Baby Rats Remember Differently: Hidden Exploratory Strategies Reveal What Standard Memory Tests Miss

For nearly four decades, one of the most popular tools in neuroscience has been deceptively simple. Place a rodent in an arena, let it inspect two identical objects, swap one of them for something new, and measure how much time the animal spends investigating the newcomer. If the animal prefers the novel object, researchers infer that it remembers the familiar one. This logic underpins both novel object recognition and its spatial cousin, novel object location memory, and it has generated an enormous literature on how memory emerges, matures, and declines. But a new perspective article published in Frontiers in Behavioral Neuroscience argues that this field has been looking at only a sliver of the data. Neuroscientists Eleanor Williamson, Jesse Cushman, and Patrese Robinson of Bryn Mawr and Haverford Colleges contend that the single discrimination ratio most laboratories rely on may be quietly hiding some of the most interesting things developing brains are doing.

The core problem, the authors explain, is a narrowing of analytical ambition. In most rodent recognition studies, memory is inferred almost exclusively from a discrimination ratio or index — a mathematical summary that collapses everything the animal did during a test session into one number reflecting preferential exploration of the novel stimulus. That number is useful, but it is also impoverished. Two animals can arrive at an identical discrimination ratio through completely different patterns of behavior: one may engage in a cautious, proximity-limited sniffing strategy while the other climbs over, sits atop, and physically manipulates the objects. If the number is the same, the difference vanishes from the record. The authors argue that this reduction may be especially costly in developmental studies, where young animals are not simply small versions of adults but are actively assembling the neural machinery of memory in real time.

Their argument draws strength from an instructive precedent in fear conditioning. For a century, freezing was the gold standard measure of learned fear, and many laboratories still report nothing else. But when researchers expanded their analyses to include an active escape-like behavior called darting, previously invisible differences emerged. Female rats dart more than males during tone-fear conditioning, and darting correlates with distinct patterns of neural activation in the lateral and basolateral amygdala and the lateral periaqueductal gray. When overtraining pushes males toward darting, it appears to reflect delayed recruitment of that same circuitry. None of this neurobiological information was available from freezing scores alone. The lesson is that a single behavioral measure can function as a bottleneck, and the fear conditioning field offers a template for breaking through it in recognition memory research.

The case for richer measurement also rests on a methodological survey that the authors compiled across the recognition memory literature. Their table of representative publications reveals striking inconsistency in nearly every procedural variable. Laboratories differ in how they define exploration: some count only nose-directed sniffing within two centimeters, others accept proximity of one or four centimeters, some include paw contact, and some simply require the head to be oriented toward the object. Handling regimens, habituation windows, and retention intervals vary widely, and the objects themselves range from Lego Duplo blocks and glass beakers to PVC pipe and plastic toy animals. Because object properties demonstrably influence exploration and performance, this variability makes cross-laboratory comparisons fragile and may mean that some published differences in memory reflect stimulus properties rather than cognitive capacity.

Why would objects matter so much? The answer lies in the ecological concept of affordances — the action possibilities that the environment offers a particular animal. A flat-topped, climbable block affords climbing and sitting; a pointed, symmetrical cone affords little more than touch. Two decades ago, researchers Chemero and Heyser warned neuroscientists that ignoring affordances leads to poor science, and they demonstrated the point empirically: mice exploring flat, climbable objects during short-term recognition tests explored more, habituated more slowly, and showed stronger recognition than mice given touch-only objects. The broader literature, the authors note, has largely failed to heed that warning. Objects in recognition tasks are frequently chosen for convenience rather than characterized for their ecological significance, even though the object itself dictates which exploration behaviors an animal can display and therefore what data the test can yield.

