A scientific dispute spanning six decades about fundamental mechanisms of visual perception in mammals has now been settled. Researchers at TUM have succeeded in observing the visual information flow from neuron to neuron. Their findings confirm the validity of the 1981 Nobel Prize-winning model by David Hubel and Torsten Wiesel, which had remained controversial in some aspects.
Astrid Eckert / TUM Already in the 1960s, Hubel and Wiesel proposed a model according to which visual perception is the result of orderly, stepwise computations in the brain - with specialized neurons in the cortex responding selectively to specific features, such as edges or the orientations of moving objects. While widely celebrated, important aspects of the theory remained an issue of debate: does this feature selectivity already originate in the thalamus, or does it emerge later in the cortex? The new study addresses this question directly by analyzing signal transmission at individual synapses between the thalamus and the visual cortex - something that had not previously been possible.
The research team, led by Prof. Arthur Konnerth, Dr. Yang Chen, and PhD student Marinus Kloos at the Institute of Neuroscience at the TUM School of Medicine and Health and the Munich Cluster for Systems Neurology (SyNergy), developed a high-resolution imaging approach to measure synaptic activity in the intact brain. Their findings directly confirm core predictions of the Hubel and Wiesel model. The new research results were published in the prestigious journal Science.
"Our results highlight how remarkably accurate and forward-looking Hubel and Wiesel's insights were," says Prof. Konnerth. "Modern neuroscience - and even artificial neural networks - continue to build on their principles. Learning from biological systems remains a powerful driver of technological innovation."
What exactly did the TUM researchers do?
When we see, signals travel from the eye first to the thalamus, a relay station deep in the brain, and from there to the visual cortex at the back of the head. In the first area of this visual cortex, known as the primary visual cortex, simple image features like edges, contrast, and orientation are processed. The TUM researchers specifically examined this segment - the connection from the thalamus to this initial visual area of the cortex - in mice.
Using two-photon microscopy, the researchers visualized individual synapses in the living brains. They employed fluorescent proteins that emit light when synaptic transmission occurs, allowing them to track activity at specific neuronal contact points in real time. At the same time, the animals were presented with simple visual stimuli, such as horizontal and vertical stripes, enabling the team to map which synapses responded to which orientations.
To distinguish direct input from the thalamus from signals generated within the cortex, the researchers used optogenetics. They equipped certain neurons with light-sensitive proteins and could thus temporarily "mute" parts of the cortex with light. So, they could determine whether synaptic activity persisted (indicating thalamic input) or disappeared (indicating intracortical processing).
This approach allowed the team to separately quantify thalamocortical and corticocortical inputs. The results were clear: signals arriving from the thalamus were robust but largely non-specific with respect to orientation. In contrast, orientation selectivity - such as distinguishing horizontal from vertical lines - emerged only through processing within cortical circuits.
These findings resolve a long-standing controversy. The new data show directly that, in mammals, the cortex constructs this information step by step from broadly tuned inputs - precisely as predicted by Hubel and Wiesel.
Implications for neuroscience and beyond
Beyond confirming a foundational theory, the study introduces a versatile method for analyzing synaptic function. According to the researchers, this technique can be applied to a wide range of neuron types and may help identify dysfunctional circuits in neurological disorders such as Alzheimer's disease.
The study also revealed a fundamental difference between synapse types. Synapses within the cortex (corticocortical synapses) exhibited calcium signals associated with learning and plasticity, whereas synapses from the thalamus (thalamocortical synapses) did not.
"This was an unexpected finding," Konnerth explains. "It suggests that not all synapses have the same capacity for adaptation and learning, challenging long-standing assumptions in neuroscience."
Yang Chen, Marinus Kloos et al: „Thalamic activation of the visual cortex at the single-synapse level", published in: Science, 26 March 2026, DOI: 10.1126/science.aec9923
Many of the researchers work at the Excellence Cluster Munich Cluster for Systems Neurology (SyNergy) , funded by the federal and state governments under the Excellence Initiative. Researchers from Kagoshima University, Japan, the Max Planck Institute for Biological Intelligence in Martinsried, and the Hebrew University of Jerusalem contributed to the study.
