🧠 neuroscience

Tracing the epistemic arc: Distinct physiological signatures for curiosity, insight, understanding and liking when viewing visual art

This study employs a "title effect" paradigm combined with behavioral, EEG, ECG, and eye-tracking measures to delineate the distinct physiological signatures and temporal dynamics of curiosity, insight, understanding, and liking within the epistemic arc of motivated art appreciation.

Original authors: Welke, D., Vessel, E. A.

Published 2026-09-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Welke, D., Vessel, E. A.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

We are constantly bombarded with new sights and sounds, and our brains are hardwired to make sense of them. When we encounter something confusing or ambiguous, we often feel a spark of curiosity, a drive to fill in the missing pieces. If we eventually find the answer, that moment of clarity can bring a sudden rush of pleasure, a feeling often described as an "aha" moment. This sequence, moving from the desire to know, through the reduction of uncertainty, to the joy of understanding, forms a fundamental pathway for how we learn and appreciate the world. Scientists call this the "epistemic arc," a journey where the brain seeks information to resolve confusion and rewards itself when it succeeds. While this process is central to human cognition, it has been difficult to track in real time, especially when dealing with complex things like art, where meaning is not always obvious.

To explore this journey, researchers at the University of Leeds and the Max Planck Institute for Empirical Aesthetics designed an experiment using visual art. They wanted to see how adding a title to a painting changes the way a person feels about it, moving from curiosity to insight and finally to liking. They recruited over one hundred participants for two separate studies. In the first, purely behavioral study, people viewed paintings on a computer screen. Before seeing the title, they rated how curious they were to know what the work was called. Then, they were shown either the painting's actual title or a fake label that said "untitled," which offered no new information. Finally, they saw the painting again and rated how much they felt they had an "aha" moment, how much they liked the image, and how well they understood it. The researchers also included a set of black-and-white images that are notoriously hard to interpret until you see the original photograph they were made from, serving as a benchmark for strong moments of insight.

The results confirmed that titles do more than just label a piece of art; they actively reshape the viewer's experience. When participants were given the real title, their feeling of understanding the artwork increased significantly, and they reported much stronger "aha" moments compared to when they saw the "untitled" label. Interestingly, while the titles boosted understanding and insight, they did not consistently make people like the paintings more. The study also revealed that the relationship between curiosity and insight depends heavily on the type of information provided. When a title offered a clear explanation of what was in the painting, the link between being curious and having an insight was strong and direct. However, when the title was vague or offered no new information, this connection weakened. The researchers found that the feeling of insight was linearly linked to how much people liked the art, but its relationship with how much they felt they understood the piece was more complex, changing depending on whether the art was abstract or realistic.

In a second experiment, the researchers took these findings a step further by recording the participants' brain activity, heart rate, and eye movements while they viewed the art. They used advanced computer models to decode these physiological signals, looking for specific patterns that matched the participants' ratings of curiosity, insight, and liking. They discovered that the brain's electrical activity could reliably predict how much a person felt they understood the artwork, with this signal building up over time as the viewer processed the image and the title. The feeling of insight, or the "aha" moment, was most clearly visible in changes in pupil size, a physical reaction that occurs when the brain is engaged in mental effort. Surprisingly, the researchers found a very fast electrical response in the brain that predicted how much a person would like the art, occurring almost immediately after seeing the image, long before the title was even shown. This suggests that while our deep understanding and moments of insight take time to develop, our initial aesthetic preference can be formed almost instantly.

The study also looked at how these processes unfold over time. The data showed that the feeling of curiosity about a title is a persistent state; the brain activity associated with this curiosity remained detectable even after the title was revealed and the viewer was looking at the painting a second time. This indicates that the drive to understand does not vanish the moment information is provided but lingers as the brain integrates the new context. While the researchers hoped to find specific brain waves associated with insight, such as rapid bursts of activity, their most robust findings came from tracking the timing of these signals across different body systems. They found that the physiological signatures of understanding were the strongest and most consistent, appearing in brain waves, heart rate, and pupil size at different stages of the viewing process.

Ultimately, this research provides a detailed map of how we move from not knowing to knowing when we engage with art. It shows that the pleasure of understanding is not a single event but a process involving distinct stages of curiosity, insight, and appreciation. The study demonstrates that while a title can dramatically improve how much we feel we understand a piece of art and trigger a sense of discovery, it does not necessarily change our immediate emotional reaction to the image itself. By combining behavioral ratings with physiological measurements, the researchers were able to trace the "epistemic arc" in real time, revealing that the brain's journey toward understanding is a complex, multi-layered experience that involves both rapid initial reactions and slower, deeper processing. This work offers a new window into the mechanics of motivated learning, showing how our brains seek out information to resolve uncertainty and how that resolution shapes our enjoyment of the world around us.

