Curiosity · Well-being · Learning Science

The Pleasure of Getting Better: Is Learning a Basic Need?

By Juliette Deyts & Matisse Poupard

In our research team, we have a lunchtime ritual: board games and video games. One of our current favorites is ROUNDS, a fast-paced, chaotic shooter we'd happily recommend to anyone looking to rekindle a healthy sense of competition with their colleagues. What makes these sessions so memorable is the intensity of the emotions involved.

There is something almost disproportionate about the joy of winning a casual game. No prize is on the line. No career depends on the outcome. And yet, beating a more experienced player, especially when you only just learned the rules, can feel genuinely exhilarating. Conversely, a string of defeats can quietly cast a shadow over the rest of your afternoon. For most of us, unless you're a professional e-sports player, nothing about that game will change your life. And still, it matters, at least in the moment.

Why? There are surely many reasons, such as the social bonding, the thrill of competition, the simple pleasure of play. But from the perspective of our team's research, one reason stands out: the feeling of learning. The sensation that you're getting better, that you’re understanding, that the gap between you and the more experienced players is slowly closing. This feeling is a powerful source of satisfaction in its own right.

The Learning Progress Hypothesis: why getting better feels good

But why, exactly, does the feeling of learning feel so rewarding? This is precisely what the Learning Progress Hypothesis (LPH) sets out to explain.

According to LPH (see Figure 1), curiosity-driven behaviors are fundamentally guided by intrinsic rewards associated with the perception of learning progress (Oudeyer et al., 2016). In more classical learning contexts, this manifests as a desire to fill a knowledge gap, motivated by the expectation of an intrinsic reward following the acquisition of new knowledge. In other words, what motivates us is not simply the resolution of uncertainty, but the feeling of improvement itself.

Figure 1: Schematic of the Learning Progress Hypothesis. Identifying a knowledge gap triggers information-seeking behavior, which leads to the acquisition of new knowledge. This process generates a perception of learning progress, which in turn delivers an intrinsic reward, reinforcing further curiosity-driven exploration.

Neuroscience offers some compelling supporting evidence. Research has shown that resolving uncertainty through acquiring new knowledge activates dopamine-rich regions of the brain (such as the striatum) which is associated with reward processing and reinforcement learning (Gruber et al., 2014; Poli et al., 2024). Recent studies in mice also suggest that midbrain dopamine release is tied specifically to the rate of learning, rather than to the mere acquisition of information (Coddington et al., 2023). This points to a profound biological basis for curiosity and intrinsic motivation: our brains may be wired not just to seek answers, but to seek progress.

|

And here's the key insight: the reward doesn’t come from learning itself, but from perceiving that you are learning. This is why direct feedback matters so much. When you survive to the end of a round in ROUNDS, beating players who've been at it far longer than you, you receive an immediate, unmistakable signal: I did better than expected. I improved.

More interestingly, this reward appears to be stronger for less experienced players, particularly when competing against more experienced ones. LPH accounts for this too. The intensity of the intrinsic reward is determined by the distance between an individual's prior prediction and the outcome obtained. The more surprising the result, the greater the update to one's knowledge, and therefore the greater the perceived learning progress. Before a game, a beginner will most likely expect to lose. So, when they win, especially against experienced players, the result is far more surprising, and the intrinsic reward is correspondingly greater. Put simply, the more unexpected your performance, the greater the perceived learning progress, the more satisfying it feels.

Self-Determination Theory: what really drives us to grow

This notion of perceived learning progress, central to curiosity-driven learning models, closely mirrors the concept of competence in Self-Determination Theory.

The joy of winning a round of ROUNDS, or the quiet frustration of losing again... These reactions reveal something deeper: we don't just want to do well, we need to. This is one of the core insights of Self-Determination Theory (SDT), a major framework in motivational psychology (Ryan & Deci, 2000).

SDT argues that human beings are not primarily driven by external rewards (beating a score or winning a trophy). Instead, we are naturally oriented toward growth for its own sake: an inner drive to develop, improve, and thrive, independent of any outside incentive. In this way, our well-being depends on the satisfaction of three basic psychological needs:

Autonomy: feeling that we are acting by our own choice, not under pressure;
Competence: feeling that we are effective in producing desired outcomes and in exercising our capacities;
Relatedness: feeling connected to others, and that we matter to them.

When these needs are met, we are motivated from within. When they are not, motivation fades, and so does well-being.

