The Etruscan shrew, Suncus etruscus (also called the white-toothed pygmy shrew), is one of the smallest mammals in the world. It leads a nocturnal lifestyle and has a limited sense of vision, relying instead on hearing, smell and especially touch, on which it is highly dependent.
The visual cues used by toads and frogs to capture prey have been studied in detail, as has the visual representation of objects by primates. But the use of tactile information for object representation is poorly understood.
A study by a team of European researchers shows that tactile object recognition in the Etruscan shrew shares some of the characteristics of human vision and, even though the shrew’s brain is 20,000 times smaller than that of a human, and weighs approximately 60 milligrams, object recognition occurs much faster in the shrew than does the processing of visual information in humans.
The Etruscan shrew is insectivorous, and has a diet consisting of large numbers of small insects. The researchers used high-speed cameras to determine how quickly the shrew can mount and complete an attack on a cricket. Large crickets are almost as big as the shrew itself (see the picture above), and killing or immobilizing them requires repeated attacks by the shrews. These attacks are mounted with incredible speed and precision. A frame-by-frame analysis revealed that during an attack the shrew protracted its snout – or rostrum – towards the cricket. Attacks took an average of 200 milliseconds, with the fastest ones taking only 80 ms. There was an interval of 200 ms between attacks; thus, the shrew can complete between 3 and 5 attacks per second.
The researchers were surprised to find that the shrews used in their experiments had a preference for attacking from the left; this is taken as evidence for lateralization in the shrew’s brain of the neural structures mediating the attack behaviour. The vast majority of attacks are aimed at the anterior thorax of the cricket, this region of the cricket was attacked >20 times more than the head. The anterior thorax of the cricket contains the thoracic ganglia, groups of nerve cell bodies; the disruption of neural activity in the thoracic ganglia leads to paralysis of the cricket. Other organisms which prey on crickets, such as the digger wasp, also target these structures. The selectivity of the shrews for the anterior thorax is quite remarkable – the head is located just 3mm away from the preferred attack location, and crickets can move extremely quickly.
The shrew has two sets of whiskers, or vibrissae, on its face: a set of long macrovibrissae protrude from the side of the face, and a set of smaller microvibrissae surround the mouth. It is primarily with the whiskers that the shrew comes into contact with prey. Like mice, shrews use their whiskers to recognize objects; the shrew can also discriminate between smooth and rough textures from the tactile data obtained via the whiskers.
Using the high-speed camera, the research team showed that, during an attack on prey, the Etruscan shrew’s whiskers are highly mobile, moving back and forth at frequencies of about 20 Hz (20 times per second) to provide tactile data that are used to produce tactile representations of the prey.
As in other species, the sensory information from the shrew’s whiskers is transmitted to an area of the brain that is devoted to processing that information. Comparative neuroanatomical studies show that shrews compensate for their small brain size by having only a small number of specialized subdivisions within the neocortex. The dependence of the shrew on the sense of touch are reflected in the organization of the cortex – the somatosensory area, which processes information from the whiskers, is much larger than the visual cortex.
Experiments in which whiskers were removed were then performed. The removal of microvibrissae led to a reduction in the number of complete attacks by about 50%; targeting of the anterior thorax was affected little; removal of the macrovibrissae led to a similar drop in the number of completed attacks, but shrews lacking macrovibrissae were very clumsy, and attacked crickets’ heads far more frequently than they normally would.
The team then examined the sensory cues used by the shrew to catch prey. They then performed three-dimensional scans of a cricket and used the data obtained to make accurate plastic dummy crickets, which were presented to the shrews along with various control objects. About 40% of attacks were carried out on the stationary dummy crickets, ruling out motion as a cue for evoking attack, and showing that the shrews do not use olfactory cues for finding targets. The shrews did not discriminate between small or large targets.
By manipulating the shape of the dummy crickets, the experimenters showed that the shrews use shape as their primary cue. The addition of an extra head to the dummy had little effect on attack behaviour, but an additional pair of legs altered the shrews’ attacks dramatically. Rather than aiming for the thorax, as they would normally, the shrews targeted most of their attacks at the legs. This support the notion that shrews use their whiskers to extract a few key stimuli – jumping legs – during the recognition of objects as prey.
The extraction of an object’s key features is necessary because although the shrew can mount and complete an attack within 200 milliseconds, its whiskers can only perform about 20 movements per second. This places an upper limit on the number of tactil object representations that can be processed during an attack, and that limit is quite low. Thus the authors conclude that “shrew behavior is guided by Gestalt-like prey descriptions”.
‘Gestalt’ can be translated as shape, figure or form. The Gestalt theorists were the first psychologists to study the perceptual organization of objects. They found that the individual parts of an object have different characteristics to the object as a whole. The human brain can perceive the parts, but normally integrates them to produce a representation of the object as a whole. In some cases, only a few parts of an object are visible, and the brain ‘fills in’ the gaps to produce a representation of the object. (The image on the left, for example, is perceived by most as a square with a circle centred on each of its corners, even in the total absence of any of the lines of the square.)
This is what the shrew appears to do when finding prey, at least in the experimental set-up described in this paper. The shrew uses its whiskers to feel potential prey and, if it recognizes jumping legs, its brain fills in the missing parts, and produces a tactile representation of a cricket, which the shrew then proceeds to attack. Successful prey capture of course depends upon the shrew’s memories of previous tactile representations, without which prey could not be recognized. Evidently, shrews are also capable of abstract sensory representations of objects, a cognitive task that was previously thought to be restricted to primates.
Reference: Anjum, et al (2006). Tactile guidance of prey capture in Etruscan shrews. Proceedings of the National Academy of Sciences. 10.1073/pnas.0605573103