Predicting Cognitive Outcomes from Stroke Based on Lesion Location

摘要 中风对患者和他们所爱的人来说都是一种可怕的经历。焦虑的主要来源 不知道哪些与中风相关的缺陷会持续下去--我所爱的人还能说话吗？这个 康复的可能性由治疗神经科医生或康复专家根据他们的 亲身经历。没有可用的工具可以使用关于笔划位置的信息， 根据成百上千个其他人的结果进行查询，并生成个性化的 中风相关缺陷的定量预测和长期康复的预后。在此，我们建议 开发这样一个工具。在爱荷华大学，我们拥有世界上最全面的病变登记系统之一 世界上有大约3,500名患有局灶性获得性脑损伤、神经成像和关于预后的广泛数据的患者。 我们建议通过开发一种工具来预测认知结果，从而利用这一独特的资源 根据病变部位进行卒中。首先，我们建议绘制与特定认知最相关的大脑区域图 1000多名患者存在缺陷，包括说话困难或注意力问题等症状。 接下来，在前瞻性收集的急性缺血性中风队列中，我们将尝试预测认知能力 通过根据前述症状‘MAP’查询病变位置的结果。我们假设这一损伤 地点将是慢性认知结果的一个重要预测因素。第二，我们最近开发了一种 将病变相关缺陷与特定大脑网络联系起来的创新策略。它结合了传统的病变 利用来自健康成年人的人类连接组数据进行映射，以推断局灶性大脑破坏了哪些网络 损伤。我们将评估这种病变网络标测方法是否可以用于补充 传统的病变标测方法预测认知结果的额外差异。最后，我们将使用 高级统计建模以评估如何优化病变位置的预测性信息 与人口统计信息和基线筛查认知表现相结合，以最大限度地提高 在纵向队列中预测慢性认知结果。通过解决这些目标，我们将 开发临床工具的基础，该工具可应用于临床获得的MRI扫描，以帮助 决定认知结果的预后，这是有助于早期管理的关键因素， 中风患者的康复和生活规划。

Abstract Having a stroke is a frightening experience for both the patient and their loved ones. A major source of anxiety is not knowing what stroke-related deficits will persist – will my loved one ever be able to talk again? The likelihood for recovery is estimated by the treating neurologist or rehabilitation specialist based on their personal experience. There are no tools available that can use information about the location of the stroke, query it against the outcomes from hundreds or thousands of other individuals, and generate a personalized quantitative prediction of stroke-related deficits and prognosis for long-term recovery. Here, we propose to develop such a tool. At the University of Iowa we have one of the most comprehensive lesion registries in the world of some 3,500 patients with focal acquired brain lesions, neuroimaging, and extensive data on outcomes. We propose to capitalize on this unique resource by developing a tool to predict cognitive outcomes from stroke based on lesion location. First, we propose to map brain regions most associated with specific cognitive deficits across over 1000 patients, including symptoms such as difficulty speaking or problems with attention. Next, in a prospectively collected cohort with acute ischemic stroke we will attempt to predict cognitive outcomes by querying lesion location against the aforementioned symptom ‘maps.’ We hypothesize that lesion location will be a significant predictor of chronic cognitive outcomes. Second, we have recently developed an innovative strategy that links lesion-associated deficits to specific brain networks. It combines traditional lesion mapping with human connectome data from healthy adults to infer what networks are disrupted by focal brain lesions. We will evaluate whether this lesion network mapping approach can be used to compliment the traditional lesion mapping approach to predict additional variance in cognitive outcomes. Finally, we will use advanced statistical modeling to evaluate how predictive information from lesion location can be optimally integrated with demographic information and baseline screening cognitive performance to maximize predictions of chronic cognitive outcomes in a longitudinal cohort. By addressing these objectives, we will lay the foundation for developing a clinical tool that can be applied to a clinically-acquired MRI scan to aid in determining the prognosis of cognitive outcomes, a key factor that will help in the early management, rehabilitation, and life planning for patients with stroke.