This is where SDT connects directly to the Learning Progress Hypothesis. The pleasure of noticing that you are improving is, at its core, a response to the need for competence. LPH identifies one specific mechanism, the perception of progress, that SDT places within a wider picture of what drives us.

But SDT goes beyond the immediate satisfaction of learning something new. It links intrinsic motivation to longer-term outcomes: sustained engagement, mental health, and overall quality of life. This is why some authors, such as Wehmeyer (1995), have proposed SDT as an educational and rehabilitation goal in itself: it’s not just a means to an end, but something to actively cultivate. From this perspective, supporting self-determination means taking individual differences into account; creating conditions that produce real, tangible benefits for the person; and sustaining this process across the lifespan, through ongoing learning, opportunities, and experiences.

The practical implication is straightforward: designing environments that support intrinsic motivation is not just about making things more enjoyable. It is about meeting fundamental human needs. This means accounting for individual differences – what feels stimulating for one person may feel frustrating or boring for another. It means giving people real choices. And it means creating opportunities for meaningful social connection as well as opportunities for learning.

These principles guided several works in our team: what happens when we design not just for performance, but for supporting intrinsic motivation ?

Example 1: Personalized cognitive training

As we age, attention is one of the first cognitive functions to decline. This affects not just memory or reaction time, but everyday autonomy and social participation: driving, following a conversation, managing daily tasks. Cognitive training programs designed to address this are nothing new: they are widely used in research settings and commercially available. And they do work, to some extent: studies report positive effects on attention and memory, but their benefits are often modest, and dropout rates are high. Most programs fail to account for the variability between and within individuals, a variability that increases with age. When everyone follows the same path, individual needs get lost.

So our team asked: what if the training adapted to you?

The result is a computerized cognitive training program (demo here) built around the Learning Progress Hypothesis (Adolphe, 2024), where the idea that perceiving your own improvement is itself a powerful source of intrinsic motivation. Concretely, this translates into several features designed to support the three basic needs of self-determination.

A virtual assistant guides users through the tasks, helping maintain their sense of competence throughout. Mini-stories and varied environments keep the experience engaging and stimulate curiosity. And real-time progression feedback allows participants to actually see themselves getting better, reinforcing their sense of achievement at every step.

Figure 2: Screenshot of the gamified cognitive training platform

At the heart of the program is the ZPDES algorithm, which continuously adapts the difficulty of tasks based on each participant's learning progress. Rooted in Vygotsky's concept of the Zone of Proximal Development, the idea is simple: keep tasks challenging enough to be stimulating, but achievable enough to maintain confidence. Not too easy, not too hard: the sweet spot where motivation and learning thrive together.

To put this to the test, this study compared two groups of older adults (Pech et al., in prep). One followed a traditional staircase learning path: the same progression for everyone, incrementally increasing in difficulty. The other was trained using the ZPDES algorithm, with difficulty adapting continuously to each individual's progress.

In total, 50 participants took part in the study: 26 in the personalized training group and 24 in the staircase training group. Participants were older adults (aged 60–85, mean age 68.6), generally in good health, with no reported sensory impairments. The two groups were comparable in terms of age, gender, and cognitive functioning.

The personalized group improved steadily and continuously across both weeks of training, while the control group progressed during the first week and then plateaued.

Figure 3. Number of transitions between activities in each group for the two studied populations. (From Pech et al., in prep)

But beyond raw performance, what matters most is what these gains did to the participants themselves. Those in the personalized group reported a significantly higher sense of competence, one of the three core needs identified by SDT. That feeling of mastering, improving, being capable, is precisely what SDT identifies as a core driver of well-being and quality of life. Although intrinsic motivation decreased over time in both groups, likely reflecting the demanding nature of the tasks, the personalized group showed less reliance on external motivation, reflecting a more self-determined dynamic overall.

Figure 4: Feeling across the learning (subjective questionnaires) with: load Index related to the NASA-TLX; SDI related to the SIMS; feeling of competence related to the TENS. (From Pech et al., in prep)

Finally, engagement turned out to be the strongest predictor of performance gains in the personalized group: the desire to keep going, to take on the next challenge, is what drove progress forward.

So when training adapts to you, you improve, you feel competent, and that feeling makes you want to keep going: a virtuous cycle, in the service of well-being.

Example 2: A self-determination-driven platform for aging at home

Our second example is HomeAssist, an assistive living platform we developed for frail older adults living alone (Sauzéon et al., 2022). The idea was simple but radical: instead of deciding what help people need, let them decide for themselves.

