Learning to Like It
A Nature paper on how mice learn to prefer a flavour after the body has felt its effects gives managers and marketers a careful analogy, which is that a first reaction to something new can be a poor guide to what people will come to value once they have had a fair second and third try.

The first sip need not be the verdict
A first taste of black coffee can be surprisingly unpleasant, as ScienceDaily’s version of the University of California, San Francisco release puts it, and yet millions of people come to look forward to the same cup. Early dislike is easy to read as a final answer, whether the thing in question is a flavour, a habit, a piece of software or a new way of working, and a new study offers one way to think about being more patient with it.
Through the STAR Framework, the first sip lands differently for each of the four lenses, and the Socialiser lens is a good place to start because, on my reading, it is more likely to try something again in company than alone. A shared second try gives the person a reason to return that does not depend on the first impression, and it lets them borrow someone else’s experience of the payoff before they have felt it themselves. The lens readings in this article are my own interpretation of the framework and have not been tested against the study’s data, which concerns mice and does not mention lenses.
What the Nature paper reported
The paper is “The neural basis of flavour-nutrient learning”, published in Nature on 4 November 2026 by James Grove, Zachary Knight and colleagues, with Knight as senior author, and ScienceDaily carried the UCSF release on 9 November. The authors describe a slow signalling pathway from the gut that lets an animal learn about the nutrient content of food after it has been eaten, and they report that nutrients delivered directly into the stomach of mice, either sugar water or a milkshake-like liquid, triggered a sustained response in a small region the paper calls the anterior basolateral amygdala. Delivering the nutrients directly allowed the team to separate the effect of what reached the gut from the experience of tasting.
The two key experiments pointed in opposite directions. When the researchers blocked the pathway, the mice carried on eating normally but lost the ability to learn that a new flavour predicted calories, and when they artificially stimulated the pathway after the mice tasted a flavour with no calories, the mice came to prefer it, so the stimulation stood in for the nutrients. Knight is quoted as saying, “We’ve shown that this slow dopamine system is what drives us to learn to like certain foods over time, and we can see where and how that happens.”
The Thinker lens, which wants to understand why a second attempt should go better, finds its answer in the sequence the paper describes, in which a flavour is tasted first and its consequences arrive later, so that, in this pathway, preference forms once the two have been linked. That is my reading, and the paper reports the sequence in mice and not in people’s decisions.
Several limits decide how far the coffee story can be pushed. The experiments used sugar water and a milkshake-like liquid, so they concern nutrients, and neither caffeine nor coffee appears in the paper’s abstract or in the preprint version of the full text, which is the version I could read. The coffee line comes from Knight’s illustration in the release, in which he says that once the connection is made between the taste of coffee and caffeine’s effects, “that coffee starts to taste a lot better”, and that is an application of the idea and not something the experiments tested. The causal experiments were in mice, and the mice showed a preference, which is a choice and not a report of liking. The human work in the paper consists of imaging, with a re-analysis of an earlier PET study of 13 men by Dana Small and an fMRI study of 38 adults led by Marc Tittgemeyer, and it found that the same region responded to nutrients, which shows a response and does not show that people learned to like anything.
Why a first reaction is a poor guide
Learning from consequences is an old idea in this field, and it is worth saying so because it makes the finding more credible and not less. Sclafani’s 1997 review notes that nutrient-based learning of flavour preferences was firmly established in animals by the 1970s, so the Nature paper adds a pathway to a phenomenon that researchers had already described.
Evidence from children points in a similar direction, although it concerns repeated exposure and not the gut pathway, with a small and early study. Birch and Marlin gave 14 two-year-olds novel cheeses in one experiment and novel fruits in another, offering each food between 0 and 20 times over 25 or 26 days across the two experiments, and when the children later chose between pairs of foods, preference rose with the number of earlier exposures, with 13 of the 14 children choosing the more familiar food in more of their choices than chance would predict. The sample was tiny, the foods were cheeses and fruits, and the children chose within ten days of the exposures, so it is best read as a signal and not as a rule, but the title of the paper, “I don’t like it; I never tried it”, captures the point neatly.
