Summary: Researchers have long known that the brain uses “GPS-like” circuits to navigate physical spaces. New research suggests those same neural systems also create an internal map of our emotions. Using fMRI and computational modeling, investigators found that the hippocampus and the ventromedial prefrontal cortex (vmPFC) work together to represent emotions along two core axes—valence (how pleasant or unpleasant a feeling is) and arousal (the intensity of the bodily reaction). This map-like organization ranges from broad categories such as “good” or “bad” to highly specific emotional states, and differences in this internal map may relate to mental health outcomes like depression and anxiety.
By combining human neuroimaging with artificial neural networks, the team demonstrated a structured hierarchy of emotion concepts. The hippocampus appears to encode the emotion categories themselves, with different hippocampal subregions supporting coarse versus fine-grained distinctions. The vmPFC contributes by tracking relationships and transitions between those categories, predicting how one emotional state might move into another. These findings offer a new framework for thinking about emotional representation in the brain and suggest potential clinical avenues for disorders that feature less differentiated emotional experience.
Key Facts
- The Emotional GPS: A hippocampus–prefrontal circuit, previously known for memory and spatial mapping, also organizes emotion concepts.
- Granular vs. Global: The interior hippocampus supports broad, global categories (for example, “this is good”), while more posterior hippocampal regions encode finer-grained, nuanced feelings.
- Relational Tracking: The vmPFC encodes relationships and transitions between emotional nodes, helping predict movement across the mental emotion map.
- Clinical Link: Higher emotional granularity—being able to distinguish subtle feelings—is associated with better mental health. In contrast, people with depression often show a compressed emotional map with fewer distinctions.
- AI Validation: The team used the Tolman-Eichenbaum Machine (TEM), a neural-network model of relational memory, to simulate agents moving through an abstract emotion graph and reproduce patterns similar to those seen in human brain data.
Source: Emory University
Psychology has long described emotions along two principal dimensions: valence (pleasantness vs. unpleasantness) and arousal (physiological intensity such as heart rate or breathing). Conceptualizing valence as longitude and arousal as latitude produces a useful metaphor: a mental map in which nodes represent emotional concepts and distances reflect similarity or difference. What remained unclear until now was how the brain constructs and organizes such a map.
New research published in Nature Communications by scientists at Emory University provides evidence that hippocampal-prefrontal circuits—regions known to form cognitive maps for space and memory—also instantiate map-like representations of emotion. The study shows that hippocampal activity reflects a hierarchical structure of emotional concepts, while vmPFC activity better captures the relationships and placements of those concepts within a two-dimensional affective space.
Philip Kragel, senior author and professor of psychology at Emory, notes the clinical implications: people with depression and anxiety often represent emotions in a more compressed, less differentiated way, while those who maintain finer emotional distinctions tend to have better outcomes. Identifying the neural basis of that compression could inform interventions that restore or expand emotional differentiation.
The research combined several approaches: analysis of the Emo-FiLM dataset (fMRI recordings and emotion ratings collected while participants watched short film clips), predictive pattern-recognition models, and simulations using the TEM artificial neural network. This multimodal strategy allowed the team to link subjective emotion reports to specific patterns of brain activity and to test whether a computational model could reproduce those patterns.
Results from predictive models showed that self-reported emotional experiences could be decoded from hippocampal–prefrontal activity. The hippocampus contained stronger signals for emotion categories, with interior regions reflecting broad affective distinctions and posterior areas encoding finer subtleties. The vmPFC, meanwhile, carried information about how emotional states relate and transition across time, particularly for broader categories.
To test the mapping idea computationally, researchers constructed an abstract graph where nodes corresponded to emotion categories derived from the film ratings. TEM agents learned that graph and generated trajectories showing how concepts relate. The model’s behavior mirrored the neural data: hierarchical, map-like organization in hippocampal representations and relational tracking consistent with vmPFC responses.
Foundational implications
This work offers a neurocomputational explanation for how the brain organizes abstract emotion knowledge at multiple levels of abstraction. The authors plan to extend the approach to study variations in emotional mapping across clinical populations, cultures, and development, asking whether broad categories are innate or learned and how more nuanced nodes are added over time.
Key Questions Answered:
A: Not a literal geographic map but a coordinate-like system. The brain arranges feelings along dimensions such as pleasantness and intensity. Emotions that are similar in valence and arousal—like anger and fear—sit near each other, while emotions with different combinations of valence and arousal—like happiness and excitement—occupy another region of the space.
A: Granularity lets you distinguish subtle differences in how you feel, which supports better emotion regulation and problem-solving. A compressed map, often observed in depression, can make many distinct states look the same—simply “bad”—which reduces the ability to respond effectively to specific emotional needs.
A: Research suggests that practices like affect labeling—naming and describing emotions with precision—may increase emotional granularity. Repeatedly attending to and labeling subtle differences in feeling could help expand the nodes on your internal emotion map and improve regulation.
Editorial Notes:
- This article was edited by a neuroscience news editor.
- The journal paper was reviewed in full for accuracy.
- Additional context was added by staff to clarify methods and implications.
About this emotion and neuroscience research news
Author: Carol Clark
Source: Emory University
Contact: Carol Clark – Emory University
Image: The image is credited to Neuroscience News
Original Research: Map-like representations of emotion knowledge in hippocampal-prefrontal systems by Yumeng Ma & Philip A. Kragel. Nature Communications. DOI: 10.1038/s41467-025-68240-z. Open access.
Abstract
Map-like representations of emotion knowledge in hippocampal-prefrontal systems
Emotional experiences involve more than momentary feelings and bodily reactions; they depend on knowledge that extends across contexts and time. Although emotions are commonly conceptualized within a low-dimensional space defined by valence and arousal, the neural mechanisms that support this organization have been unclear. This study tests whether hippocampal-prefrontal circuits—regions implicated in cognitive mapping—also support structured abstraction of emotional experience. Using fMRI data recorded while participants viewed emotionally evocative film clips, the researchers found that hippocampal activity represents emotion concepts in a hierarchical structure, while ventromedial prefrontal cortex more precisely tracks positions within a two-dimensional affective space. Computational models further showed that these neural responses can be predicted from the statistical regularities of emotional transitions across multiple time scales. Together, the results indicate that hippocampal-prefrontal systems represent emotion knowledge in a map-like way across levels of abstraction, offering insight into how the brain organizes affective information.