Built around a user-centered approach, it covers three broad areas of everyday life: daily activities (monitoring sleep, meal preparation, dressing, with personalized reminders if needed), home safety (alerts for open doors, unusual outings, or appliances left on), and social participation (internet browsing, photo sharing with family, collaborative games).

Figure 5: Content of HomeAssist platform—a set of sensors, a web-based catalog of assistive applications, and two touch screen tablets (main and secondary). ADL: activities of daily life. (Figure from Sauzéon et al., 2022)

What makes this platform different is not what it does, but rather how it was designed. Instead of imposing a fixed set of features, the platform allows each user, together with a technician and their caregivers, to choose which activities they want assistance with, and in what form. Features can be added, paused, or removed at any time, as needs and people evolve.

This design philosophy maps directly onto the three core needs identified by Deci and Ryan:

Autonomy is supported by giving users real control over their assistive environment: choosing which apps to activate, personalizing notifications, and being able to pause the system entirely for privacy;
Competence is raised through daily feedback on activity monitoring, helping users develop an accurate picture of their strengths and limitations, and build confidence in their ability to act on their environment;
And relatedness is fostered through features that maintain social connection: a digital picture frame shared with family, communication apps, collaborative games, keeping users embedded in their social world even when physical mobility is limited.

Does it work? Our early findings suggest it does, and in ways that go beyond simple ease of use.

A first study (Dupuy et al., 2016) tested this with 34 older adults living at home, in their early eighties on average (mean age ≈ 82), with moderate but typical age-related limitations, consistent with the kinds of everyday challenges this type of technology aims to support.

Half used the platform for six months; the other half did not.

Those who used HomeAssist reported improvements across every dimension of self-determination (feeling more autonomous, more capable, more in control). In contrast, those who didn‘t showed no change, or even a slight deterioration (empowerment).

Figure 6: Evolution of self-determination dimensions (A. Autonomy; B. Self-regulation; C. Empowerment; D. Self-realization) for equipped and control groups from T0 (in pale blue) to T6 (in dark blue) (from Dupuy et al., 2016)

This study also showed that self-determination is a determining factor of technology acceptance, with strong relationships revealed between self-determination dimensions and technology acceptance dimensions. These results suggest that designing for self-determination is a powerful lever for increasing the acceptance of assistive technologies, a persistent challenge in the field.

Crucially, a second study based on the same protocol (Dupuy et al., 2017) reported a stabilization of everyday functioning in the platform group, that would otherwise be expected to decline. Caregiver burden was also reduced, a finding that matters not just for the individuals themselves, but for the wider system of care around them: a reminder that supporting one person’s autonomy often lightens the load on those around them.

And our most recent analyses revealed that participants who reported higher levels of self-determination were also those who reported a better quality of life (Deyts et al., in prep). It sounds intuitive once you hear it. But it takes real work — in research, in design, in care — to make it happen. In the end, feeling better is not just about doing more: it's about feeling that what you do, you chose.

Learning, self-determination, and well-being: what the evidence tells us

The two examples above illustrate a broader pattern that is increasingly supported by the literature. Across very different contexts (cognitive rehabilitation, assistive technology, educational design) satisfying the basic psychological needs identified by SDT consistently predicts better outcomes, not just in terms of performance, but in terms of well-being and quality of life (Kindelberger et al., 2025). Well-being arises from having supportive relationships, engaging in meaningful activities, and achieving a sense of mastery over one's life, precisely the conditions that SDT-informed design tries to create. Satisfaction of these three basic needs contributes to enhanced engagement and pleasure in daily activities, thereby supporting overall well-being.

And learning, specifically, appears to play a direct role in this. Beyond its effects on motivation and performance, the experience of learning has been linked to broader health outcomes. Analyses of fieldwork data suggest that learning can develop psychosocial qualities that actively promote well-being (Hammond, 2004), just like curiosity (Priemysheva et al., 2025).

