Dimitriadis George, Svahn Ella, MacAskill Andrew F, Akrami Athena
Sainsbury Wellcome Centre, University College London, London, United Kingdom.
Department of Neuroscience, Physiology and Pharmacology, University College London, London, United Kingdom.
Elife. 2025 Jul 16;13:RP91915. doi: 10.7554/eLife.91915.
To realise a research project, experimenters face conflicting design and implementation choices across hardware and software. These include balancing ease of implementation - time, expertise, and resources - against future flexibility, the number of opaque (black box) components and reproducibility. To address this, we present Heron, a Python-based platform for constructing and running experimental and data analysis pipelines. Heron allows researchers to design experiments according to their own mental schemata, represented as a Knowledge Graph - a structure that mirrors the logical flow of an experiment. This approach speeds up implementation (and subsequent updates), while minimising black box components, increasing transparency and reproducibility. Heron supports the integration of software and hardware combinations that are otherwise too complex or costly, making it especially useful in experimental sciences with a large number of interconnected components such as robotics, neuroscience, behavioural sciences, physics, chemistry, and environmental sciences. Unlike visual-only tools, Heron combines full control (of instrument and software combinations) and flexibility with the ease of high-level programming and Graphical User Interfaces. It assumes intermediate Python proficiency and offers a clean, modular code base that encourages documentation and reuse. By removing inaccessible technical barriers, Heron enables researchers without formal engineering backgrounds to construct sophisticated, reliable and reproducible experimental setups - bridging the gap between scientific creativity and technical implementation.
为了实现一个研究项目，实验人员在硬件和软件方面面临相互冲突的设计和实施选择。这些选择包括在易于实施（时间、专业知识和资源）与未来的灵活性、不透明（黑箱）组件的数量以及可重复性之间进行权衡。为了解决这个问题，我们推出了Heron，这是一个基于Python的平台，用于构建和运行实验及数据分析管道。Heron允许研究人员根据自己的心智模式设计实验，这种心智模式以知识图谱的形式呈现——知识图谱是一种反映实验逻辑流程的结构。这种方法加快了实施速度（以及后续更新），同时将黑箱组件数量减至最少，提高了透明度和可重复性。Heron支持整合那些原本过于复杂或成本过高的软硬件组合，这使其在机器人技术、神经科学、行为科学、物理学、化学和环境科学等拥有大量相互关联组件的实验科学领域中格外有用。与仅具备可视化功能的工具不同，Heron将（对仪器和软件组合的）完全控制权和灵活性与高级编程及图形用户界面的易用性结合在一起。它假定用户具备中级Python水平，并提供一个简洁、模块化的代码库，鼓励进行文档记录和代码复用。通过消除难以企及的技术障碍，Heron使没有正规工程背景的研究人员能够构建复杂、可靠且可重复的实验装置——弥合科学创造力与技术实施之间的差距。
