The short version
- Carrion crows completed a visual motion task in roughly 235 milliseconds, approximately half the time required by primates.
- Neural recordings indicate that avian brains use the same mathematical logic for evidence accumulation as mammalian neocortices, despite structural differences.
- The findings suggest that compact, clustered brain tissue can process complex decisions rapidly without sacrificing accuracy compared to layered cortical structures.
A recent study led by neurobiologist Andreas Nieder at the University of Tübingen has demonstrated that carrion crows evaluate visual information with remarkable speed, outpacing primates in controlled decision-making tasks. The research addresses a long-standing question in neuroscience regarding how brains with different evolutionary architectures process similar cognitive challenges. While much of our understanding of decision-making stems from studies on monkeys and humans, avian brains lack the layered neocortex found in mammals. Instead, bird brains are organized into dense clusters without layering, a structural divergence that occurred over 360 million years ago.
To compare these systems directly, Nieder and colleague Philipp Schmidbauer subjected two hand-raised crows to a random-dot-motion task, a standard benchmark used extensively with primates. In this experiment, subjects view a cloud of flickering dots on a screen, where a specific fraction moves coherently in one direction while the rest drift randomly. The subject must determine the overall direction of movement. For the monkeys, this typically involves making an eye movement to indicate a choice; for the crows, the birds pecked at targets flanking the display. Correct responses were rewarded with mealworms.
The results showed that the crows solved the task using the same underlying mathematical logic as primates, accumulating visual evidence until a threshold was crossed. However, they did so significantly faster. From the moment the dots appeared to the initiation of the peck, the crows averaged 235 milliseconds. In comparable studies involving monkeys, reaction times ranged from roughly 410 to 600 milliseconds. This speed difference persisted even though the birds were not forced to rush; they operated at a self-paced rhythm, suggesting that their rapid processing is an inherent feature of their neural architecture rather than a result of external pressure.
Researchers varied the difficulty of the task by adjusting the percentage of dots moving coherently. The conditions ranged from 81 percent coherence, which is easily discernible, down to just 3 percent, where the signal is barely distinguishable from noise. A zero-percent condition was also included to test performance when no correct answer existed. Both crows performed better than chance even at the lowest coherence levels, demonstrating high sensitivity to subtle visual cues. This accuracy indicates that the birds did not sacrifice precision for speed, a common trade-off in rapid decision-making scenarios.
To understand the neural mechanisms behind this performance, the team recorded activity from 248 neurons in the nidopallium caudolaterale (NCL). This region is considered the functional counterpart to the primate prefrontal cortex, despite its different structural organization. The recordings revealed that approximately 58 of these neurons fired differently depending on the direction the bird would eventually choose. These neurons began registering the decision roughly 70 milliseconds after the dots appeared, about 170 milliseconds before the bird physically moved to peck.
The pattern of neuronal firing closely mirrored evidence accumulation seen in mammalian neocortices. As the birds processed the visual input, their neural activity climbed until it crossed a threshold, triggering a decision. This finding suggests that convergent evolution has led to similar computational strategies for handling uncertainty and making choices, even in brains that evolved independently over hundreds of millions of years. The compact nature of the avian nervous system appears to facilitate this rapid processing without compromising the integrity of the decision.
Initially, data suggested that crows might operate on a fixed deadline, as their reaction times varied only slightly between easy and difficult tasks—ranging from 220 milliseconds for high-coherence trials to 254 milliseconds for low-coherence ones. However, closer statistical analysis revealed a slight increase in reaction time as task difficulty increased, similar to the pattern observed in mammals. Nieder noted that this effect was less pronounced simply because the entire process occurs much faster in birds. This indicates that crows are not merely guessing on a timer but are genuinely accumulating sensory evidence.
The study underscores the efficiency of avian brain architecture. While mammalian brains rely on stacked sheets of cells, the dense clustering in bird brains allows for rapid information processing. This has implications for understanding cognitive evolution and the diversity of neural solutions to complex problems. The ability of crows to match or exceed primate performance in speed while maintaining accuracy challenges assumptions that larger, layered brains are necessary for sophisticated decision-making.
Future research may explore how these findings apply to other avian species and whether similar efficiencies exist in other non-mammalian vertebrates. The work also highlights the importance of studying diverse animal models to gain a fuller picture of neural computation. By demonstrating that different brain structures can achieve similar cognitive outcomes through distinct pathways, this study expands our understanding of the biological basis of thought and decision-making.
As neuroscience continues to bridge gaps between species, these insights may inform artificial intelligence development, particularly in designing systems that balance speed and accuracy. The crows’ ability to process visual information rapidly without error offers a compelling model for efficient computation. Ultimately, this research illustrates that intelligence is not confined to a single architectural blueprint but can emerge from various evolutionary paths, each with its own unique advantages.
Sources behind this briefing
Go to the original reporting
- Ars Technica↗Crows can evaluate visual info twice as fast as mammals
