A conversation with DR joe macinnes
Joseph MacInnes Eye Movements Researcher Dr. Joseph MacInnes is an Associate Professor at the Psychology Department at Higher School of Economics and the Head of Vision Modelling Laboratory in Moscow |
What brought you to the Eye Movements research?
It was a lucky combination of opportunity and interesting research.
I started as a programmer of Eye Tracking experiments at Ray Klein's Psychology lab in Canada. I was extremely interested in the research being done, and Ray was patient enough to answer all my questions and include me on a few projects. I ended up doing a PhD under his supervision combining computer science and psychology.
I started as a programmer of Eye Tracking experiments at Ray Klein's Psychology lab in Canada. I was extremely interested in the research being done, and Ray was patient enough to answer all my questions and include me on a few projects. I ended up doing a PhD under his supervision combining computer science and psychology.
How do you conduct experiments?
I start with the research question first, and then decide what tool is best suited to answer that question. Most of my research questions involve the interplay of attention and eye movements, so the Eye Tracker is an important tool in most of my studies.
I start with the research question first, and then decide what tool is best suited to answer that question
What matters most in the choice of Eye Tracker?
Eye Trackers differ in spatial accuracy, temporal resolution, latency, type of information that you can ask it, and of course the price. Eye Trackers that do most of what we need in the lab are currently quite expensive.
Some of my experiments, like gaze contingent studies, require extremely high spatial precision and low latency. If we want to limit the viewing area to a small circle around the current gaze position, we need to query the gaze position and redraw the screen at a very fast resolution. Sometimes we change positions of objects on the screen when participants are less likely to observe it.
Gaze contingent studies change the display based on where and when the participant moves their eyes
During saccades we suppress much of the incoming visual information to prevent motion blur. If we want to change the screen during this time, we need to detect the onset of that saccade with millisecond accuracy. We need a tool that saves the data for all of these effects. We also do computational modelling of the visual system and this wealth of data is extremely important for training and testing our models.
Saccades are ballistic eye movements that may only last 20-30 milliseconds
How does your workflow look these days,
when the lab is closed and you can't collect data?
when the lab is closed and you can't collect data?
The research pipeline for our experiment projects work from idea to experiment design to data collection to analysis, and finally to publication. This can take a year or longer and we always have multiple projects at different stages of the pipeline. The shutdown has actually been great for finishing up projects that are in later stages since everyone in the lab has been focused on finishing older work. It has been hard on some of our master's students who were still collecting data on their projects, and I think we will notice a problem more next year because we aren't starting any new projects now.
The research pipeline for our experiment projects work from idea to experiment design to data collection to analysis, and finally to publication
Where do you see the field in 10 years?
I think Eye Trackers will become much more ubiquitous in research and less intrusive. Smaller cameras, cheaper and faster computers will open up their use in VR/AR research and real-life tasks. Eye Tracking research which allows for head and body movement is still in its infancy. I think in the future more researchers will have access to inexpensive, high quality eye tracking data for nearly any task.
Editor-in-Chief — Kirill Korotaev