This brings us back to where we started: dopamine. Earlier, we discussed the neural evidence linking curiosity-driven learning to dopaminergic activity in the brain. There’s also some evidence of a positive relation between dopamine and well-being in certain conditions. For example, Rutledge et al. (2015) showed that dopamine increased happiness from some rewards and increased approach behavior toward potential gains, providing direct causal evidence that dopamine modulates subjective wellbeing in humans. A related study by Sharot et al. (2009) found that dopamine enhanced participants' expectations of future hedonic pleasure when imagining positive life events, suggesting dopamine shapes not just current mood but anticipatory wellbeing. However, as demonstrated in some neuroscience research but also in curious-driven learning literature, dopamine seems to contribute to wellbeing largely through motivation and reward pursuit rather than through pleasure itself, and both too little and too much dopaminergic disruption can impair subjective wellbeing.

The right kind of difficulty

Taken together, the evidence seems clear: learning, particularly when intrinsically motivated, is genuinely good for our well-being (see Priemysheva et al., 2025). And crucially, the intrinsic reward it generates is proportional to the perceived learning progress (Poli et al., 2024). So why not simply give every learner the most complex tasks possible, and wait for the enormous reward that comes with finally cracking them?

Naturally, things are more complicated than that. Recall that to maximize learning progress, difficulty needs to sit within the Zone of Proximal Development: challenging enough to be stimulating, but achievable enough to remain within reach. As demonstrated in some of our recent work, when task complexity is too high, reducing it actually improves intrinsic motivation and self-determination (Poupard et al., 2025; 2026). But the opposite is equally true: tasks that are too easy offer no meaningful learning progress, and therefore little intrinsic reward.

This is particularly well illustrated by a some finding from our team and others' experiments: when given a free choice, participants naturally gravitate toward tasks of intermediate difficulty relative to their own current competence level (Ten et al., 2021; 2025). In other words, we intuitively seek out the challenges that maximize our learning progress, and we adjust the difficulty based on our own prior knowledge. Then, difficulty is not the enemy of motivation and, in the right dose, is very engine. This relationship takes the form of an inverted-U curve between difficulty and curiosity: too easy, curiosity fades, too hard, it collapses. The sweet spot is in between.

Figure 7. The relationship between task difficulty and curiosity/learning progress (LP) depends on the type of difficulty involved. When difficulty is intrinsic to the task (goal-relevant elements, left panel), the relationship follows an inverted-U curve: curiosity and LP are maximized at an intermediate level of challenge, and drop off when tasks are either too easy or too difficult. When difficulty is extraneous to the task (goal-irrelevant elements, right panel), the relationship becomes negative and linear: any increase in difficulty reduces curiosity and LP, with no optimal zone.

However, not all difficulty is equal. Some theories of learning that we have recently begun to connect to our work on curiosity draw an important distinction here. Cognitive Load Theory, for instance, differentiates between intrinsic difficulty (i.e., the kind that is inherent to the task and directly useful for learning) and extraneous difficulty, which burdens the learner without contributing to their progress. Our recent work suggests that these two types of difficulty have meaningfully different effects on curiosity (Poupard et al., 2025; 2026). The inverted-U relationship between difficulty and curiosity appears to hold when difficulty is intrinsic. But when difficulty is extraneous, the relationship becomes linear and negative: more difficulty, less curiosity, with no sweet spot in sight.

|

Back to ROUNDS: raising the difficulty of the terrain, or matching you against stronger opponents, modulates the intrinsic difficulty of the task. Success under those conditions produces genuine learning progress, following the inverted-U curve. But imagine playing with an audience watching over your shoulder. The distraction, the social pressure, the divided attention, add difficulty too, but of a different kind. It has no direct bearing on the task itself and therefore generates no learning progress. It simply makes things harder, without making them more rewarding.

The practical implication is important: when designing learning environments, it is not enough to simply calibrate difficulty. It matters what kind of difficulty you are introducing. Supporting well-being through learning means maximizing the intrinsic challenge while minimizing the extraneous noise, keeping the learner in the zone where effort and reward go hand in hand.