Technical Summary: Tracing the Epistemic Arc in Visual Art Interaction

Problem Statement
The paper addresses the need to characterize the "epistemic arc"—a hypothesized motivational pathway for learning that proceeds from curiosity (a drive state) to insight (uncertainty reduction) to pleasure (reward). While previous research has linked these states, particularly using low-level perceptual tasks like Mooney images, there is a lack of empirical data tracing these processes in the context of complex, naturalistic stimuli such as visual art. Specifically, the study seeks to determine how semantic information (titles) influences the transition from curiosity to insight ("aha" moments), understanding, and aesthetic liking, and to identify the distinct (neuro)physiological signatures and temporal sequences of these states.

Methodology
The authors conducted two preregistered experiments utilizing a paradigm based on the "title effect," where additional semantic information alters the interpretation of an artwork.

Experiment 1 (Behavioral): An online study with $N=55$ participants. Participants viewed visual artworks (categorized as figurative, semi-abstract, or abstract) and rated their curiosity for the title. They were then presented with either the original title (providing semantic information) or a dummy "untitled" label. Following a second viewing of the image, participants rated their "aha" experience, aesthetic liking, and feeling of understanding. The stimulus set included 120 artworks with varying title types (descriptive, explanatory, evocative) and 20 Mooney images with their solution cues.
Experiment 2 (Physiological): A laboratory study with $N=49$ participants using a subset of stimuli selected from Experiment 1 based on strong title effects. This experiment employed a multimodal recording approach including EEG (64 channels), pupillometry (binocular gaze and pupil size), and ECG (heart rate). The same experimental paradigm was used, but data were analyzed using a time-resolved multivariate decoding approach (sliding window decoding) to determine when and if specific ratings (curiosity, aha, liking, understanding) or experimental conditions could be decoded from the physiological signals.

Key Contributions

Paradigm Extension: The study successfully extends the investigation of the epistemic arc from low-level perceptual tasks (Mooney images) to complex visual art, validating the "title effect" as a tool to induce insight in this domain.
Behavioral Mapping: It establishes the complex, non-linear relationships between curiosity, insight, understanding, and liking, demonstrating that these relationships are shaped by the type and amount of information provided by the stimulus (e.g., title style and art abstraction).
Physiological Signatures: The study provides distinct physiological correlates for these internal states. Notably, it identifies that "aha" experiences can be decoded from pupil size but not EEG, while "understanding" is robustly decodable from EEG and heart rate over extended timespans.
Temporal Decoding: Using time-resolved decoding, the paper outlines the temporal sequence of processing, revealing that aesthetic liking has a fast initial cortical signature, whereas understanding builds up over time.

Results

Behavioral Findings (Exp 1):
- Title Effect: Providing a title significantly increased ratings of "aha" (slope = 17.89) and felt understanding (slope = 11.06) compared to untitled conditions. However, the title effect on aesthetic liking was marginal and not statistically significant.
- Stimulus Comparison: Mooney images elicited higher "aha" ratings than most art categories, though semi-abstract art with explanatory titles approached Mooney levels. Figurative art received the highest aesthetic liking ratings.
- Correlations: Curiosity predicted "aha" ratings, but the strength of this linear relationship varied by title type (strongest for Mooney images and explanatory titles, weaker for descriptive/untitled conditions). "Aha" ratings showed a positive linear relationship with liking across all categories. The relationship between "aha" and understanding was non-linear (U-shaped) for figurative and semi-abstract art, but linear for abstract art.
Physiological Findings (Exp 2):
- Decodability: Experimental conditions (1st vs. 2nd presentation, art vs. Mooney, title vs. no-title) were robustly decodable from time-domain EEG, pupil size, and heart rate.
- Understanding: Felt understanding was the most robustly decodable rating, showing significant signals in time-domain EEG (both pre- and post-title) and heart rate (post-title).
- Insight (Aha): "Aha" ratings were significantly decodable from pupil size in early post-title trials but could not be reliably decoded from EEG.
- Liking: Aesthetic liking showed a brief, significant decoding peak in the time-domain EEG very early in post-title trials (approx. 0.46–0.68s after onset), suggesting a fast cortical response component.
- Curiosity: Curiosity ratings were decodable from late pre-title pupil signals and early post-title EEG, suggesting a sustained cortical state that persists through the presentation of the title cue.

Significance and Claims
The authors claim that their study provides a "new window into core behavioral and physiological properties of motivated information foraging." By integrating behavioral ratings with multimodal physiological data, they offer a more mechanistic understanding of the epistemic arc.

The paper modestly concludes that while the study outlines the contours of this arc and traces temporal sequences, merging these results into a single coherent temporal-sequence model remains difficult. They emphasize that the findings reveal complex, partly non-linear relations between the states, shaped by stimulus information content. The study does not claim to have fully resolved the causal mechanisms of the arc but suggests that "aha" acts as a valid indicator of information gain and that "understanding" may hold a primacy in neural processing, evidenced by its strong and sustained physiological signature. The authors note that while they did not confirm all a priori hypotheses regarding specific EEG frequency bands (gamma/alpha) for insight and liking, the time-resolved decoding approach successfully identified distinct physiological signatures for different epistemic states.

More like this