A BGU study has demonstrated that the way fish perform visual search resembles the way humans do and provides an important insight in to the way the visual system works.
Visual search, the ability to find an object of interest against a background, is a task humans encounter on a daily basis. While looking for a car in the parking lot or searching for a key among the many items on a desk, we rely on visual features that can make distinct the object of interest from its surroundings. For animals, from detecting food items to locating lurking predators, this ability has to be accurate and fast to ensure survival.
In mammals, including humans, the need for fast visual search led to the development of an important brain capability that enables the scanning of the visual scene in a parallel fashion. That is, a concurrent processing of the entire visual scene as opposed to serial processing of object by object. This ability, called pop-out visual search, enables the observer to rapidly detect the object of interest. Pop-out search is believed to be based on a population of brain cells that together generate a map that analyzes every location in the visual scene to determine how different it is from its surroundings. This map simplifies the representation of the visual field and highlights the location which differs the most from its nearby regions. During casual viewing, the observer’s attention is drawn towards this location. Over the years, much work has been devoted to showing that these specialized brain cells are located in the cortex, which is the largest and most recently evolved area in the mammalian brain.
Currently, most research has chosen to focus on monkeys and humans to study visual search. However, studying an animal which is an earlier evolutionary life form can offer a new perspective on understanding the human brain by exploring the evolutionary and developmental perspectives of important behaviors.
In a recent study published in Nature Communications, the BGU group followed this approach and focused on the archer fish, an animal which lacks a fully developed cortex. The archer fish exhibits complex visual hunting behavior as it shoots down prey found on foliage above the water level and is able to learn to distinguish between artificial targets presented on a computer monitor in an experimental setting. The BGU group has levered this exceptional visual behavior to study visual search in lower vertebrate.
The archer fish, as this study reveals, exhibits a pop-out visual search mode of moving targets. That is, the fish was able to detect a target that moved faster than the other objects presented on the screen. Importantly, this detection was not affected by the number of objects in the display, an indicative property of pop-out. This is the first evidence for pop-out search mode in a non-mammal and in particular, non-primate animal.
To understand the computation needed for this behavior in the archer fish brain, the researchers recorded the activity of the archer fish brain cells from the optic tectum. The optic tectum is the main visual area and one of the largest areas in the archer fish brain. This area also exists in primates. However, due to the massive expansion of the cortex, this area comprises a much smaller fraction of the primate’s brain. The new study shows that the activity patterns of the archer fish brain cells are similar to those found in the mammalian visual cortex. This implies that the visual cortex may not be the crucial component in generating the map of the unique locations in the visual field.
The BGU team results show that a pop-out search mode exists in fish and it may indicate that parallel search mode may be a common, and perhaps even universal, mechanism across vertebrates’ visual systems.
The team included Prof. Ronen Segev and PhD student Mor Ben-Tov, from the Department of Life Sciences, Prof. Opher Donchin from the Department of Biomedical Engineering and Prof. Ohad Ben-Shahar from the Department of Computer Science.
