文章的摘要十分关键，因为很多研究人员只读摘要而不读全文。因此，摘要提供准确而详尽的研究总结十分重要：它可以帮助研究人员了解你所开展的工作、你的研究目的和研究发现以及研究结果的益处和重要性。摘要必须能够独立成文，具备研究概要的功能，使人不看全文就能读懂。在阅读摘要后对文章细节感兴趣的读者自然会继续阅读全文。因此摘要不必太面面俱到，例如，可不必列举方法细节。

尽管摘要是论文的第一部分，但事实上应最后撰写。在完成其他部分后应尽快写摘要，因为这些内容依然清晰地印在你的脑海中，使你能够对自己的工作进行简明而全面的总结，而不至于忽略任何重要的内容。不同期刊对摘要的撰写要求有所不同，因此应参照目标杂志的《稿约》了解具体要求。尽管杂志要求不同，但依然存在一些普遍应遵守的惯例：

•   应注意对字数的限制。通常来讲摘要的字数限制平均为250个词，但许多杂志要求更短些(如《Nature》和《BBRC》对摘要的篇幅限制为150个字)，而许多杂志(如《BioMed Central》)允许摘要篇幅稍长些。这充分说明了为什么应在写文章之前确定目标杂志。
•   应避免使用技术行话，从而使摘要更易懂，更具可读性。不同目标期刊的“技术行话”取决于杂志的读者情况(可以通过期刊网站查询)。例如：“焦虑测试”一词通常比“高架十字迷宫实验”更容易理解，除非该杂志专门针对行为研究人员。通常摘要因受篇幅所限不能对技术术语进行定义和解释。如果术语使用不可避免，应在首次提到时用简单的措辞加以定义。
•   如同技术术语一样，应尽量不使用缩略语，其可用性也取决于不同的目标期刊。例如，大多数杂志接受HIV的使用。相比之下，RT-PCR对于分子生物学技术的杂志是可以接受的，但绝大多数杂志要求在首次使用时给出完整拼写(reverse transcriptase polymerase chainreaction)。许多杂志在网页上列出可使用的缩略语。反复使用三次或以上的必要的缩略语应在首次使用时给出完整拼写。只使用一次或两次的缩略语应使用全称，除非这样做超出了字数要求。摘要中已给出全称的缩略语在正文中首次使用时也应给出全称。
•    尽管一些杂志允许在摘要中引用文献，但绝大多数杂志不允许引用文献。因此，除非你要投稿的杂志允许这样做，否则不应在摘要中引用文献。

以下是BBRC杂志作者须知给出的指导性意见：
•    摘要应放在第2页，即标题页之后
•    摘要应采用一段式，总结文章的主要发现，篇幅不超过150字
•    摘要后应列出10个用于收录和检索的关键词

一些杂志要求采用结构式摘要，分为背景、目的、方法、结果和结论。临床期刊可能要求额外或不同段落，如“patients”。因此，再次强调，在动笔之前应查阅目标杂志的《稿约》，确定杂志的具体版式或格式要求。

摘要后经常需要列出由作者选择的关键词。《稿约》会指出要求列出多少个关键词，甚至提供可供参考的关键词清单。选择合适的关键词很重要，因为他们可作检索之用。选择合适的关键词可以使你的文章更容易被发现和引用。因此，关键词越切合你的文章内容越好，应避免选择多数研究所适用的一般性术语。

实例：让为这个题目选择合适的关键词：
“Region-specific neuronal degeneration after okadaic acid administration”

好的关键词：okadoic acid、hippocampus、neuronal degeneration、MAP kinase signaling以及mouse (或是rat或其他实验动物)。

差的关键词：neuron、brain、OA (简写)、regional-specific neuronal degeneration以及signaling。这些词过于笼统。

英文原文
The snapshot: abstract and keywords

Yourpaper’s abstract is critical because many researchers will read thatpart only, rather than reading the entire paper. Therefore, it iscritical that it provides an accurate and sufficiently detailed summaryof your work so that those researchers can understand what you did, whyyou did it, what your findings are, and why your findings are usefuland important. Your abstract must be able to stand alone, that is, tofunction as an overview of your study that can be understood withoutreading the entire text. Readers who become interested in learning moredetails than can be included in the abstract will inevitably proceed tothe full text. Therefore, the abstract does not need to be overlydetailed; for example, it does not need to include a detailed methodssection.