The new developmental data that anchor this perspective make the stakes vivid. The team tested pre-weanling rats around postnatal day 21 and adult rats on a novel object location task using 3D-printed objects scaled to the animals. One object was asymmetrical with a flat, climbable top that provided clear spatially orientable cues; the other was radially symmetrical and pointed, impeding climbing and offering no spatial structure. On the traditional discrimination ratio measure, the story seemed straightforward. Both ages showed comparable short-term spatial memory when the climbable, asymmetrical object was used, and neither age showed a novelty preference with the pointed object. A conventional analysis would stop there and conclude that there were no age differences in spatial memory.

But when the researchers examined the animals’ behavior beyond object-directed exploration, a striking developmental difference appeared. Adult rats, when not directly investigating the objects, spent most of their time rearing on their hind legs in the open arena. Pre-weanlings did something entirely different: they switched their exploratory strategy depending on which object was present. With the flat, climbable object, pups spent significantly more time sitting on or climbing the objects than rearing in the arena. With the pointed object, the pattern reversed and became more adult-like, with rearing in the open arena dominating. Crucially, this object-dependent switching appeared only in the spatial task; during a non-spatial object identity test, pre-weanlings behaved like adults, and the object-type effects disappeared. The behavioral shift was therefore tied specifically to spatial processing, not to a general preference for climbing.

The authors propose that this switch is not merely a response to a physical challenge, but a signature of neurobehavioral ontogeny. The timing is suggestive: around postnatal days 20 to 23, place cell distributions and firing patterns in the developing hippocampus shift toward adult-like organization, coinciding with the period when rats first reliably show object location memory. If the young spatial mapping system is still noisy, the additional spatial information gained by climbing on and rearing at objects may help fine-tune it — essentially, the pups may be using vertical exploration to boost the signal in an immature hippocampal network. The authors are careful to frame this as an intriguing hypothesis that requires recording from spatially tuned cells during tasks with and without climbable objects. They also acknowledge a parallel possibility: visual circuits, including dorsal and ventral streams and binocularly guided behavior, are still maturing at this age, and sensory immaturity could also shape how young animals engage with objects and their affordances.

The practical recommendations that follow are refreshingly concrete. Alongside discrimination ratios, developmental studies should routinely report environmental exploration measures — rearing, climbing, sitting — across habituation, sample, and test phases, since these behaviors show ontogenetic shifts sensitive to object cues. Anxiety-associated measures such as thigmotaxis, freezing, darting, and general locomotion could likewise reveal how environmental sampling interacts with novelty across ages. The authors also endorse calls to question whether discrimination ratios truly index memory at all, or whether they should be reinterpreted as novelty preferences, given that the recognition memory nomenclature is often used without directly testing recognition. Standardizing exploration definitions, using consistently characterized stimulus objects, and scaling apparatus to the animal would further strengthen the field’s developmental benchmarks.

The deeper message is philosophical as much as methodological. Animals of different ages may reach the same discrimination ratio through fundamentally different strategies and neural mechanisms, which means developmental memory research may be overlooking its most informative data by fixating on novelty preference. The challenge ahead, as the authors put it, is not simply to determine whether young animals remember, but to identify how they remember and whether the behavioral routes to successful memory change across development. Recognition tasks already provide a powerful window into cognitive development; realizing their full potential will require treating behavior not as a single outcome number, but as a rich, multi-dimensional record of the developing brain at work.

Subject of Research: Developmental recognition memory and the role of object affordances in rodent exploratory behavior

Article Title: The memory we miss: exploratory behavior as a window into developmental recognition processes

Article References: Williamson, E. G., Cushman, J. D., & Robinson, P. A. (2026). The memory we miss: exploratory behavior as a window into developmental recognition processes. Frontiers in Behavioral Neuroscience, 20, Article 1919577. https://doi.org/10.3389/fnbeh.2026.1919577

Image Credits: AI Generated

DOI: 10.3389/fnbeh.2026.1919577

Keywords: recognition memory, novel object recognition, exploratory behavior, affordances, developmental neuroscience, rodent behavior, hippocampus, spatial memory, discrimination ratio, fear conditioning, behavioral neuroscience, pre-weanling rats