7 Minutes
Cuttlefish and the 'Marshmallow' Paradigm
An experiment adapted from the classic human cognitive test commonly known as the 'marshmallow test' has produced compelling evidence that cuttlefish possess advanced decision-making abilities. Originally developed to measure delayed gratification and future-oriented choices in children, the marshmallow paradigm can be retooled for nonverbal animals by substituting food rewards and using learned cues to signal timing. When researchers applied such a protocol to common cuttlefish, the results added to a growing body of work showing sophisticated cognition among cephalopods.
Cuttlefish are known for rapid color change, complex camouflage, and flexible hunting strategies. These behaviors suggest neural computations that support perception, memory, and behavioral control. The study discussed here examined whether cuttlefish can forgo an immediately available, lower-value prey item in favor of a delayed, higher-value live prey — a direct test of delay of gratification, future reward anticipation, and adaptive self-control in an invertebrate.
Experiment Design and Methods
The research team placed six common cuttlefish in an experimental tank fitted with two transparent, enclosed chambers. Each chamber contained one type of food: a less-preferred raw king prawn in one chamber and a highly preferred live grass shrimp in the other. The chambers were separated from the animals by transparent doors displaying distinct, trained symbols: a circle signaled an immediately opening door; a triangle signaled a door that would open after a delay (ranging from 10 to 130 seconds); and a square, used only in control trials, indicated that the door would remain closed and the prey inaccessible.
The key test condition presented the prawn behind the immediate-open (circle) door and the live shrimp behind the delayed (triangle) door. If a cuttlefish entered and consumed the prawn before the delayed door opened, the live shrimp was removed and became unavailable. In control trials the shrimp was shown behind a permanently closed (square) door so animals could see but could not access it. This design prevented simple preference for visible prey from confounding evaluations of waiting behavior.
The researchers also ran a separate associative-learning task to assess cognitive flexibility. Animals were trained to associate one visual cue with a reward and, after reaching criterion, the reward contingency was reversed. Reaction times, choice accuracy, and the number of trials to learn the reversed association provided metrics of learning speed and flexibility.
Results: Delays, Learning, and Correlations
The cuttlefish in the delayed-reward condition consistently waited for the live shrimp, tolerating delays between 50 and 130 seconds in many trials. By contrast, those in the control condition did not display the same waiting behavior when the preferred item was demonstrably inaccessible. This pattern indicates that the animals were not simply habituated to waiting but were making decisions contingent on the anticipated availability of a superior reward.
Cuttlefish performance on the associative-learning reversal task varied between individuals. Notably, individuals that adapted fastest to the switched visual cue (i.e., learned the new contingency quickly) were also the ones most likely to wait longer for the delayed shrimp reward. This correlation suggests an association between cognitive flexibility and tolerance for delayed gratification: individuals with faster rule-switching may also exert stronger behavioral control in pursuit of future outcomes.
The experimental setup and illustration of stimuli are shown in the original publication and captioned as follows:
The experimental setup. (Schnell et al., Proc. R. Soc. B, 2021)
Interpretation and Comparative Context
Researchers pointed out that the delay durations tolerated by cuttlefish are comparable to those reported in large-brained vertebrates such as chimpanzees, corvids, and parrots. In species where delayed gratification has been linked to tool use, food caching, or complex social behavior, the evolutionary drivers are relatively intuitive: planning, memory for stored resources, or social coordination can reward patience. Cuttlefish, however, do not typically use tools or cache food, nor are they strongly social, so those conventional explanations are less applicable.
The team proposed an ecological hypothesis: cuttlefish spend extended periods camouflaged and motionless while waiting to strike prey. Breaking camouflage during active foraging increases exposure to predators. Selection may therefore favor restraint and optimized decision rules that minimize unnecessary foraging events and prioritize higher-quality prey when available. In this view, delayed gratification emerges as an adaptive byproduct of a sit-and-wait foraging strategy and the associated predation risk management.
Evidence of episodic-like memory in cuttlefish has also been reported, and follow-up studies have suggested even more complex memory dynamics, including reports of false-memory-like phenomena in 2024. Together, these findings portray a cephalopod mind capable of flexible learning, memory-based decisions, and behavioral inhibition.
Expert Insight
Dr. Elena Morales, a cognitive ecologist (fictional), comments: 'These data help reposition cephalopods in the comparative cognition landscape. The combination of delay tolerance and rapid associative learning suggests neural mechanisms for planning-like behavior, even in the absence of mammalian-style social or caching pressures. Future work should probe neural correlates and whether cuttlefish can use prospective cues to plan multi-step actions.'
This expert perspective underscores two research priorities: linking behavior to neural processes in cephalopods, and testing whether delayed choices reflect true future planning or a different computational strategy such as updated value estimation based on learned contingencies.
Implications and Future Directions
The study highlights convergent evolution of cognition: disparate taxa can evolve similar decision-making strategies when ecological pressures favor certain behavioral solutions. For cephalopod researchers, the findings motivate experiments that test longer delays, alternative reward modalities, and the extent to which cuttlefish can generalize waiting behavior to novel contexts. Neurobiological studies could identify circuits that support inhibition and prospective evaluation in cephalopods, providing comparative insight into how complex cognition arises in diverse nervous systems.
Researchers also recommend longitudinal and developmental studies to determine when delay tolerance emerges during cuttlefish ontogeny, and whether experience or individual temperament predicts patience and learning speed. Integrative approaches combining behavior, ecology, and neurophysiology will be essential to establish whether cuttlefish are planning for future states in a way analogous to some vertebrates, or using different proximate mechanisms to achieve similar outcomes.
Cuttlefish can also change their body's color patterns to camouflage or signal.
Conclusion
This adaptation of the marshmallow paradigm for cuttlefish provides robust evidence that Sepia officinalis can delay gratification for a superior food reward and that this ability correlates with cognitive flexibility. The results expand our understanding of animal intelligence and illustrate how ecological constraints — foraging strategy and predation risk — can shape the evolution of self-control and future-oriented decision-making. Further experimental and neurobiological work will clarify whether cuttlefish truly plan for anticipated future states or accomplish similar outcomes through alternative cognitive processes.
Source: sciencealert
Comments