What Is the IQ of a Mouse? Unraveling Their Cognitive Potential

Mice do not have an IQ in any meaningful sense. IQ is a human psychometric construct, calibrated to human populations, and there is no valid way to convert a mouse’s performance on a learning task into a number on the human scale. What mice do have, and what researchers have been studying for decades, is a measurable general cognitive ability that functions a lot like the “g factor” psychologists use to describe broad intelligence in people. This factor, extracted from batteries of diverse cognitive tests, accounts for roughly a third of the variation in how well individual mice perform across tasks. The science of mouse cognition turns out to be far richer than a single number could capture.

Why You Cannot Assign an IQ to a Mouse

Human IQ tests measure a person’s cognitive performance relative to other humans of the same age, then express that rank as a standardized score with 100 as the average. The tests are designed around human sensory abilities, human language, and human cultural contexts. A mouse cannot take a vocabulary subtest or rotate a mental image of a three-dimensional block. More fundamentally, cognitive tests developed for non-human animals have rarely been evaluated for the kind of measurement validity that human intelligence testing demands. Non-cognitive factors like anxiety, motivation, motor ability, and sensory acuity all influence how an animal performs, and teasing those apart from genuine cognitive skill is an ongoing challenge.¹

So when someone asks “what is the IQ of a mouse,” the honest answer is that the question doesn’t apply. But a more interesting version of the question does: can mice vary in overall smartness the way people do? And if so, how much?

The Mouse g Factor

In human psychology, the g factor is the statistical observation that people who score well on one kind of cognitive test tend to score well on others. It is the closest thing we have to a single number capturing general mental horsepower. Researchers have now found something strikingly similar in mice. When genetically diverse mice are run through batteries of different learning tasks, a common factor emerges from the data. In one study using heterogeneous stock mice, all measures of cognitive ability loaded positively onto a single factor that accounted for about 31% of the total variance across tasks.² That factor held up after removing outliers and correcting for sex differences, and it was unrelated to anxiety, suggesting it reflects genuine cognitive ability rather than just temperament.

Other research has replicated and extended this finding. In genetically heterogeneous CD-1 mice, a general cognitive factor accounted for about 38% of the variance across tasks, alongside a separate domain-specific factor tied to spatial and hippocampal processing.³ A sibling study found the g factor was heritable, with an upper-limit heritability estimate of around 40%.⁴ In other words, smart parents tend to produce smart pups, just as in humans. The parallel is not perfect, but it is real enough to make researchers take mouse general intelligence seriously as a biological phenomenon.

How Researchers Actually Test Mouse Brains

Since you cannot hand a mouse a pencil-and-paper exam, researchers rely on behavioral tasks that probe specific cognitive abilities. The most iconic is the Morris water maze, a circular pool of opaque water with a hidden platform just below the surface. A mouse is placed in the pool and must use visual landmarks around the room to navigate to the platform and escape the water. Over repeated trials, a cognitively intact mouse will find the platform faster and take more direct routes, demonstrating spatial learning. When the platform is removed, researchers measure how much time the mouse spends searching in the area where it used to be, revealing how well it remembers.⁵ The Morris water maze has been a workhorse of neuroscience for decades and remains one of the most widely used methods for assessing learning and memory in mice.⁶

More recently, touchscreen-based systems have brought mouse cognitive testing closer to what you might see in a human neuropsychology lab. These platforms present visual stimuli on a screen, and the mouse responds by nose-poking the correct image to earn a food reward. From this basic setup, researchers can run a surprising range of tests: extinction tasks that measure how quickly a mouse stops doing something that no longer pays off, reversal learning tasks that test mental flexibility, and the five-choice serial reaction time task, which gauges attention and impulse control.⁷ These touchscreen tasks allow researchers to measure constructs like impulsivity and compulsivity with a precision that would have been hard to imagine a generation ago.

