When Picking Dinner, Cuttlefish Can Make Some Complex Decisions


Tom Hale

Tom is a writer in London with a Master's degree in Journalism whose editorial work covers anything from health and the environment to technology and archaeology.

Senior Journalist

Cuttlefish are part of an order of animals known as cephloads, which also includes squid and octopus. DiveIvanov/

Cephalopods might just look like a squiggly blob of rubbery tentacles, but there’s a wealth of evidence that shows that this strange class of animal possesses incredible cognitive abilities. In their latest display of intelligence, scientists have found that cuttlefish are able to make some sharp, informed decisions even when tempted by the prospect of more food, a bit like a cephalopod “marshmallow-test.”

As reported in the journal Royal Society Open Science earlier this month, a duo of scientists from the National Tsing Hua University of Taiwan assessed whether pharaoh cuttlefish (Sepia pharaonis) were able to learn that more food is not always the best choice to make. 


Optimal foraging theory explains that an animal looking for food will adopt a strategy that provides the most energy for the lowest cost. Given the choice, you’d expect an animal to always pick two easily-caught snacks over just one. However, as these cuttlefish show, some animals can employ a much more complex strategy that’s not strictly driven by utility, but guided by a bunch of other brainy decision-making abilities. 

To understand how cuttlefish make decisions while foraging, the researchers put some of these small cephalopods through a “training phase” involving a tank where they could pick one of two chambers: one containing a single shrimp, and another that was empty. Obviously, the cuttlefish most often went towards the chamber containing a shrimp at first. If the cuttlefish chose one shrimp, they were not fed the shrimp within the chamber and only received a tiny shrimp as a reward. According to the researchers, this “priming raises the ‘value' of one shrimp significantly,” effectively drilling it into their heads that a one-shrimp chamber was a safe bet.  

The second stage of the experiment saw the cuttlefish choose between a chamber of one shrimp or a chamber of two shrimps. Cuttlefish that hadn’t gone through the “training phase” picked two shrimp over one, as you might have expected. However, those that had received the training were far more likely to choose the single shrimp or two shrimp, based on their previous experience that this choice tends to pay off even if it isn’t immediately apparent. 

Effectively, the researchers had manipulated the cuttlefish to prefer one shrimp in a choice between one and two shrimps. This might not sound like much to a wise human such as yourself, but the researchers argue this shows some complex “value-based decision-making.” The cuttlefishes’ foraging is not simply guided by simple impulse – “more food = good” – but a relative value perception and judgment that’s based on their recent experience. 


As the New York Times pointed out, this new study has some comparisons to the Stanford marshmallow-test,” a classic psychological experiment first carried out in the 1950s in which kids were offered a choice between one immediate reward or two rewards if they waited for a period of time. While there are some similarities to this cuttlefish study, there are notable differences, namely that the animal is not displaying any real sense of self‐control or “future thinking,” but simply reinforced learning.

Nevertheless, scientists have previously toyed with the idea of carrying out delayed‐gratification studies on cephalopods, including cuttlefish. Whether these brainy freaks of the sea are up for the task, however, remains unanswered for now.