> Publications > CENL Writing tips


There are many ways to go about writing a scientific paper in the field of neuroscience. The following are guidelines that have worked well for us so far. Writing is an iterative process. The following are steps you can take in sequence, but you will find yourself fine-tuning and going back from time to time, as you are moving from one step to the next. This is totally normal and healthy.
Additional comments or links are welcome.

Step1: Make a set of FIGURES

- In CorelDraw or PowerPoint, 1 figure per tab/page, 1 inch margin all around.
- Include figure captions on the figure. Make sure to label all axes. Make sure tick-marks labels are large and readable. Use the 'Arial' font for figure captions and text inside figures.
- In a multi-panel figure, clearly label the panels with letters (A, B, C, D... Arial, bold font), from left to right and top to bottom.
- Use color ONLY if necessary (readers will probably print out your paper using a black-and-white printer).
- Using pictures/graphs from other published work: If you cut-and-paste from a published paper, include 'from (Smith...2010)' in the caption. Any other edits, as minor as they may seem, should be indicated as 'adapted from (Smith...2010)'. In general it is always best to get the original editable figure from the authors. If not possible (or in the meantime, to get you going), just open the Smith paper in a PDF viewer, zoom in as much as possible on the figure, then copy-paste in Corel Draw.
- Importing figures from Matlab, Excel or other programs: Always import your figure in an 'editable' format (i.e. you should be able to access the text of the labels on the axes, change the color of the lines ...). The easiest way is to copy and 'paste special' using 'enhanced windows metafile' format. Corel Draw/PowerPoint should then allow you to 'ungroup all', so you can access all the components of the figure.
- When presenting a result, show an example trace (raw data, your 'best example'), then show group data. Don't forget to assess statistical significance.
- Think about the flow: From one figure to the next, you should be telling a 'story'.
- Think about a 'conceptual model' figure, either as first, or last figure of the set. This could be a cartoon, or a 'box-and-arrow' figure. This figure should give the overall framework (or conclusion(s)) of the study.
- Include a set of 'supplemental figures' at the end, in which you will put all the other figures that you have. Typically, those figures are the ones that are interesting but not necessarily useful in the paper. They may end up as actual supplemental material, or as figures in a powerpoint presentation of your work.
- It is also a good idea to follow this set of figures by a 'Other' or 'Trash' section. You will put there figures or bits of figures you have decided not to show in the paper (but, again, might be useful in a powerpoint presentation!).

Step2: Write the RESULTS section

- Take the set of figures, and start a narative. Open each description of each figure with a short paragraph motivating what the figure/experiment is about and why it was conducted. Cite the relevant literature (but minimize citations here. They will come in INTRODUCTION and/or DISCUSSION).
- Start decribing the results, panel after panel. Include whatever details are necessary in the text (number of rats/data points, statistical analyses ...). Use the past tense consistently.
- For each figure (perhaps even panel), make sure to conclude the description with a sentence or two on what has been learned ('these results showed that...'). As part of these sentences, try and introduce the next figure ('However, the interpretation could be that... hence we tested whether ...'). It is useful to start the narative for a given figure with one sentence-title announcing what the point of the figure (conclusion) actually is; this nicely partitions the result section.
- You may find that you need to edit the figures and add a few graphical items to help the readers (e.g. an arrow pointing to something...).
- Make sure to refer to the figures in the proper order (i.e. don't refer to figure 3 before refering to figure 2). All panels of all figures should be explicitely mentioned and discussed.

Step3: Write the METHOD section

- To some extent, this section is somewhat independent from the others. If you are bored in step 1 and 2, start this section. The overall goal of this section is to give enough information to the reader so s/he can reproduce your results. You will have to find a balance between enough and too much details...
- If your method is somewhat different from others, think about a figure (figure 1 or 2, if 1 is a conceptual model) that shows the main aspects of the methods. This could be a new maze, a time line of the experiments/tests...
- Clearly separate the method section into subsections (e.g. animals, behavioral task, electrophysiology, neuropharmacology, data acquisition, data analyses and statistics...).
- In general, only cite papers here because of their method section(s), or because their results influence your methodological choices.

