Summary: A precision developmental neurobiology and behavioral study from the University of Oregon identifies a distinct group of neurons that act like a biological “disappointment meter.” These cells, located deep in the lateral habenula, activate when an animal expects a reward but receives less than expected or none at all. The intensity of their firing scales with the size of the shortfall, revealing a concrete neural mechanism for encoding negative reward prediction errors and guiding learning and decision-making.
The research shows how the brain records expectation-outcome mismatches and uses that information to update behavior, refine strategies, and support adaptive learning. Mapping these specialized neurons opens avenues for more targeted studies of psychiatric conditions where reward processing is disrupted, such as depression and addiction.
Key Facts
- The Anti-Reward Center: The focus is the lateral habenula (LHb), an evolutionarily ancient brain structure known for increased activity during unexpected negative events, leading to its reputation as the brain’s “anti-reward center.”
- The Disappointment Gradient: In mice trained to seek sugar water, researchers found a subset of LHb neurons that become active when an anticipated reward is reduced or withheld. The neural response strength scaled directly with how much reward was missing.
- Not a General “Bad News” Detector: These cells did not respond strongly to unrelated unpleasant surprises, like a sudden puff of air, showing they are specialized for signaling shortfalls in expected reward rather than registering all aversive events.
- Accidental Discovery: Senior author Emily Sylwestrak first noticed these signals while recording stray activity from neighboring tissue during an experiment aimed at a different region, which led to targeted investigation of this cell type.
- Error Correction Machine: Lead author Kana Suzuki emphasizes that this circuit specificity lets the brain distinguish different negative experiences and use the record of successes and failures to shape future choices, perseverance, and strategy changes.
- Therapeutic Potential: Identifying a molecularly defined cell population offers a precise “knob” for future research. That specificity could guide development of treatments that target problematic circuits in neuropsychiatric disorders while avoiding broad off-target effects of conventional drugs.
- Next Steps: The research team plans to move from observational recordings to active manipulation—silencing or altering these neurons during reward tasks—to determine how their dysfunction contributes to conditions such as addiction and depression.
Source: University of Oregon
University of Oregon neuroscientists have identified a group of brain cells that function as a precise biological “disappointment meter,” signaling when outcomes fall short of expectations.
Published May 8 in Current Biology, the study describes a specific population of lateral habenula neurons that become active when mice expect a reward but receive less than anticipated or none at all. The neurons’ activity provides a quantitative readout of the gap between expected and actual reward, which researchers could use to infer the magnitude of the missing reward.
Assistant Professor Emily Sylwestrak explains that understanding which cell types encode particular aspects of reward processing is essential for designing targeted interventions for neuropsychiatric disorders. “If a particular cell type is compromised in conditions like depression, identifying it gives researchers a focused target to correct circuit dysfunction without affecting unrelated neurons,” she said.
The lateral habenula contains many neuron types, and prior work implicated the region broadly in responding to negative outcomes. This study drills down to a transcriptionally defined subpopulation—neurons expressing tachykinin 1 (Tac1)—and shows they are selectively tuned to negative reward prediction errors (nRPEs). Tac1-expressing LHb neurons increase firing when rewards are worse than expected and scale their response with the size of the deficit, while showing little modulation to other task events or robust responses to aversive stimuli.
The researchers trained mice to perform a reward-seeking task where nose-pokes could yield sugar water. Once the animals learned to expect a reward, the experimenters sometimes reduced the reward amount or withheld it entirely. The identified LHbTac1 neurons reliably increased activity at those moments, and the amplitude of that activity corresponded to how much reward was missing—providing a graded signal of disappointment.
This specificity matters for learning and behavior. As doctoral student Kana Suzuki notes, treating all negative experiences the same would hamper the brain’s ability to adapt: different negative outcomes call for distinct behavioral responses. A circuit that specifically encodes expectation shortfalls enables the animal to update predictions, adjust strategies, and persist or change course as needed.
Understanding the cell-type-specific logic of negative reward signaling can clarify how normal decision-making breaks down in neuropsychiatric disorders. The team intends to inhibit or otherwise manipulate Tac1 LHb neurons in future experiments to map causal roles in reward-guided behavior and to explore how their malfunction might underlie aspects of addiction, depression, and maladaptive persistence or avoidance.
Existing psychiatric medications often affect broad swaths of the brain, creating unwanted side effects. A more precise molecular and circuit-level understanding of neurons that encode negative reward prediction errors could enable the development of targeted therapies that correct specific dysfunctional pathways while minimizing collateral impacts.
Key Questions Answered:
A: These neurons adjust their electrical firing rate in proportion to the size of the missing reward. In controlled experiments, the neural response reliably matched the shortfall in sugar water, allowing researchers to infer the amount withheld from the activity alone.
A: Their selectivity shows the brain uses separate channels to represent different negative situations. That separation enables accurate learning from specific types of mistakes—reward prediction errors—without confusing them with other aversive experiences that require different responses.
A: By identifying a transcriptomically defined population (Tac1-expressing LHb neurons) linked to negative reward prediction errors, researchers gain a precise molecular target. Therapies that modulate this defined group may correct maladaptive reward processing with fewer off-target effects than broadly acting drugs.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full by the editorial team.
- Additional explanatory context was provided by staff.
About this neuroscience and disappointment research news
Author: Molly Blancett
Source: University of Oregon
Contact: Molly Blancett – University of Oregon
Image: Image credited to Neuroscience News
Original Research: Closed access. “Tachykinin 1 neurons in the lateral habenula signal negative reward prediction error” by Kana E. Suzuki, Tharusha A. Seagoe, Blake Holcomb, Jacqulyn R. Kuyat, and Emily L. Sylwestrak. Current Biology. DOI: 10.1016/j.cub.2026.04.032
Abstract
Tachykinin 1 neurons in the lateral habenula signal negative reward prediction error
Accurately evaluating outcomes and predicting which actions produce reward is essential for adaptive behavior. Discrepancies between expected and actual outcomes—reward prediction errors (RPEs)—serve as teaching signals that update future predictions. Neural signatures of RPEs appear across brain regions, including the lateral habenula (LHb), where some neurons encode negative RPEs (nRPEs) by increasing activity for worse-than-expected outcomes and decreasing activity for better-than-expected outcomes.
The LHb influences dopaminergic activity and plays a central role in reward learning and decision-making, but its many cell types have mixed functions. Using cell-type-specific recordings in mice performing reward-guided tasks, the study demonstrates that LHb neurons expressing tachykinin 1 (Tac1) are selectively tuned to nRPEs. LHbTac1 activity is sensitive to both expected and realized reward value and scales with their difference. These neurons show minimal responses to other task events and are weakly driven by aversive stimuli.
Overall, the data indicate that Tac1 identifies a subpopulation of LHb neurons that encodes valence-biased prediction errors—preferentially signaling when appetitive outcomes are worse than expected—and create a foundation for targeted manipulations to probe their role in reward-guided behavior.