References

Adolphe, M. (2024). Development and evaluation of AI-based personalization algorithms for attention training (Issue 2024BORD0176) [Theses, Université de Bordeaux]. https://theses.hal.science/tel-04884647
Coddington, L. T., Lindo, S. E., & Dudman, J. T. (2023). Mesolimbic dopamine adapts the rate of learning from action. Nature, 614(7947), 294–302. https://doi.org/10.1038/s41586-022-05614-z
Dupuy, L., Consel, C., & Sauzéon, H. (2016). Self determination-based design to achieve acceptance of assisted living technologies for older adults. Computers in Human Behavior, 65, 508–521. https://doi.org/10.1016/j.chb.2016.07.042
Dupuy, L., Froger, C., Consel, C., & Sauzéon, H. (2017). Everyday Functioning Benefits from an Assisted Living Platform amongst Frail Older Adults and Their Caregivers. Frontiers in Aging Neuroscience, Volume 9-2017. https://doi.org/10.3389/fnagi.2017.00302
Gruber, M. J., Gelman, B. D., & Ranganath, C. (2014). States of Curiosity Modulate Hippocampus-Dependent Learning via the Dopaminergic Circuit. Neuron, 84(2), 486–496. https://doi.org/10.1016/j.neuron.2014.08.060
Hammond, C. (2004). Impacts of lifelong learning upon emotional resilience, psychological and mental health: Fieldwork evidence. Oxford Review of Education, 30(4), 551–568. https://doi.org/10.1080/0305498042000303008
Kindelberger, C., Chauchard, E., Robin, F., Bret, A., & Guignard, J.-H. (2025). Contribution of Basic Psychological Needs to Students’ Well-Being Through Positive and Negative Affect. Scandinavian Journal of Psychology, (1–14). https://doi.org/10.1111/sjop.70067
Oudeyer, P.-Y., Gottlieb, J., & Lopes, M. (2016). Intrinsic motivation, curiosity, and learning. In Progress in Brain Research (Vol. 229, pp. 257–284). Elsevier. https://doi.org/10.1016/bs.pbr.2016.05.005
Poli, F., Koolen, M., Velázquez-Vargas, C. A., Ramos-Sanchez, J., Meyer, M., Mars, R. B., Rommelse, N., & Hunnius, S. (2024). Autistic traits foster effective curiosity-driven exploration. PLOS Computational Biology, 20(10), e1012453. https://doi.org/10.1371/journal.pcbi.1012453
Poupard, M., Larrue, F., Bertrand, M., Liguoro, D., Sauzéon, H., & Tricot, A. (2026). Reducing Load, Fostering Curiosity: Empirical Validation of the IMCLM-XR. https://hal.science/hal-05219664
Poupard, M., Larrue, F., Bertrand, M., Liguoro, D., Tricot, A., & Sauzéon, H. (2025). Using virtual reality for enhancing neuroanatomy learning by optimizing cognitive load and intrinsic motivation. Computers & Education, 235, 105332. https://doi.org/10.1016/j.compedu.2025.105332
Priemysheva, A., Akhtar, S., & Gruber, M. (2025). The relationships between curiosity and wellbeing: A systematic review (Rcjv7_v1). PsyArXiv. https://doi.org/10.31234/osf.io/rcjv7
Rutledge, R. B., Skandali, N., Dayan, P., & Dolan, R. J. (2015). Dopaminergic Modulation of Decision Making and Subjective Well-Being. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 35(27), 9811–9822. https://doi.org/10.1523/JNEUROSCI.0702-15.2015
Ryan, R. M., & Deci, E. L. (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being. American Psychologist, 11.
Sauzéon, H., Edjolo, A., Amieva, H., Consel, C., & Pérès, K. (2022). Effectiveness of an Ambient Assisted Living (HomeAssist) Platform for Supporting Aging in Place of Older Adults With Frailty: Protocol for a Quasi-Experimental Study. JMIR Res Protoc, 11(10), e33351. https://doi.org/10.2196/33351
Sharot, T., Shiner, T., Brown, A. C., Fan, J., & Dolan, R. J. (2009). Dopamine Enhances Expectation of Pleasure in Humans. Current Biology, 19(24), 2077–2080. https://doi.org/10.1016/j.cub.2009.10.025
Ten, A., Kaushik, P., Oudeyer, P.-Y., & Gottlieb, J. (2021). Humans monitor learning progress in curiosity-driven exploration. Nature Communications, 12(1), 5972. https://doi.org/10.1038/s41467-021-26196-w
Ten, A., Oudeyer, P.-Y., Sakaki, M., & Murayama, K. (2025). The Curious U: Integrating Theories Linking Knowledge and Information-Seeking Behavior. Open Mind, 9, 1763–1785. https://doi.org/10.1162/OPMI.a.41
Wehmeyer, M. L. (1995). Intra-individual factors influencing efficacy of interventions for stereotyped behaviours: A meta-analysis. Journal of Intellectual Disability Research, 39(3), 205–214. https://doi.org/10.1111/j.1365-2788.1995.tb00503.x

Page updated

Google Sites

Report abuse