Step4: Write the INTRODUCTION section

- Identify the 2-4 most important issues that your work addresses. Find the key references that have addressed them in the past, and briefly discuss them.
- As you discuss the papers/issues, clearly identify the 'open' questions, or 'unknowns', especially if some of your figures partially answer them (but do not say 'as we will see below', or 'as figure 2 shows'...Let the reader 'hang'...).
- Finish the introduction with a statement of your hypotheses, and a sentence or two about the methods.

Step5: Write the DISCUSSION section

- Go back to the key references in introduction, and the 'open' questions. Try and briefly repeat them, then answer them using your results.
- Try and structure the discussion into 2 or 3 thematic sections. They do not need to be explicitely labeled as such.
- Discuss the limitations and possible alternative interpretations of your results. Cite other papers that might be relevant to these limitations and alternatives.
- Include a short paragraph on possible follow up studies, or new experiments. These should relate to the 'broad implications' in the last sentence of the abstract (below).

Step6: Write an ABSTRACT

- 2-3 sentences to summarize the problem area.
- Summarize the main findings (this should come naturally from the figures). Avoid the mention of control experiments or conditions. Avoid the mention of the methods (or make it less than 1 sentence worth). Just focus on the main results.
- Conclude with a couple of sentences on the possible broad implication(s) of your findings.

Step7: Acknowledgements

- You can a short (2-3 sentences) section of acknowledgements. These should mention:
- Any and all institutional financial support that was used (even in part) to support the work. Ask JM if you are not sure. UofA programs (UROC, UBRP) certainly count. Also, if someone donated a drug or a piece of equipment, they should be acknowledged here.
- Any significant technical help that contributed to the work (e.g. someone who helped many times with running the animals, or who spent time teaching an experimental or data analysis technique). Note: If the contribution was significant and scientific, this person may be better listed as co-author.

--------- NOTES: -----------

- On citations: We use EndNote in the lab. It should be installed on all the machines. When communicating your paper for editing to your co-authors, make sure to email the text (Word), the figures (Corel Draw) and the reference files (Endnote). Avoid citing papers in the middle of a sentence; insert the citation(s) at the end of it instead. This unless you are explicitely refering to a paper as in
Our results are therefore comparable to those of Smith et al. (2009) except for the control group which shows more variability in our study than in theirs, as others have found (Doe, Smith, 2008; Hanibal and Fountain, 2006).
- In the lab we use: Microsoft word (text), Corel Draw or PowerPoint (figures and posters), Matlab and Excel (data analyses), SPSS (statistics), EndNote (citations), Adobe Premiere (movies) and Powerpoint (oral presentations). Please stick to these programs.
- To make things easier and keep track of your work, let us adopt the following file naming convention: To the name your .docx file, append a dash-separated list of the initials of all your co-authors, with an index. The index shows how many times the co-author has edited/reviewed the manuscript and it constantly increases. For example:
PTSDpaper-JD1-TH0-JM2.docx
means that this file contains text that JD has seen/edited once, TH has not had a chance yet, and JM has seen/edited it twice so far.
- When editing a manuscript, keep the 'track changes' option on, so that the co-authors can see what/where the changes are.
- On editing etiquette: All comments you receive from your co-authors should be accounted for and discussed, down to punctuation marks. It is considered 'bad etiquette' to ignore a suggested change without discussion/justification.
- Undergraduate semester reports: If you signed up for research credits (summer or regular semester), your grade will be based on: 1)Your work/attendance/participation to lab activities, 2) a 5-10 page report (format above, single spaced, including figures, excluding references) and 3) a 20-30 min presentation to the laboratory at the end of the semester (make sure to schedule it before finals week).

Other tips:

- The use of the passive voice.
- Techniques for clear scientific writing and editing (European Journal of Neuroscience).
- There are many writing classes around. Among others: 'NRSC 595B -- Scientific Writing Strategies, Skills and Ethics', 'NRSC 701 -- Communication in Neuroscience'.
-
An article on how to ensure reproducibility of computational models: Towards reproducible descriptions of neuronal network models, Nordlie, Gewaltig and Plesser, PLOS 2009.
- Other Resources:
Understanding Research Methods; An overview of the essentials (7th ed). M.L.Patten. Pyrczak Publishing, Glendale CA, 2009.