Research

I started working on numerosity perception in the first year of my PhD, and I am still working on it. There are many questions that can be asked on this topic. One of the central questions is where it comes from: is it an inherent ability – the number sense – or does it come from something else? How do we know the number of objects we are looking at without explicitly counting them? What does the visual system do when perceiving numbers?

Our research shows that numerosity perception is systematically shaped by the spatial structure of vision. The same number of objects, when arranged differently, can be perceived as having a different number of items. In many of our experiments, we found that when items are arranged in a way that makes them interfere more strongly with each other, they appear less numerous than when they are arranged in a way that produces weaker interference.

The former arrangement is called radial, and the latter tangential. This difference is known as the radial–tangential anisotropy. It has been shown to occur in many visual phenomena, and we were the first to show that it also occurs in numerosity perception.

We are now extending this work from behavior to the brain. By combining visual psychophysics with EEG and neural decoding, we are investigating when spatial arrangement begins to influence neural representations of number.

I became interested in redundancy masking because the phenomenon itself is quite striking: you present three clearly separated letter Ts in peripheral vision, but people often report seeing only two. Three becoming two is kind of crazy—one entire object seems to disappear from conscious perception. This phenomenon was first demonstrated by my PhD supervisor and PI, Bilge, and colleagues in 2019 (Sayim & Taylor, 2019).

Redundancy masking occurs when similar or repeated items are presented close to each other in peripheral vision. Instead of representing every item separately, the visual system seems to compress redundant information. In this project, we ask what actually happens to the information that seems to be “lost.” Is the missing item simply discarded, or does the visual system preserve some of its information while compressing the scene?

We first extended this phenomenon to a more meaningful type of visual object: faces. Faces are usually detected very efficiently. However, even with faces, people frequently reported seeing only two when three were presented. Losing an item from conscious report does not necessarily mean losing all the information carried by that item.

We also asked whether this compression could be beneficial. In experiments with repeated bars, we found that when participants perceived fewer items than were actually presented, their perception of the remaining items’ width was more accurate. This suggests that redundancy masking may not simply be a failure of peripheral vision. Instead, the visual system may trade numerical accuracy for a better representation of other visual features.

We are also investigating what happens in the brain when redundancy masking occurs. Using EEG and frequency tagging, we found that the “missing” item still leaves a neural signal.

There are now many ongoing studies in this project, approaching redundancy masking from different directions. We still have a lot to understand about when, how, and why the visual system compresses redundant information. Perhaps there really is something like a “compression switch” in the brain, a mechanism dedicated to deciding when redundant information should be compressed, just waiting for us to discover it.