Scientific presentations: PowerPoint structure, templates and slides for research

Scientists are trained to write for readers who can reread. A talk has listeners who cannot. That single difference explains most bad research talks: methods slides as dense as the methods section, figures copied from the paper at a size nobody can read, and the main result arriving on slide 23 of 25. The fixes are simple, but they require cutting material you worked hard on.

The structure of a research talk

A typical 12- to 15-minute conference talk follows the paper’s logic in a lighter form. A sensible time budget:

For a thesis defence or a seminar of 45 minutes, the same skeleton scales up: more motivation and background, one section per study or chapter, each with its own small results and summary, and a final synthesis. For a lab meeting, flip it: the problem you are stuck on first, then the data.

Figures: redraw them for the screen

Paper figures are designed to be read up close, with many panels, small labels and every condition shown. On a slide, each figure should make one point the audience can see in a few seconds.

• One panel per slide. Split multi-panel figures; reveal panels one at a time if they belong together.
• Large labels: axis labels and tick labels at 18 pt or more. Re-export from your analysis software at slide size, or rebuild simple charts in PowerPoint.
• Highlight the comparison you talk about; grey the rest.
• Direct labels instead of legends where possible.
• Error bars and significance shown as in the paper, with what they mean (SD, SE, 95 % CI) in a short note.
• Colour-blind-safe palettes: avoid red–green contrasts as the only difference.

Equations, symbols and units

Show an equation only if you will talk through it. PowerPoint has a full equation editor: Insert → Equation (Alt+= on Windows) opens it, and it understands LaTeX-like input — type \frac{a}{b} or \alpha and press the space bar. In recent versions you can switch the input to LaTeX on the Equation tab. For many equations, a LaTeX tool that exports pictures, inserted as SVG, keeps them consistent with your paper.

Write units with a space before them (“5 mg”, “20 °C”), use the same number of decimals within a comparison, and spell out abbreviations the first time they appear — the audience may come from neighbouring fields.

References and credits on slides

• Short citations on the slide where a claim or figure comes from: “(Smith et al., 2023)” in small text at the bottom.
• Figure credits for every picture you did not make, with its licence.
• A full reference slide at the end, after the conclusion, for those who want to read on — or a link or QR code to a reference list.
• Your own paper on the title or conclusion slide, with its DOI if it is published.

What a good scientific template gives you

Many universities and institutes provide PowerPoint templates in their branding; use them where required. A good scientific template has a calm content layout with a wide area for figures, a title-only layout, a two-column layout for figure plus text, a small reference line in the footer, slide numbers (people ask questions by number), and fonts that read well at a distance. Dark backgrounds suit darkened lecture halls and microscopy; light ones suit bright rooms and printed handouts.

Delivering a research talk

• Rehearse with a timer at least three times; conference chairs cut speakers off.
• Point with words, not a laser: “the dark bars on the left show…”. A laser pointer jitters and is invisible on many screens.
• Repeat the take-home message at the start, after the results and at the end.
• Prepare for the three questions you fear most, each with a backup slide.
• Bring the deck as PDF and .pptx on a stick and in the cloud.

Methods slides people can follow

The methods section is where most research talks lose the room. The audience needs to know enough to trust the results — not enough to replicate the study. Three slides usually do it:

1. The design as a diagram. Groups, conditions and time points as boxes and arrows, drawn with shapes or SmartArt. A flow diagram of participants (recruited, excluded, analysed) works well for clinical and survey studies.
2. What was measured. The key measures with one line each, and a picture of the task, instrument or setup if it helps — a photo of the apparatus, a screenshot of the test.
3. How it was analysed. The main model or test in plain words, with the details in a backup slide.

Ask yourself for each methods detail: would a sceptical listener doubt the results without it? If not, it belongs in the paper or the backup slides. Specialists in the audience will ask, and a prepared backup slide answers better than a crowded main slide.

From paper to talk with ExpertSlides

The most tedious part of a research talk is turning a twenty-page paper into fifteen slides. ExpertSlides does the first draft: attach the paper as a PDF (or your manuscript as Word), say who the audience is and how long the talk is, and the AI plans a deck in the structure above — motivation, question, methods as a diagram, results one per slide with the finding as the title, discussion and conclusion. It keeps to what the paper says; tables of figures become real charts with their values, and nothing is invented.

You then check every statement, insert your own exported figures into the picture frames, and rehearse. Present from the browser, share a Web link with the session chair, or download the deck: PowerPoint export is part of Pro and Business. Every plan exports PDF; on Free the PDF carries the ExpertSlides mark.

Related: presenting data, posters in PowerPoint, and the rules for good slides. See also document to presentation and the Design Templates.