What Happens Inside a Thinking Mouse’s Brain

Mouse brains are tiny compared to ours, but they share the same basic architecture in the regions most critical for learning and memory. The hippocampus is central. When researchers damage it in mice, the animals lose the ability to integrate the “what,” “where,” and “when” of an experience, essentially the building blocks of episodic memory. Mice with hippocampal damage couldn’t remember which objects they had explored, where they encountered them, or when.⁸

The medial prefrontal cortex plays a more selective role. Mice with prefrontal damage could still remember what they had seen and when, but they lost track of where events had occurred. This division of labor between brain regions mirrors what we see in humans, where the hippocampus binds experiences into coherent memories while the prefrontal cortex handles source monitoring and strategy. Research using an ethological search task found that the mouse prefrontal cortex tracks the animal’s default strategy and adjusts as the animal learns to follow a more reliable cue, reflecting a competition between habitual and flexible behavior.⁹

Not All Mice Are Equally Smart

Lab mice are not a monolith. Different inbred strains, bred for genetic uniformity, show strikingly different cognitive profiles. C57BL/6 and BALB/c mice are two of the most commonly used strains in behavioral research, and they differ in tasks assessing memory, anxiety, and learning. These differences stem from their distinct genetic backgrounds, and they matter for researchers designing experiments because a result obtained in one strain may not replicate in another.¹⁰

Genetics can also be tweaked directly. In a well-known line of experiments, researchers engineered mice to overexpress a component of a brain receptor involved in learning (the NR2B subunit of the NMDA receptor). These transgenic mice showed greater electrical activity at synapses and improved performance across a variety of learning and memory tasks.¹¹ The popular press dubbed them “Doogie mice” after the fictional TV prodigy, and while the nickname was playful, the finding was real: a single genetic change could broadly enhance cognitive function. This underscored that mouse intelligence has identifiable biological levers that can be pushed.

Environment Shapes the Mouse Mind

Genes set a range, but experience fills it in. Even among mice with identical genetic backgrounds raised in the same cage, individual differences in behavior emerge and persist into adulthood. Research tracking genetically identical mice in enriched environments found that individuals developed distinct behavioral trajectories over time, and these were tied to lasting differences in hippocampal neurogenesis and epigenetic marks on neuronal plasticity genes.¹² In plainer terms: two genetically identical mice raised side by side can end up with measurably different brains and behaviors, driven by subtle differences in their early experiences. This is a powerful reminder that “mouse intelligence” is not fixed at conception.

Enriched environments, those with tunnels, running wheels, novel objects, and social opportunities, have been shown repeatedly to boost cognitive performance. The Doogie mouse experiments mentioned earlier also examined environmental enrichment and found that it, too, improved learning and memory, working through partly overlapping biological pathways as the genetic manipulation.¹¹

Mice as Problem Solvers

One of the more impressive demonstrations of mouse cognition comes from mechanical puzzle-solving. Researchers presented mice with a lockbox, a multi-step mechanical puzzle requiring several distinct manipulations to open. The mice engaged spontaneously with the task and learned to solve it within just a few trials. Detailed tracking of their behavior revealed that early progress was driven by improved motor skills, like learning how to physically manipulate the mechanisms, while a genuine cognitive strategy for the correct sequence of steps emerged only later.¹³ This separation of motor learning from cognitive strategy is something researchers also observe in humans tackling novel physical problems.

Evolution appears to have shaped mouse problem-solving ability in response to environmental demands. A study comparing different subspecies of house mice found that populations with the longest history of living alongside humans, up to 11,000 years, outperformed those with shorter commensal histories across seven different food-extraction puzzles. The differences were not explained by exploration, motivation, persistence, or impulse control, suggesting that the cognitive ability itself had evolved.¹⁴ Living in human environments, with their novel obstacles and food sources, seems to have selected for smarter mice over millennia.

How Mice Compare to Rats

Rats are the other major rodent in the lab, and they are often assumed to be the smarter of the two. There is some truth to this, but it depends on the task. In an adaptive decision-making study, rats learned the basic discrimination task significantly faster than mice, reaching the performance criterion in about 10 sessions compared to 18 for mice. The largest gap appeared in sound discrimination, where rats were about seven sessions ahead.¹⁵ But once both species had learned the task, they adapted to shifting reward rules equally well, and mice actually showed slightly steeper psychometric slopes, meaning they were making sharper perceptual distinctions between stimuli. Rats may be faster learners in many contexts, but mice can match them in flexibility once they have mastered the basics.

