Study: archerfish can learn to accurately recognize human faces

Fish are generally regarded as dumb, but maybe that's not a fair assumption. At least one type of fish, the archerfish, is able to recognize human faces, according to a new study. Researchers with the University of Queensland in Australia and the University of Oxford found that archerfish can learn to recognize different people with a pretty high level of accuracy, doing so despite its lack of a superior visual cortex, the kind found in primates.

Though the archerfish's brain does not contain a neocortex, researchers found the fish is able to recognize a specific human face from varying numbers of faces up to 44. How does the fish let researchers know which face it is choosing?

The archerfish, as you may be aware, is able to squirt a jet of water out of its mouth, doing so typically to knock prey out of the sky (like a bug, for example). In working with the fish, researchers trained them to choose a face from a group of faces by shooting it with a jet of water. Not only could the fish choose a specific face, but it could still find it even if some big element was taken away, such as the shape of the head.

When picking a specific face out of a group of 44 faces, the fish had an accuracy of 81-percent, a number that increased in a second experiment where face color and brightness was processed to be similar across the board.

The findings call into question whether it is necessary to have a specific and complex portion of the brain found in primate to distinguish human faces. Said Oxford's Dr. Cait Newport:

Being able to distinguish between a large number of human faces is a surprisingly difficult task, mainly due to the fact that all human faces share the same basic features. All faces have two eyes above a nose and mouth, therefore to tell people apart we must be able to identify subtle differences in their features. If you consider the similarities in appearance between some family members, this task can be very difficult indeed.

SOURCE: Phys.org