I am very happy to share our new paper, “Spin-based modelling of perception as emergent from contextualized internal evaluation,” recently published in Royal Society Open Science.
This project is quite far from my usual research on strongly correlated electrons and superconductivity. I was brought into this field by my friend and colleague Belén Valenzuela, herself a theoretical physicist, who became interested in these questions for personal reasons. For me, one of the most exciting aspects of the project was precisely the possibility of taking tools I know very well – the Ising model, phase transitions and Landau field theory – and using them in such a different context.
The starting question is simple: how does the perception of a sensation emerge from the many evaluations that surround it?
Take back pain as an example. The final experience of the sensation may depend on many different elements: previous episodes of pain, beliefs about what movement is safe or harmful, medical advice, bodily signals and the situation in which the sensation occurs. These evaluations are not independent, but influence one another and together shape the perceptual state that eventually emerges.
In our model, we represent these local evaluations as interacting spins. A spin can encode a positive or negative evaluation, for example, whether something is interpreted as reassuring or threatening, and, in the spin-1 version, it can also take a neutral value. The collective configuration of all these local evaluations then determines the macroscopic perceptual state.
The surrounding context enters through the analogue of temperature. Here, temperature does not literally mean heat: it represents how informative the context is with respect to the sensation. For example, when we are absorbed in a pleasant or engaging situation, the surrounding context may carry little information about the back pain, and the sensation can lose salience or fade from awareness. In a context strongly associated with threat or previous pain, instead, the same sensation may become much more prominent.
This is where statistical physics becomes useful. Starting from the microscopic spin models, we coarse-grain the system and derive an effective Landau-type landscape for the macroscopic perceptual state. In this way, quantities introduced phenomenologically at the macroscopic level can be connected to microscopic ingredients such as the interaction between local evaluations, the contextual information and the availability of a neutral state.
One of the main results is that introducing a neutral state changes the collective behavior in a non-trivial way. It modifies the transition threshold and the shape of the perceptual landscape, making the system more sensitive to context. In the Blume–Capel model, tuning the energetic weight of the neutral state can even change the nature of the transition, producing bistability and hysteresis: persistent perceptual states emerging from a fully symmetric microscopic model.
What I found particularly stimulating in this work was seeing how some of the most familiar ideas of statistical physics like emergence, collective behaviour, phase transitions and coarse graining can acquire a completely different meaning when applied to another scientific problem.
Spin-based modelling of perception as emergent from contextualized internal evaluation
Royal Society Open Science 13, 252221 (2026)
