Skip to main content

Transforming the understanding
and treatment of mental illnesses.

Celebrating 75 Years! Learn More >>

Photo of Joshua A. Gordon, M.D., Ph.D.

What Can Animals Tell Us About Mental Illnesses?


This Director’s Message on the use of animal models in mental health research is one of two focused on this topic. The companion message, directed toward scientific researchers, is titled: A Hypothesis-Based Approach: The Use of Animals in Mental Health Research.

“How can you tell if a mouse has schizophrenia?”

I get this question a lot. You see, in my research lab, I study the brains of mice engineered to carry a genetic mutation that increases the risk for schizophrenia in humans. So, when I meet new people, the first question I get is the usual, “so, what do you do?” And the second question... well, now you know.

I used to respond with a smart retort, like, “ask a mouse psychiatrist.” But it turns out it is a great question, for two reasons. First, it is an opportunity to talk about my science, which I love to do (almost as much as I love doing the science in the first place!); second, the answer is simultaneously obvious and nuanced, and important for our field to get right.

The obvious answer

Mice do not have schizophrenia.

Or any other mental illness. The human brain is a complex organ, dizzyingly complex, and quite different from the mouse brain in many ways. These differences definitively preclude us from re-creating a mental illness in its totality in mice. One small fact: people with schizophrenia have changes in the function of the dorsolateral prefrontal cortex, a piece of the brain that sits roughly above the temples. Mice don’t have a well-defined dorsolateral prefrontal cortex, above their temples or anywhere else!

Of course, there are a lot of other differences between mice and people besides just the presence of a dorsolateral prefrontal cortex — different genes, molecules, cells, circuits, and behaviors. Indeed, the differences in behaviors are particularly challenging — and not just because mice can’t tell us if they hear voices! Variations in how human patients express emotion, perform cognitive tasks, seek out rewards, avoid punishments, and interact with other people (including their doctors and therapists) are how we recognize and categorize mental illnesses. The profound difference in these behaviors between mice and people prevents the straightforward mapping of psychiatric syndromes across species.

Take emotional expression, for example. In people, facial expressions reflect their inner emotional experience — smiles for happiness, frowns for sadness, furrowed brows for anger, etc. Not so in mice. But we can recognize various states that seem to mirror these human emotions by observing other behaviors in mice — their movement patterns, the tone of their vocalizations, the repetitiveness of their grooming, and their bowel movements and eating habits. Accordingly, scientists have used these behaviors to try to “model” human disorders in mice. For example, scientists have studied overgrooming as a potential model of obsessive-compulsive disorder and avoidance of food in a brightly lit open arena as a model of anxiety, etc.

None of these models can accurately reflect the human condition. For example, anxiety in humans is much more than avoidance of food in a brightly lit area — anxiety can be accompanied by changes in energy, sleep, and other behaviors. Similarly, individuals with mental illnesses such as schizophrenia typically suffer from a range of symptoms, including episodic hallucinations and delusions, but also more enduring social and cognitive dysfunction. No mouse model of schizophrenia can accurately reproduce all these features, nor should we expect one to do so. Mice are not people. Mice do not have schizophrenia or anxiety.

The nuance

Nonetheless, mice (and other species) can help us understand and develop treatments for schizophrenia, anxiety, and other mental illnesses. How? By teaching us about molecules, cells, circuits, and systems of the brain; how they work; and how they produce behavior. We can then use that information to identify better targets for treating impairments or symptoms associated with mental illnesses.

One example: in my research, I don’t study a mouse model of schizophrenia (remember, there is no such thing). Instead, I study how certain genetic mutations increase the risk of schizophrenia and how we might use that understanding to develop new treatments. So, when I was starting my career, I teamed up with colleagues at Columbia University (Drs. Joseph Gogos, M.D., Ph.D. , and Maria Karayiorgou, M.D. , among others) to try to study the effect of one of these mutations, the 22q11.2 microdeletion—a deletion of more than 20 genes that results in a neurodevelopmental syndrome. About 30% of people with this microdeletion will develop schizophrenia, making it one of the strongest genetic risk factors for the disease.

Several groups, including those led by Drs. Gogos and Karayiorgou, engineered the genome of mice to create a similar deletion. But they didn’t create mice with schizophrenia. They created mice that could be used to study the effects of this microdeletion on brain development and function. Before I came along, they had already demonstrated significant effects of the microdeletion on molecular function, neural development, and working memory. My lab worked with theirs to demonstrate that these effects could all be linked together by changes in the function of a particular circuit involving the hippocampus and prefrontal cortex, two brain regions often implicated in schizophrenia. We also showed that targeting one particular molecular mechanism — a protein called a kinase that is important for neurons as they develop — reversed the effects on neurodevelopment, thereby rescuing the circuit and behavioral changes. We continue to work together to try to figure out whether this molecular mechanism might be a good therapeutic target to help people with the microdeletion and/or others with schizophrenia.

This is the approach we use not only in my lab but throughout the NIMH portfolio: carefully choosing animal models to explore the mechanisms underlying biological processes relevant for mental illnesses and then using this mechanistic knowledge to begin the challenging process of developing novel interventions. As I’ve written about recently, this approach can work. But it requires careful experimental design to ensure that the questions being asked can be properly answered in an animal.

This last point is really important to consider. We have an obligation to carefully consider the need for animal studies, for both ethical and scientific reasons. Ethically, we must remember that animals are deserving of being treated with respect and care for their well-being. Scientifically, given the difficulty of translating across species, if there is knowledge that can be gained safely and ethically from healthy humans and individuals with mental illnesses, we should study them instead of animals. But to understand the full complexity of the brain--particularly how the circuits of the brain drive behavior, and the role of molecular and cellular processes in the development and function of those circuits — animals will continue to play a crucial role. We must, therefore, ensure that they are used appropriately to answer those questions that would be impossible or unsafe to answer using human subjects.