All of us involved in the mind and brain sciences, in whatever capacity, are increasingly aware of the importance of ensuring the credibility of our research. This credibility—which for experimental work turns on its reliability—is particularly salient for consciousness research, given the at-times precarious perception of consciousness science within the wider landscapes of psychology and neuroscience (Michel et al. 2019). We are therefore very pleased to introduce a number of ‘open science’ initiatives that have recently been implemented in Neuroscience of Consciousness. One key objective of open science is to actively resist the tendency to search for, and to preferentially publish, eye-catching findings that fit with compelling narratives. This means supporting studies that have been designed to report unbiased results that are reproducible and replicable, whether or not they are narratively convenient. The gold standard approach here is the Registered Report (RR) format. Intuitively, the idea is to allow researchers to retain complete control over their hypotheses, methods and interpretations, but not over the results (Hardwicke and Ioannidis 2018; Chambers 2019). RRs divide the peer review process into two stages. In the first, reviewers evaluate the rationale, hypotheses, methods and proposed analyses for a study. These are described in a ‘stage 1’ manuscript. This stage takes place before any data collection, other than pilot data collected purely to inform study design. Submissions can be rejected, sent for revision, or accepted at this stage. Importantly, stage 1 peer review offers opportunities to enhance study design—including correction of potentially fatal confounds and flaws—while there is still time. This benefits the researchers and is rewarding for the reviewers. Once a stage 1 submission is accepted, it is preregistered in a time-stamped public repository such as the Open Science Framework (https://osf.io/), with or without a temporary embargo. Following this ‘in principle acceptance’, the researchers carry out their planned data collection and perform their preregistered analyses, as well as any additional, clearly demarcated, exploratory analyses. They then submit a stage 2 manuscript incorporating these results and a discussion. The introduction and methods remain unchanged from the stage 1 in-principle-accepted manuscript. This stage 2 submission is sent back to the original reviewers—who cannot at this point raise post hoc objections to the study rationale or design. Critically, the journal is committed to publish this stage 2 manuscript no matter how the results turn out, so long as the researchers followed the procedures that they described in their stage 1 submission, and have written up their results and discussion sections satisfactorily, as judged by the reviewers. The wow-factor of the results is simply not relevant to the acceptance decision. In this way, publication of a RR does not depend on the appeal of a post hoc story based on eye-catching findings, but instead on the significance of the research question and the adequacy of the experimental design to address this question. As mentioned, exploratory analyses are not excluded by the RR format. In fact, they are encouraged—so long as they are clearly marked as exploratory and reported separately from the primary preregistered analyses. Neuroscience of Consciousness has been accepting RR submissions since June 2019, and we are delighted that Prof. Zoltán Dienes (University of Sussex, UK) has joined our editorial team specifically to handle these manuscripts. We strongly encourage consciousness researchers to take advantage of this innovative format. The journal webpages contain further information about the format, including detailed instructions on how to prepare RR submissions (https://academic.oup.com/nc/pages/ General_Instructions). Despite the many benefits offered by RRs, both to individual researchers and to the wider community, we recognize that it is not always possible to follow this process to its fullest extent. This may be the case when data collection cannot be timed to the researcher’s convenience; where the research is purely exploratory, or focuses on methodological innovations, or is otherwise constrained in ways that preclude the RR procedure. Some kinds of research—such as computational modelling—fit uneasily in present RR formats. Fortunately, RRs are not the only tool by which open science can be facilitated. Experimental designs and proposed analyses can still be—and wherever possible should be—preregistered prior to data collection (Nosek et al., 2018). This means that the critical distinction between (preregistered) confirmatory and (non-preregistered) exploratory analyses can still be maintained. As with RRs, this separation mitigates against the selective reporting or ‘cherry picking’ of results, as well as against p-hacking (conducting repeated analyses until a desired outcome is obtained, without sufficient statistical correction) and HARK-ing (hypothesizing-after-the-results-are-known)—all of which constitute seductive traps for non-preregistered research. Several options for preregistration are available. One flexible format is provided by www.aspredicted.org—with the Open Science Framework again providing a suitable repository. Other open science initiatives focus on transparency. Experimental data can be shared, along with code

