New research reveals how estrogen amplifies the brain’s dopamine signals

A new study on laboratory rats has found that the brain’s ability to learn from rewards is enhanced by the hormone estrogen. The work shows that naturally cycling estrogen levels alter the brain’s dopamine system, making the chemical signals that guide learning more potent. The research was published in the journal Nature Neuroscience.

Hormones are known to produce wide-ranging effects on the brain, influencing everything from mood to decision-making. The precise mechanisms behind these changes, especially for complex cognitive processes, are still being mapped out. A team of scientists led by researchers at New York University sought to understand how natural hormonal fluctuations could affect the fundamental process of learning.

“Despite the broad influence of hormones throughout the brain, little is known about how these hormones influence cognitive behaviors and related neurological activity,” says Christine Constantinople, a professor in New York University’s Center for Neural Science and the paper’s senior author. She notes a growing recognition that “changes in estrogen levels are related to cognitive function and, specifically, psychiatric disorders.” The research team included scientists from NYU’s Center for Neural Science, NYU Grossman School of Medicine’s Neuroscience Institute, and Virginia Commonwealth University.

The researchers focused on a concept called reinforcement learning. This is a form of learning from trial and error, where an animal or person adjusts their behavior based on whether outcomes are better or worse than expected. This process is thought to be driven by the brain chemical dopamine, which acts as a “reward prediction error” signal, broadcasting the difference between an anticipated reward and the actual one. This signal helps the brain update its expectations for the future.

To investigate this, the scientists trained hundreds of female rats on a task designed to measure their expectations and learning. In each trial, a rat initiated the process by poking its nose into a port. It then heard one of five distinct audio tones, each signaling a different volume of a water reward. The reward was delivered after an unpredictable delay, and the rat could choose to wait or opt out to start a new trial.

Free daily newsletter

The experiment was structured in blocks of trials. In “low” blocks, only small rewards were offered, while in “high” blocks, only large rewards were available. This design allowed the researchers to manipulate the rats’ expectations. The key measurement was the time it took for a rat to initiate the next trial. A shorter initiation time suggested the rat was more motivated and expected a better outcome.

The team tracked the rats’ four-stage reproductive cycle, which involves predictable fluctuations in estrogen levels. They found that the rats’ learning was significantly affected by their hormonal state. When in the “proestrus” stage, where estrogen levels peak, the rats were more sensitive to changes in reward availability. Their trial initiation times changed more dramatically between the low-reward and high-reward blocks, suggesting they were updating their expectations more strongly based on recent outcomes.

To understand the neural basis for this behavioral change, the researchers used a technique called fiber photometry. They injected a genetically engineered sensor into a brain region called the nucleus accumbens, a key reward center. This sensor becomes fluorescent in the presence of dopamine, allowing the team to measure dopamine activity in real time. The recordings confirmed that dopamine signals in this region functioned as reward prediction errors. Dopamine levels surged when rewards were larger than expected and dipped when they were smaller.

In alignment with the behavioral findings, these dopamine signals were amplified during the high-estrogen proestrus stage. The difference in the dopamine response between large and small rewards was greater, meaning the signal had a wider dynamic range. Specifically, the dopamine spikes associated with large, positive prediction errors were enhanced.

The scientists then tested if this dopamine signal directly caused the changes in behavior. Using optogenetics, a technique that uses light to activate specific neurons, they stimulated dopamine-releasing terminals in the nucleus accumbens at the moment a reward was signaled. This artificial boost of dopamine caused the rats to initiate subsequent trials more quickly, confirming that the dopamine signal directly influences this learning-related behavior.

The next step was to identify the molecular mechanism connecting estrogen to the amplified dopamine signals. The team analyzed protein levels in the nucleus accumbens of rats at different stages of their cycle. They found that in the high-estrogen stages, there were significantly lower levels of two key proteins: the dopamine transporter (DAT) and the serotonin transporter (SERT). These transporter proteins act like vacuums, removing dopamine from the spaces between neurons after it has been released.

With fewer of these transporters present, dopamine would linger longer after being released. The researchers used a computational model to simulate this effect. The model showed that slower dopamine removal would allow the effects of consecutive neural signals to accumulate, resulting in larger and more sustained dopamine peaks. This process would especially amplify the signals for the largest rewards, matching what the researchers observed in their dopamine recordings.

To establish a causal link between estrogen signaling and learning, the team performed one final experiment. They used a virus to deliver a short-hairpin RNA, a tool for genetic suppression, into the ventral tegmental area, the midbrain region where the dopamine-producing neurons that project to the nucleus accumbens originate. This tool specifically knocked down the expression of a primary estrogen receptor, effectively blocking estrogen’s ability to act on these cells.

After this intervention, the rats’ learning was impaired. Their behavior became less sensitive to the changes between high and low reward blocks, an effect similar in magnitude to the difference between the high-estrogen and low-estrogen states. This result demonstrated that estrogen signaling within this specific midbrain circuit is a direct cause of the enhanced reinforcement learning.

The study has some limitations. The experiments were conducted in rats, so direct translation to human cognition requires more research. The work also focused on dopamine dynamics in the nucleus accumbens, leaving open questions about how estrogen might also affect the activity of dopamine neurons at their source.

Future research could explore the precise molecular pathways through which estrogen receptors regulate the expression of transporter proteins. A better understanding of this hormone-neuromodulator interaction may provide insight into why the severity of symptoms for some neuropsychiatric disorders can fluctuate with hormonal cycles.

“Our results provide a potential biological explanation that bridges dopamine’s function with learning in ways that better inform our understanding of both health and disease,” said Carla Golden, an NYU postdoctoral fellow and the paper’s lead author.

The study, “Estrogen modulates reward prediction errors and reinforcement learning,” was authored by Carla E. M. Golden, Audrey C. Martin, Daljit Kaur, Andrew Mah, Diana H. Levy, Takashi Yamaguchi, Amy W. Lasek, Dayu Lin, Chiye Aoki & Christine M. Constantinople.