Neurotrauma as a big-data problem.
Neurotrauma as a big-data problem.
复制标题
DOI：
10.1097/wco.0000000000000614
复制
发表时间：
2018-12
影响因子：
4.8
通讯作者：
Ferguson AR
中科院分区：
文献类型：
--
作者：
Huie JR;Almeida CA;Ferguson AR
神经创伤研究领域面临着可重复性危机。作为回应，创伤性脑损伤（TBI）和脊髓损伤（SCI）的研究领导者正在利用数据管理和分析方法来鼓励透明度，并提高严谨性和可重复性。在这里，我们回顾了当前的挑战和机遇，这些挑战和机遇来自于将神经创伤的大数据转化为知识的努力。三个平行运动正在推动神经创伤中的数据驱动发现。首先，大型多中心联盟正在收集大量的神经创伤数据，完善可用于跨研究的通用数据元素（CDE）。研究人员现在正在不同的研究环境中测试CDE的有效性。其次，数据共享计划正在努力使神经创伤数据可查找，可访问，可互操作和可重用（FAIR）。最近的临床前和临床神经创伤开放数据库项目反映了这些努力。第三，机器学习分析使研究人员能够发现新的数据驱动假设，并在多维结果空间中测试新的治疗方法。我们正处于数据收集、管理和分析的新时代的门槛上。大数据在神经创伤研究中的下一阶段将需要负责任的数据管理，数据共享文化以及“暗数据”的照明。
The field of neurotrauma research faces a reproducibility crisis. In response, research leaders in traumatic brain injury (TBI) and spinal cord injury (SCI) are leveraging data curation and analytics methods to encourage transparency, and improve the rigor and reproducibility. Here we review the current challenges and opportunities that come from efforts to transform neurotrauma’s big data to knowledge. Three parallel movements are driving data-driven-discovery in neurotrauma. First, large multicenter consortia are collecting large quantities of neurotrauma data, refining common data elements (CDEs) that can be used across studies. Investigators are now testing the validity of CDEs in diverse research settings. Second, data sharing initiatives are working to make neurotrauma data findable, accessible, interoperable and reusable (FAIR). These efforts are reflected by recent open data repository projects for preclinical and clinical neurotrauma. Third, machine learning analytics are allowing researchers to uncover novel data-driven-hypotheses and test new therapeutics in multidimensional outcome space. We are on the threshold of a new era in data collection, curation, and analysis. The next phase of big data in neurotrauma research will require responsible data stewardship, a culture of data-sharing, and the illumination of ‘dark data’.