feedback-based approaches that limit disease expression. Two SCI models will be used to quantify variability in emergent dysfunction with the temporal correspondence of alterations in measured biometrics: [1] T9-10 contusion SCI and [2] T2-3 complete transection. For both experimental series, variables will be continuously captured in specialty instrumented home cages located in environmentally controlled chambers both before and for 10 weeks after SCI or sham surgery. Captured biometrics will be further categorized for machine learning based on measures of SCI -induced dysfunction from more conventional tests of sensory and autonomic dysfunction to link noninvasive biometric digital biosignatures with established measures physio-behavioral dysfunction after SCI. If successful, capturing digital biosignatures of dysfunction in real time may have translational impact on individualized medicine applications in SCI individuals. This is because acquired biosignatures may then serve a template recognition function from analogously captured biometrics obtained from embedded/wearable sensors in clinical populations.