All animals need to search their environment – for food, for predators, for mates. When humans search for a specific target, the target often pops out from the background. For example, in the image below, the fish is obviously different than the people. No matter how many people were present, the fish would always be easy to spot.
In contrast, think about a typical Where’s Waldo. Here the search process is slow and difficult because you are looking for a man in a red and white striped shirt among many similar objects. The more objects there are, the slower your search will be.
Psychologists explain this difference using Feature Integration Theory. According to the theory, some basic aspects of a visual stimulus are processed quickly and automatically (e.g., color, shape, movement). However, combining those individual features of an object together is a slower process that requires attention.
Thus, in the image on the left, people are able to quickly find the grey fish among white fish, no matter how many white fish are present. The grey fish pops out. However, in order to find the grey fish that is facing right in the other image, people need to individually search through each fish. The more fish there are, the slower they’ll be at finding the correct fish.
Feature Integration Theory was first introduced by Treisman and Gelade in 1980 and its 40th anniversary was recently celebrated in a special issue of the Psychonomic Society journal Attention, Perception, & Psychophysics.
In one of the articles, authors Adam Reichenthal, Ronen Segev, and Ohad Ben-Shahar (pictured below) examine how non-human animals search for objects. Do the same features “pop out” for all animals or are there differences due to brain anatomy and evolutionary pressures?
Researchers have examined this question with a number of different species including pigeons, blue jays, bees, monkeys, and rats using a variety of ingenious methods. For example, in order to study the visual search behavior of barn owls, researchers place a tiny video camera on the owl’s head. Barn owls can make only very restricted eye movements, instead of moving their entire head to track objects. This feature allows researchers to use the head-mounted camera to determine exactly where the owl is looking.
The owl then flies over a collection of objects, looking for the odd one out. When the objects differed by just one dimension (orientation) the owls fixated on a similar number of items, regardless of how many objects were present.
However, when the target object was identified by a conjunction of two features (orientation and contrast), the number of fixations increased with the number of objects presented.
However, my favorite results come from the archerfish. Archerfish are amazing animals who shoot down insects using jets of water.
This ability allows researchers to train them to shoot at the target objects. Whenever the fish accurately shoots the target, they get a food pellet.
In the current studies, Reichenthal and colleagues had both humans and archerfish complete the same visual search tasks. For four of the tasks, the target object would be predicted to “pop out” of the display. The target was either a different color, a different size, a different orientation, or moving in a different direction (Feature search). For another two tasks, the target object was a different shape – either a disk among Pac-men or a Q among disks (Shape search). Then, the final few tasks required searching for two features at once, either a large and blue circle or a small and blue circle (Conjunction search) among distractors that varied in both shape and color.
Overall, the humans and the fish performed the tasks similarly. For example, their speed to spot a differently-colored circle did not change with additional distractors. Similarly, both groups were slower to find a small blue circle when there were more distractors.
But, there were some interesting differences. While humans were equally accurate at both the feature (e.g. color) and shape (e.g. Q) searches, the archerfish were much less accurate on the shape tasks. In addition, both shape tasks elicited pop-out effects in the humans, but the archerfish had to search serially for the disk among Pac-men. The authors speculate that the difference may be due to the archerfish’s unfamiliarity with the shapes or even that the Pac-men may look like the open mouth of a predator fish.
It’s a rare theory that is still spurring new research 40 years later, but the insights documented by Treisman decades ago are still influencing research questions today. Feature integration theory illustrates not just how the human visual system operates, but also how many animals search their visual world, even the archerfish.
Psychonomic Society article considered in this post:
Reichenthal, A., Segev, R., & Ben-Shahar, O. (2020). Feature integration theory in non-humans: Spotlight on the archerfish. Attention, Perception, & Psychophysics, 1-23. https://doi.org/10.3758/s13414-019-01884-4
