We are used to thinking of vision as something instantaneous: we open our eyes, and the world is already assembled. A table remains a table, a face a face, a line a line. Behind this apparent simplicity lies the remarkably complex work of the brain. Different areas of the cortex simultaneously analyze individual properties of an image, exchange signals, and gradually arrive at a shared state. A new study has revealed one of the mechanisms by which such a "consensus" can arise right inside the visual cortex.
The study, published on 18 September 2026 in the journal Nature Neuroscience, was conducted by neuroscientists from three institutions. The lead author — Mitra Javadzadeh (now a Fellow at Cold Spring Harbor Laboratory) — headed a team joined by researchers from the University of Cambridge and University College London. Together they studied the interaction of two visual areas of the mouse brain — the primary visual cortex V1 and the lateromedial area LM. Both participate in processing visual information, but they do this work at different levels. V1 is the first to receive the sensory signal from the thalamus, while LM is involved in more complex processing of the image. Information constantly circulates between them in both directions.
The researchers simultaneously recorded the activity of 194 V1 neurons and 228 LM neurons in seven mice. The animals performed a task distinguishing the orientation of visual stimuli — they had to tell apart lines tilted at different angles and received a reward for a correct answer. The scientists observed how the two areas respond to the same image and how their joint activity changes over time.
It was here that an interesting effect emerged, which they called "consensus building." When the activity of V1 and LM added up into a coordinated pattern, such a state persisted considerably longer — about 400 milliseconds. When the signals diverged, the dynamics quickly transitioned into another state. The system seemed to naturally favor those configurations in which both areas arrived at a similar "decision."
To describe this behavior, the researchers used a mathematical model of dynamical systems. In it, the mutual excitatory connections between V1 and LM formed what is called an approximate line attractor. If we translate this term from the language of mathematics into the language of everyday experience, it refers to a space of states in which certain combinations of neural activity are able to persist considerably longer than others — thanks to the mutual reinforcement of signals between the regions.
The resulting picture is quite striking: agreement between areas literally changes the flow of time within the neural network. Some states replace one another quickly, while others become more stable and continue to exist for hundreds of milliseconds. On the scale of neural processes, this is already a substantial amount of time — an opportunity for information to take hold, travel further through the network, and influence subsequent processing, and therefore perception as well.
The researchers tested the model experimentally by temporarily suppressing the activity of individual areas using optogenetics — a method that makes it possible to control the activity of certain nerve cells with light by acting on specially modified light-sensitive proteins. The data obtained confirmed the theory's key prediction: the dynamics of the system did indeed depend on the mutual connections between V1 and LM. Coordinated states developed more slowly, while divergent ones developed faster, which matched the model's predictions.
This result offers a different view of how the brain creates a unified perception. Instead of a single center assembling ready-made pieces of information, we see a network of areas that continuously influence one another. Unity arises not somewhere in a higher governing center, but right within the very interaction between regions.
One can imagine an ensemble in which each musician hears the others and at the same time changes their own playing. The music is born right in the process of mutual tuning, without a conductor setting the tempo. In a similar way, visual areas can form a shared image: the signal moves forward, comes back, changes again, and gradually stabilizes into a coordinated configuration. Each time, the brain literally negotiates with itself.
The study focuses on the visual system of mice, so for now its significance lies primarily in understanding the mechanisms of perception and information processing in the cerebral cortex. At the same time, the idea itself opens up a far broader path for research. If similar dynamics operate between many areas of the neocortex, they may help explain how distributed neural processes become a unified perception of what is happening.
The next question sounds even more intriguing: what happens when different senses offer the brain different versions of reality? For example, the eyes report one thing, hearing another, and bodily sensations add a third piece of information. Perhaps here too a unified experience is born through the gradual reconciliation of many streams — the same mechanism of consensus, now operating at a higher level of the brain's organization.
The work of Mitra Javadzadeh and her colleagues brings neuroscience closer to understanding one of the most fundamental principles of how the brain works. Unity may arise as a dynamic property of the network — through continuous exchange, mutual influence, and the stabilization of shared states. This is no longer a passive gathering of information, but an active, living reconciliation.
What we experience as a single, whole world may be born every second out of a vast number of such small acts of reconciliation. Millions of neurons continuously "talk" to one another, and out of this living dialogue the reality we see before us gradually takes shape. The brain works not as a rigid hierarchy, but as a democracy in which every area has a voice.