中文翻译：

---

社论：意识研究中的开放科学

无论身在何处，参与心理和脑科学的所有人都越来越意识到确保我们研究的可信度的重要性。鉴于在广泛的心理学和神经科学领域对意识科学的认知perception可危，这种可信度（对于实验工作而言，取决于其可靠性）对于意识研究尤为重要（Michel et al.2019）。因此，我们非常高兴地介绍一些最近在“意识神经科学”中实施的“开放科学”计划。开放科学的主要目标之一是积极抵制寻找和优先发表与引人入胜的叙述相匹配的醒目的发现的趋势。这意味着要支持旨在报告可重现和可复制的无偏结果的研究，无论它们在叙述上是否方便。此处的黄金标准方法是注册报告（RR）格式。从直觉上讲，这个想法是让研究人员能够完全控制他们的假设，方法和解释，但不能完全控制结果（Hardwicke和Ioannidis 2018; Chambers 2019）。RR将同行评审过程分为两个阶段。首先，审阅者评估研究的基本原理，假设，方法和建议的分析。这些在“第一阶段”手稿中进行了描述。该阶段发生在任何数据收集之前，而不是纯粹为了研究设计而收集的试验数据。在此阶段，提交可以被拒绝，发送以进行修订或接受。重要的，第1阶段同行评审提供了改进研究设计的机会-包括纠正可能致命的混淆和缺陷-同时还有很多时间。这使研究人员受益，并使审稿人受益。一旦阶段1提交被接受，就可以在有或没有临时禁运的情况下，在带有时间戳的公共存储库（例如Open Science Framework（https://osf.io/））中预先注册。在“原则上接受”之后，研究人员进行了计划的数据收集并进行了预先注册的分析，以及所有其他明确标定的探索性分析。然后，他们提交了包含这些结果和讨论的第二阶段手稿。从第一阶段原则接受的手稿开始，介绍和方法保持不变。第2阶段的提交意见将发回原始审稿人，他们现在无法对研究原理或设计提出事后异议。至关重要的是，无论研究结果如何，该期刊都致力于发布该阶段2的手稿，只要研究人员遵循他们在阶段1提交中描述的程序，并且令人满意地写下他们的结果和讨论部分即可，由审阅者。结果的惊人因素与验收决定根本不相关。这样，《无线电规则》的发表并不取决于基于引人注目的发现的事后故事的吸引力，而是取决于研究问题的重要性以及实验设计是否足以解决该问题。如前所述，RR格式不排除探索性分析。实际上，鼓励使用这些方法，只要将其明确标记为探索性的，并与主要的预先注册分析分开报告即可。自2019年6月以来，意识神经科学已开始接受RR提交的论文，我们很高兴ZoltánDienes教授（英国苏塞克斯大学）特别加入了我们的编辑团队来处理这些手稿。我们强烈鼓励意识研究者利用这种创新形式。期刊网页包含有关格式的更多信息，包括有关如何准备RR提交的详细说明（https://academic.oup.com/nc/pages/General_Instructions）。尽管《无线电规则》为个人研究人员和整个社区都带来了许多好处，但我们认识到，并非总是可以最大程度地遵循这一过程。当无法及时收集数据以方便研究人员使用时，可能会出现这种情况。研究纯粹是探索性的，或侧重于方法的创新，或者以其他方式限制了RR程序的地方。某些研究（例如计算模型）难以适应当前的RR格式。幸运的是，RR并不是促进开放科学的唯一工具。实验设计和建议的分析仍可以在可能的情况下在数据收集之前进行预注册（Nosek等，2018）。这意味着（预注册的）确认性分析和（非预注册的）探索性分析之间的关键区别仍然可以保持。与RR一样，这种分离减轻了对结果的选择性报告或“樱桃采摘”的影响，以及针对p-hacking（进行重复分析，直到获得所需结果，而没有进行足够的统计校正）和HARK-ing（已知结果后假设）-所有这些都构成了诱骗性陷阱，预注册的研究。有几种预注册选项。www.aspredicted.org提供了一种灵活的格式-开放科学框架再次提供了一个合适的存储库。其他开放式科学计划注重透明度。可以与代码共享实验数据 www.aspredicted.org提供了一种灵活的格式-开放科学框架再次提供了一个合适的存储库。其他开放式科学计划的重点是透明度。可以与代码共享实验数据 www.aspredicted.org提供了一种灵活的格式-开放科学框架再次提供了一个合适的存储库。其他开放式科学计划的重点是透明度。可以与代码共享实验数据