Even though the abstract is one of the first parts ofyour paper, it should actually be written last. You should write itsoon after finishing the other sections, while the rest of themanuscript is fresh in your mind, enabling you to write a concise butcomprehensive summary of your study without overlooking anythingimportant. Requirements for abstracts differ among journals, so thetarget journal’s instructions for authors should be consulted forspecific details. Despite differences among journals, there are a fewgeneral rules that should be obeyed when writing an abstract:

•   The word limit should be observed; 250 words is probably about averageand commonly adopted as a word limit for the abstract, but manyjournals request shorter abstracts (for example, Nature Articles andBBRC both have a 150-word limit) while many others (for example, BioMedCentral journals) allow longer ones. This is one good reason why thetarget journal should be identified before you write your paper.
•   Technical jargon should be avoided so that the abstract isunderstandable for a broad readership, although what is considered“technical” may vary depending on the target journal’s audience (checkthe journal’s website for details of their readership). For example, “atest of anxiety” would generally be clearer than “elevated plus-mazetest” in an abstract unless the journal was specifically targeted tobehavioral researchers. Usually, there simply isn’t enough space in theabstract to define and explain technical terminology. If suchterminology is unavoidable, it should be defined in simple terms whereit is first used.
•    Like technical jargon, abbreviations shouldbe limited as much as possible, although their acceptability may againdepend on the target journal. For example, HIV is likely to beacceptable in abbreviated form by most journals. By contrast, RT-PCRmight be considered acceptable by a journal reporting molecular biologytechniques, but would it need to be spelt in full (reversetranscriptase polymerase chain reaction) in most journals at first use.Many journals provide a list of acceptable abbreviations on theirwebsites. Necessary abbreviations used three or more times should bedefined at first use; however, abbreviations used only once or twiceshould be spelled out in full unless doing so causes the word limit tobe exceeded. Abbreviations that are defined in the abstract will needto be defined again at first use in the main text.
•    Althoughsome journals do allow references to be cited in the abstract, the vastmajority do not. Therefore, unless you plan to submit to a journal thatallows it, you should not cite references in your abstract.

If we look at the instructions to authors for BBRC, we can see the following guidelines:
•    The Abstract should be on page 2, i.e., after the title page
•    The Abstract must be a single paragraph that summarizes the main findings of the paper in fewer than 150 words.
•    A list of up to 10 keywords useful for indexing or searching should be included after the Abstract.

Somejournals request structured abstracts divided into sections such asbackground, objectives, methods, results, and conclusions. Clinicaljournals may require additional or alternative sections, such as‘patients’. Therefore, it is again necessary to check the targetjournal’s instructions for authors to determine the particularformatting/outline requirements prior to writing.

Abstracts arefrequently followed by a list of keywords selected by the authors. Theinstructions for authors will state how many keywords are required andmay even provide a list of recommended keywords. Choosing appropriatekeywords is important, because these are used for indexing purposes.Well chosen keywords enable your manuscript to be more easilyidentified and cited. Thus, the keywords should be as specific to yourmanuscript as possible, and general terms, which could apply to anenormous number of studies, should be avoided.

Examples:
Let’sconsider some appropriate keywords for the following title:“Region-specific neuronal degeneration after okadaic acidadministration”.

Good keywords would be: okadaic acid,hippocampus, neuronal degeneration, MAP kinase signaling, and possiblymouse (or rat or whatever experimental animal was used).

Poorkeywords would be: neuron, brain, OA (as an abbreviation),regional-specific neuronal degeneration, and signaling. These terms aresimply too general.

Dr Daniel McGowan
分子神经学博士
理文编辑学术总监