For people working on change, the careful claim is an analogy. New tools, routines and ways of working often meet an early dislike that is read as a verdict, and the evidence above suggests that early dislike can be a poor predictor of where liking ends up, although the Nature paper does not say that about workplaces and some things people never come to like. Dear Algorithm makes a related point about the systems that profile us: “The algorithm’s error isn’t that it categorises people. It’s that it categorises them once and calls it done.” A first reaction that is filed once and never revisited makes the same error.
The Adventurer lens, which values autonomy, exploration and freedom to act, is on my reading the one most drawn to try the unfamiliar again by choice, which is useful and also a risk, because a change that only works for the Adventurer lens may never give the other three the repeat exposure they need, and a rollout that depends on volunteers may see mostly the people who needed no encouragement.
Giving it a fair second try
The Realist lens, which values security, stability and reliability, along with simplicity and ease, needs the payoff to arrive quickly and smoothly, since on my reading it will not keep returning to something that is awkward while the benefit is still hidden. For a change that matters, that suggests making the second attempt easier than the first, and the suggestions below are mine and not findings of the paper:
- Agree in advance how many tries something gets before a decision is made, so that a first reaction is recorded as a first reaction and not as the outcome.
- Put the payoff early and make it visible, because liking in the paper followed the consequences and a benefit that nobody can see cannot do that work.
- Pair a first try with someone who has already reached the payoff, which gives the Socialiser lens a reason to return and the Thinker lens a worked example.
- Ask what exactly put people off, and separate “this is unfamiliar” from “this does not work”, because only the first is likely to fade with repetition and the second should be fixed.
Dual Process Theory offers one way to read the pattern, in which a first reaction is a fast response while a settled liking builds more slowly from repeated experience of the consequences. Appraisal Theory of Emotion adds that the way a first experience is judged shapes whether a second one happens, which is why a poor first impression deserves an explicit plan and not a shrug. Neither theory was tested in the paper, and both readings are mine.
If you would like to see your own decision-making style, and which first impressions you tend to trust too much, the STAR Cognitive Profile Assessment (33 questions) at starassess.com is a sensible place to begin.
David Chadderton is a behavioural strategist, Chief Marketing Officer and creator of the STAR Framework. He writes The Unoptimised Human.
References
Grove, J. C. R., Hakimi, A. M., Li, Q., Zhang, J., Backes, H., Kuzmanovic, B., Choi, J., Ubadiah, V., Qiu, L., Liu, Z., Small, D. M., Tittgemeyer, M., & Knight, Z. A. (2026). The neural basis of flavour-nutrient learning. Nature. Published 4 November 2026.
University of California, San Francisco (2026). How your brain learned to love coffee, chocolate, fat and sugar. Medical Xpress, 4 November 2026.
University of California, San Francisco (2026). Scientists finally discover why we learn to love bitter coffee. ScienceDaily, 9 November 2026.
Sclafani, A. (1997). Learned controls of ingestive behaviour. Appetite, 29(2), 153-158.
Birch, L. L., & Marlin, D. W. (1982). I don’t like it; I never tried it: Effects of exposure on two-year-old children’s food preferences. Appetite, 3(4), 353-360.
Evans, J. St. B. T., & Stanovich, K. E. (2013). Dual-process theories of higher cognition: Advancing the debate. Perspectives on Psychological Science, 8(3), 223-241.
Moors, A., Ellsworth, P. C., Scherer, K. R., & Frijda, N. H. (2013). Appraisal theories of emotion: State of the art and future development. Emotion Review, 5(2), 119-124.
Chadderton, D. (2026) Dear Algorithm, It’s Not Me, It’s You: Putting the Personal Back Into Personalisation. 2nd edn. 16 May 2026.
Eight minutes. Thirty-three questions.
A precise picture of how you operate: motivations, decision style, and blind spots. Free to start.
- S
- T
- A
- R
The canonical source.
Essays, when there is something worth sending. Choose the subjects you want. I will ask you to confirm the address before I write.