Summary: If a behavior reliably earns a reward, most models of learning predict it will be repeated and eventually become dominant. A new study from the University of Iowa challenges that view. Researchers rewarded pigeons with food when they pecked five buttons in any order out of 120 possible sequences. Rather than converging on a single perfect sequence, the birds favored some patterns but continually switched among them—staying flexible and exploratory instead of settling into a fixed routine.
Over thousands of trials the pigeons demonstrated persistent variability in their choices. The authors interpret this as behavior that operates “at the edge of chaos”: stable enough to exploit successful patterns but variable enough to adapt if conditions change. This finding calls for a refined understanding of the century-old Law of Effect and highlights a biological inclination toward strategic unpredictability that may have broad relevance to cognition and creativity.
Key Facts
- The experiment: Six pigeons completed the five-button task roughly 30,000 times across eight months. Any of the 120 possible five-button sequences produced a food reward.
- Challenge to the Law of Effect: Classic theory predicts animals will concentrate on the sequence that yields reward most efficiently. The pigeons did narrow their repertoire but never fixed on a single sequence.
- Edge of chaos: Lead author Ed Wasserman describes the pattern as an adaptive balance—enough regularity to benefit from success, enough variability to avoid being “machinelike.”
- Persistent exploration: Even after many repetitions, pigeons continued to sample less-common sequences, maintaining a dynamic range of behaviors.
- Broader implications: The researchers suggest this adaptive variation may be a biological basis for flexible, creative behaviors in other species, including humans.
Source: University of Iowa
If a specific pattern always brings a reward, why not stick with it?
The answer from this study is nuanced. The birds did reduce the total variety of sequences they used, consistent with reinforcement learning in principle, but they never abandoned exploration entirely. Instead, their preferred sequences rose and fell in popularity over time. This shifting preference indicates an intrinsic tendency to keep multiple successful options available rather than converging on only one.
Researchers placed each pigeon in an individual chamber equipped with a touchscreen displaying five lit buttons. Each button showed a distinct colored geometric pattern that disappeared when pecked. Once a pigeon pecked all five buttons—regardless of order—it received a food reward. Because every one of the 120 possible sequences produced the same outcome, the setup created ideal conditions to test whether animals would lock onto a single, consistently rewarded action sequence.
Consistent with the Law of Effect, pigeons gradually emphasized some sequences more than others. But contrary to expectation, none of the six birds became fully committed to a single sequence. Over the course of the study, each bird produced every possible sequence at least once and repeatedly switched among favored patterns. These fluctuations persisted across the 30,000 trials, producing what the authors call “dramatic behavioral instability.”
Ed Wasserman, professor in the Department of Psychological and Brain Sciences and corresponding author, emphasizes that this instability is likely adaptive. “There appears to be a built-in resistance to becoming fully machinelike,” he explains. “Keeping some variability in behavior may protect an animal if the environment changes and the previously successful habit no longer works.”
Odysseus Orr, a co-author and graduate student, notes that the phenomenon extends the idea of the edge of chaos beyond evolutionary explanations. “Adaptive variation may underlie more complex creative behaviors,” he says. “Studying pigeons in tightly controlled conditions gives us a window into the biological mechanisms that encourage deviation from habitual success—an engine of novelty that could also be relevant for music, art, and innovation.”
The findings align with earlier reward-based studies showing animals reduce their action sets when one option reliably pays off. However, this experiment makes clear that reduction in variety does not imply elimination of exploration. Pigeons maintained an exploratory tendency, repeatedly sampling less-common sequences and avoiding permanent fixation on any single pattern.
Funding:
Research was supported by Wasserman’s Comparative Cognition Laboratory.
Key Questions Answered:
A: The surprising result is that simple efficiency would be risky in a changing world. By keeping many successful sequences available, pigeons hedge against potential changes in the environment. That flexibility likely has survival value.
A: The study suggests animals possess an evolved exploratory drive—an innate neural tendency to avoid getting stuck in repetitive loops. It’s less about conscious strategy and more about a built-in mechanism promoting adaptive variability.
A: Both processes require deviating from what worked previously. The pigeons’ persistent variation hints at a basic biological capacity for adaptive novelty. By studying this capacity in controlled tasks, researchers can better understand the foundational mechanisms that enable complex creative behaviors.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full.
- Additional context was added by staff.
About this cognition and neuroscience research news
Author: Richard Lewis
Source: University of Iowa
Contact: Richard Lewis – University of Iowa
Image: The image is credited to Neuroscience News
Original Research: The findings will appear in Journal of Experimental Psychology: Animal Learning and Cognition