Monday, November 24, 2008

Optimizing SfN Poster Design Using Eye Tracking Software

Scan Path.jpg


This week I attended my second SfN meeting in Washington D.C. In my spare time between poster sessions, I walked the entire vendor exhibition section browsing for companies that I know of. One vendor that caught my eye (pun to be appreciated shortly) was SensoMotoric Instruments, a developer of eye-tracking software.

A very helpful company representative took time to discuss the details of how the software works despite no obvious benefit of implementing their software in my research.

Then it hit me. If this eye tracking software can be used by web-designers to optimize the layout of their web pages, surely it can be used by researchers to optimize the layout of their conference posters.

The following day I returned to their exhibition booth with a PDF of my poster loaded on a USB drive. The gentlemen helped me load my poster into their eye tracking software called BeGaze and a new "experiment was launched". I was the subject and I was seated in front of a 17 in monitor that had an infrared eye camera that could automatically detect eye gaze direction and head position.

The video below shows the path my eyes took when scanning my poster:



The larger the circle diameter, the longer time was spent focusing on that point.

Once this data was collected, various options exist to describe the way I viewed my own poster. One of which is the colorful Attention Map. The idea here is that the areas of my poster that had a longer gaze time have "warmer" colors:

Attention Map.jpg


Another view is the Focus Map, which shows all areas that I looked at ignoring the time spent. This way I can see what areas I didn't even pay attention to:

Focus Map.jpg


To go to another level of analysis, you can assign regions of interest (e.g., title, intro, results, discussion, etc.):

AOI Editor.jpg


Then you can compare values of dwell time for a group of subjects and perform statistical analysis. I have made a graph here:

dwell time graph.png


The data above are from one person, myself, but with more people viewing you may be able to get a real feel for what areas of your poster people focus on. You can also assess the path people take as they move through your poster and compare that to what you envisioned when you were putting it together in the lab.

I'm afraid I don't have the time to look into the application of this software to really fine-tuning your posters for next neuroscience. However, the representative at the booth was very excited at the prospect of setting up a kiosk at the SfN conference, having people load their poster file onto a computer, and then have the next 20 people that come scan their poster and collect data.

So if you see this next year at neuroscience you now know it came to be.

6 comments:

Drugmonkey said...

that is totally and completely AWESOME!!!!

Mike Pascoe said...

Thanks DrugMonkey!

brembs said...

cool! Must've missed that booth...

matt said...

Thanks, that's a great suggestion to us who are already using eyetrackers already.

Hilary said...

Hi Mike,
I'm just preparing a training session on how to design a conference poster, and found your post interesting.

I wonder if it the eyetracker results would be slightly different if you were reading a poster that was unfamiliar to you. The analysis video showed that you carefully read all the way through your poster, I suspect that if you were reading a poster that was less familiar, there may be some interesting differences between the two sets of results

Best wishes,
Hilary

Mike Pascoe said...

Thanks for the comment Hilary.

I have no doubt that the results would differ if a stranger viewed my poster or if I viewed an unfamiliar poster. That would have been my ideal test but I couldn't ask the vendor to stop their demo schedule just to plug in my lowly poster. Also, results would vary between those familiar with my area of research and those from other disciplines.

I do think this type of analysis is helpful for the basic organization of a poster. Arrange content in columns, so readers go top to bottom and then right to the next column. It would also be helpful in seeing if eyeballs can easily move from figure 1 to figure 2 and so on. This is why I have begun numbering my content to guide the reader.

Best to you on your training session. I highly encourage you to post materials online (slides on scribd, narrated talk on vimeo, etc...) and share your expertise with the online community.

My slides from my how to design a conference poster are on scribd and are the most viewed of all my docs on scribd! See here - http://bit.ly/Gyxpp

Best - Mike Pascoe