Harnessing machine learning and cloud computing to test biological models of the role of white matter in human learning

该奖项是美国国家科学基金会社会、行为和经济科学（SBE）博士后研究奖学金（SPRF）计划和SBE的学习科学和增强智能计划的一部分。SPRF计划的目标是为学术界、工业界或私营部门和政府的科学事业准备有前途的早期职业博士级科学家。SPRF奖励包括在知名科学家的赞助下进行为期两年的培训，并鼓励博士后进行独立研究。美国国家科学基金会寻求促进科学界各阶层的科学家，包括那些未被充分代表的群体的科学家，参与其研究项目和活动；博士后阶段被认为是实现这一目标的一个重要的专业发展阶段。每个博士后必须解决各自学科领域的重要科学问题。在印第安纳大学的Franco Pestilli博士的赞助下，这个博士后奖学金奖支持一位早期职业科学家，研究人类大脑中白质通讯通路在学习和概括中的作用。先前的工作将人类白质的个体差异与当前能力联系起来，作为过去学习的衡量标准；相反，拟议中的研究将利用人类白质的个体差异来预测未来的学习能力。这项工作提出的假设是，感觉运动训练改变了白质沟通途径，从而允许对未经训练的行为进行概括。研究人员将通过使用机器学习方法和实施明确的模型测试方法来验证这一假设。这项研究将为该领域提供有关大脑学习相关变化的重要信息，这些信息将适用于教育和神经康复实践。该项目将大脑白质通讯途径的尖端测量与新颖的行为评估相结合。这项提议的工作建立在一个充分记录和可重复的发现之上：感觉运动学习导致一般化的学习（例如，手写增加字母识别）。该项目有三个目标。第一个目标是证明在感觉运动任务上的训练（例如，绘制新的符号）会导致白质沟通通路的组织特性发生任务特异性的变化。我们将采用参与者之间的训练操作，并评估实验组之间学习相关白质微观结构的差异。第二个目标是证明与感觉运动学习相关的白质变化支持对未经训练的行为的概括。我们将使用机器学习来建立一个学习相关的白质组织微观结构变化与感觉运动学习之间关系的模型。然后，我们将量化该模型预测视觉识别学习（即学习识别新符号）的效果。我们的期望是，全球白质组织特性的个体差异将预测感觉运动学习和泛化。这项工作的最终目标是利用云计算平台——大脑生活。提供开放科学和可重复的方法，以及公开可用的分析和服务。数据、分析和结果将在“大脑生活”上共享。IO有可能影响对学习感兴趣的多个科学家群体：行为科学家、计算机科学家和神经科学家。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This award was provided as part of NSF's Social, Behavioral and Economic Sciences (SBE) Postdoctoral Research Fellowships (SPRF) program and SBE's Science of Learning and Augmented Intelligence Program. The goal of the SPRF program is to prepare promising, early career doctoral-level scientists for scientific careers in academia, industry or private sector, and government. SPRF awards involve two years of training under the sponsorship of established scientists and encourage Postdoctoral Fellows to perform independent research. NSF seeks to promote the participation of scientists from all segments of the scientific community, including those from underrepresented groups, in its research programs and activities; the postdoctoral period is considered to be an important level of professional development in attaining this goal. Each Postdoctoral Fellow must address important scientific questions that advance their respective disciplinary fields. Under the sponsorship of Dr. Franco Pestilli at Indiana University, this postdoctoral fellowship award supports an early career scientist investigating the role of white matter communication pathways in the human brain in learning and generalization. Prior work has related individual differences in human white matter to current abilities as a measurement of past learning; the proposed work will, instead, use individual differences in human white matter to predict future learning. The hypothesis addressed in the proposed work is that sensorimotor training changes white matter communication pathways in ways that allow for generalization to untrained behaviors. The investigators will test this hypothesis by using machine-learning methods and implementing an explicit model testing approach. This research will provide the field with important information concerning learning-related changes in the brain that will be applicable to educational and neuro-rehabilitation practices. This project integrates cutting-edge measurements of white matter communication pathways in the brain with novel behavioral assessments. The proposed work builds from a well-documented and repeatable finding: sensorimotor learning leads to learning that generalizes (e.g., handwriting increases letter recognition). The project has three goals. The first goal is to demonstrate that training on a sensorimotor task (i.e., drawing novel symbols) leads to task-specific changes in the tissue properties of white matter communication pathways. We will employ a between-participants training manipulation and assess differences in learning-related white matter microstructure among training groups. The second goal is to demonstrate that the white matter changes associated with sensorimotor learning support generalization to an untrained behavior. We will use machine-learning to build a model of the relationship between learning-related changes in white matter tissue microstructure and sensorimotor learning. We will then quantify how well that model predicts visual recognition learning (i.e., learning to recognize the novel symbols). The expectation is that individual variability in global white matter tissue properties will predict sensorimotor learning and generalization. The final goal of the work is to leverage the cloud computing platform–brainlife.io–to deliver open-science and reproducible methods as well as publicly available analyses and services. Data, analyses, and results will be shared on brainlife.io with the potential to impact multiple communities of scientists interested in learning: behavioral scientists, computer scientists, and neuroscientists.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.