Rats have also demonstrated something close to metacognition, the ability to monitor the strength of their own memories. In an odor-matching task, rats given a “decline” option performed significantly better on the tests they chose to take than on tests they were forced to take, suggesting they could distinguish between remembering and forgetting.¹⁶ Whether mice share this metacognitive ability is still an open question, though their demonstrated episodic-like memory capacities suggest it is at least plausible.

Social and Episodic Memory

Mice are not just spatial learners. They also form memories that integrate social information with context. In a recent study, mice explored familiar companions placed in either matching or mismatched contexts from prior encounters. They spent more time investigating a companion who appeared in the wrong context compared to one who appeared in the expected setting, suggesting they had formed a memory binding the identity of the other mouse to the place and context in which they originally met.⁹ This kind of “what-where-when” memory in a social context is exactly the type of integrative recall that was once considered uniquely human.

When Mouse Brains Age

Like humans, mice experience cognitive decline as they get older. In the widely used C57BL/6J strain, signs of age-related memory loss in the Morris water maze and novel object recognition tests appeared from about six months of age onward. Discrimination and recognition scores declined steadily, providing a baseline for what “normal” aging looks like in mice as opposed to pathological neurodegeneration.¹⁷ Six months in a mouse corresponds very roughly to early middle age in a human, so the timeline of decline is compressed but the trajectory is familiar.

This age-related cognitive trajectory is one reason mice are so heavily used in Alzheimer’s disease research. But it also creates a complication: distinguishing normal cognitive aging from disease-model effects requires solid baseline data, and that data varies by strain, housing conditions, and testing protocol.

The Gut-Brain Connection

One of the more surprising recent findings in mouse cognition research involves the gut microbiome. The community of bacteria living in a mouse’s intestines communicates with its brain through immune, metabolic, and neural pathways. Disruptions to this microbial community can trigger inflammation that reaches the hippocampus and impairs learning and memory.¹⁸

This connection has been tested directly. In an Alzheimer’s disease mouse model, transplanting gut bacteria from a human donor carrying a genetic protective factor against Alzheimer’s improved the mice’s short-term recognition memory compared to mice receiving a transplant from a donor without that protection. The improvement was associated with changes in specific bacterial genera and increased immune activity in the hippocampus.¹⁹ Meanwhile, research in another Alzheimer’s model found that depleting gut bacteria with antibiotics confirmed that information flows bidirectionally between gut and brain. Certain short-chain fatty acid-producing bacteria were sharply reduced in the Alzheimer’s mice, and a metabolite called iso-butyric acid was elevated in their feces, blood, and brain, correlating with inflammation and amyloid buildup.²⁰ The takeaway is that a mouse’s cognitive ability is not solely a product of its neurons; its gut bacteria are part of the equation.

Why Testing Conditions Matter More Than You’d Think

Here is a wrinkle that rarely makes it into popular summaries of mouse intelligence: mice are nocturnal. Their brains are wired for peak performance in darkness. Yet for over a century, most behavioral testing has been done during the daytime, under bright laboratory lighting, simply because that is when human researchers are at work. Only recently have some labs begun testing rodents under red light (which mice cannot see) during their natural active period. The time of day at which rodents are tested is still often unreported or inappropriate to the species’ known active periods, even in highly cited studies.²¹ Imagine taking your most important exam at 3 a.m. after being woken from deep sleep under fluorescent lights. That’s what many cognitive tests ask of mice. This means we may be systematically underestimating mouse cognitive performance.

From Mouse Maze to Human Medicine

The entire enterprise of testing mouse cognition exists partly because researchers hope findings will translate to human brain diseases. The track record here is sobering. In Alzheimer’s disease, more than 99% of clinical trials have failed despite positive results in animal models. A systematic evaluation found that animal efficacy results correlated with human clinical outcomes only about 58% of the time. Individual mouse models gave divergent results across different drug interventions, and individual drugs gave divergent results across different models.²² The fundamental problem is that no single mouse model captures the full complexity of a human neurological disease. Each model mimics certain symptoms but misses others.

This does not mean mouse cognitive research is useless. It means the “IQ” question has real stakes beyond curiosity: understanding how mouse cognition works, where it parallels human cognition, and where it diverges is essential for knowing when a promising drug result in a mouse is likely to hold up in people. The mice that look smart in the lab may be solving problems in ways that only partly overlap with how a human brain would handle the same challenge, and that gap is where translational failures live.