Invertebrate Cognition

The bodies, nervous systems, behavioral repertoires, ecology, social organization, and evolution of the plethora of invertebrates are so different from those of vertebrates in general, and of mammals and birds in particular, that the question of “cognition in invertebrates” sounds farfetched. Despite these differences, invertebrates also face problems in procuring food, attracting mates, finding safe locations, and other familiar actions from the study of vertebrate behavior. We also know that although their nervous systems are organized in unique ways, their neurons have familiar properties. And, ultimately, their environment includes two of the same dimensions common to all environments: time and space. These are the physical dimensions underlying the evolution of behavioral plasticity.

As with nonassociative (Chapter 4) and associative learning (previous sections), a meaningful start is to ask what types of behavioral phenomena described in any invertebrate species could qualify as cognitive. The question of mechanism will arise once dependable behavioral data on invertebrate cognition are available. Animal cognition is fully characterized in Chapter 8, but for the purpose of this chapter, cognition is here identified with problem solving. Not surprisingly, when looking at potential cognitive tasks in invertebrates, the

Table 7.3 Partial list of learning phenomena known from experiments with honeybees and vertebrates

Note: The control groups were omitted for brevity. See Bitterman (1988, 1996) for reviews of learning in honeybees.

attention focuses on insects and cephalopods, two groups that exhibit relatively complex nervous systems and behavior. Two examples are described next.

Consider the following problem pose to bumblebees (Bombus terrestris) foraging for sucrose (Alem et al., 2016). If placed on top of a blue artificial flower with a drop of sucrose in the center, the bee will drink the sucrose immediately. If then the sucrose is covered partially, but the bee has access to it, it will still drink it. Now, the flower is moved progressively under the cover, but if the bee pulls from a string attached to the flower, then it can bring it closer and access the reward. As the flower is moved farther away, the animal needs to pull more to get to the sucrose (Figure 7.6A). The authors report that 23 of 40 bees trained in this task actually learned to pull from the string. Bees would grab the string with their forelegs or mandibles to move the flower closer. However, when tested with the sucrose fully covered, bees were unable to solve this problem in their first trial, so a response-shaping procedure was required. Moreover, naive bees that could not see the flower moving closer did not acquire the behavior, whereas trained individuals were capable of pulling the string in the absence of visual feedback from the moving flower. Interestingly, 50 bees were given experience with the sucrose on the blue flower and then were shown the fully covered target with the string; two individuals actually solved the problem on their own after two sessions of exposure. These innovators, individuals that exhibit problem-solving capacity with minimal pretraining, displayed an unusual amount of exploratory behavior, trying several methods until they were successful. This behavior is reminiscent of the problem-solving behavior of Thorndike’s (1911) cats, which was characterized as trial-and-error learning (Chapter 6). The behavior of these innovators that solved the string-pulling problem without shaping comes close to an example of cognition.

Figure 7.6 Problem-solving behavior. (A) A bumblebee (Bombus terrestris) has learned to pull from a string to move an artificial flower with a drop of sucrose on the center (round disc) that was made inaccessible by a translucent cover. (B) An octopus (Octopus vulgaris) explores and attacks a clear jar, opens the lid, inserts an arm in the jar, and seizes the crab. See text for references.

Source: (A) Photo by Sylvain Alem, courtesy of Lars Chittka. (B) Reproduced with permission from Elsevier.

Bumblebees are highly social and also visually oriented. Would naive individuals that observe trained demonstrators pull the string to obtain sucrose be able to imitate them? After pretraining to obtain sucrose from the blue flower, observers were allowed to see a trained bee pull the string and drink the sucrose 10 times, although no direct interaction was allowed. When placed in the problem situation, 15 of 25 observers successfully pulled from the string on the first trial. A group of observers that did not see a trained bee pull from the flower (which was pulled by the experimenter using a thin nylon thread), but saw a bee drinking the sucrose once the flower was in place, did not learn the string-pulling task.

Cephalopods (octopuses, cuttlefish) also show problem-solving capacity associated with sustained exploratory behavior (Mather & Dickel, 2017). In one experiment with octopuses (Octopus vulgaris) housed in individual tanks (Fiorito, von Planta, & Scotto, 1990), a glass jar with a lid containing a crab (a typical prey) was placed in the center of the tank and left there for 5 min. For a group of 20 octopuses, it took them about 5 trials to approach, open the jar, and seize the crab (only one animal failed to seize the crab). Repeating this procedure across days resulted in the development of a more efficient predatory sequence, which can be attributed to instrumental conditioning. Once the crab was seized for the first time, all the subsequent trials ended with the octopus grabbing and eating the crab (Figure 7.6B). Unscrewing the lid for the first time is a candidate for a cognitive phenomenon, although the fact that these animals displayed a persistent exploratory behavior, even after solving this problem many times, suggests the possibility that the solution occurred by chance, as a byproduct of tactual exploration of the jar and lid.

Intriguing evidence of possible tool use in the octopus Amphioctopus marginatus comes from field observations (Finn, Tregenza, & Tregenza, 2009). Individuals have been observed to carry discarded coconut shell halves for some distance under their body to later use them as shelters. Upon arriving at some location, an individual was observed to arrange the halves to create a housing device. This was interpreted as tool using. Other animals were observed to dig coconut shells partially buried in the substrate by jetting water to clean the mud. The octopus Tremoctopus violaceius was found to collect fragments of tentacles of the Portuguese man-of-war (Physalia physalis) perhaps to serve predatory or defensive functions (Jones, 1963).

These possible cognitive abilities of insects and cephalopods may be dismissed as instances of associative learning. However, as shown in Chapters 5 and 8, complex problem-solving abilities often derive from a combination of behavioral flexibility and conditioning to develop into capacities used for solving new problems (e.g., see Figure 5.20 and 5.21).