Summary: In the 1960s, Nobel laureates David Hubel and Torsten Wiesel proposed that visual perception is built in a stepwise, “bottom-up” fashion: the brain constructs complex images from simple elements such as edges and lines. For decades researchers debated whether this orientation and feature selectivity originates in the thalamus—the eye’s relay station—or whether it is assembled later in the cortex.
A recent high-resolution study has now settled that debate. Combining advanced two-photon imaging with optogenetics to temporarily silence cortical activity, the research team demonstrated that thalamic inputs deliver broad, non-specific signals while the cortex refines those signals into precise orientation-selective responses. In short, the cortex is the site where orientation selectivity is constructed.
Key Facts
- Validation of a classic model: The findings provide direct evidence for the feedforward, stepwise computation model originally proposed by Hubel and Wiesel.
- Thalamic inputs are broad: Signals arriving from the thalamus are strong but broadly tuned, not distinguishing specific line orientations on their own.
- Cortex creates selectivity: Orientation selectivity—our ability to distinguish vertical from horizontal lines—emerges only after cortical circuitry processes thalamic input.
- Different learning roles: Synapses within the cortex exhibit calcium signals associated with plasticity and learning, whereas thalamocortical synapses do not show those same calcium dynamics.
- New experimental tools: The team used two-photon microscopy and fluorescent reporters to monitor individual synapses firing in living brains, combined with optogenetic silencing to separate input sources.
Source: TUM
Background: Hubel and Wiesel’s model, formulated in the 1960s, proposed that visual perception arises from orderly, hierarchical computations. In that framework, neurons in primary visual cortex respond selectively to elementary features such as edges and orientations. While influential, the model left open whether feature selectivity originates in the thalamus or is constructed within cortical circuits.
The new study addresses that question by examining signal transmission at the level of individual synapses between the thalamus and the primary visual cortex—an analysis that was previously not possible at this resolution in vivo.
The research was led by Prof. Arthur Konnerth, Dr. Yang Chen, and PhD student Marinus Kloos at the Institute of Neuroscience, TUM School of Medicine and Health, in collaboration with the Munich Cluster for Systems Neurology (SyNergy). Using an approach that combines two-photon glutamate imaging with optogenetic cortical silencing, the team measured activity at single synapses in intact mouse brains. The full results were published in the journal Science.
“These results show just how prescient Hubel and Wiesel’s model was,” says Prof. Konnerth. “Their ideas about hierarchical sensory processing remain foundational for modern neuroscience and continue to inspire computational approaches, including artificial neural networks.”
What the researchers did
Visual signals travel from the eye to the thalamus, and from there to the primary visual cortex (V1). In V1, simple features—edges, contrast, orientation—are analyzed. The team focused on layer 4 of mouse V1, where thalamic axons make synapses onto cortical neurons, and tracked how those inputs contribute to orientation tuning.
Using two-photon microscopy, the researchers visualized individual dendritic spines and synapses in living mice. They expressed fluorescent reporter proteins that emit light when synaptic glutamate is released, enabling direct measurement of synaptic activation in real time. While presenting simple visual patterns such as horizontal and vertical gratings, the team mapped which synapses responded preferentially to particular orientations.
To separate incoming thalamic signals from activity generated within the cortex, the team used optogenetics to silence intracortical activity transiently. By selectively switching off cortical circuits with light, they could determine whether a given synaptic response persisted—indicating direct thalamic drive—or disappeared, indicating it depended on intracortical processing.
This combination of tools allowed the researchers to quantify thalamocortical versus corticocortical inputs at single-synapse resolution. The results were decisive: thalamic inputs were robust yet broadly tuned for orientation, whereas orientation selectivity only became apparent after intracortical processing. These observations directly support the core prediction of the Hubel and Wiesel feedforward model.
Implications
Beyond resolving a longstanding debate about where orientation selectivity arises, the study introduces a flexible experimental method to probe synaptic function in the intact brain. This approach can be applied to other circuits and cell types and may help identify circuit-level dysfunction in neurological conditions such as Alzheimer’s disease.
A notable and unexpected discovery was the difference in plasticity markers between synapse types: corticocortical synapses exhibited postsynaptic calcium signals associated with learning and adaptability, while thalamocortical synapses lacked those calcium dynamics. This suggests that the cortex, rather than thalamic inputs, is the primary locus for experience-dependent plasticity that sculpts sensory representations.
“The finding challenges the assumption that all synapses are equally plastic,” Konnerth notes. “Instead, it points to specialized roles for different synapse classes in computation and learning.”
Key Questions Answered:
A: It reveals how sensory processing is organized in the brain. If the thalamus performed detailed feature extraction, the cortex would be a passive receiver. The study shows instead that the cortex performs hierarchical integration—combining simple inputs into complex representations—which is a fundamental principle that also underlies many artificial neural network designs.
A: Researchers found that thalamocortical inputs appear relatively fixed, while intracortical synapses show calcium signals linked to plasticity. In effect, the cortex contains the adaptive circuitry that learns and refines sensory responses, whereas thalamic inputs provide stable, broadly tuned drive.
A: The team used optogenetics, a method that makes selected neurons light-sensitive. Delivering light to those neurons temporarily suppresses intracortical activity, allowing experimenters to observe thalamic input in isolation and determine which synaptic responses depend on cortical processing.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The underlying journal paper was reviewed in full.
- Additional context was added by editorial staff.
About this visual neuroscience research news
Author: Ulrich Meyer ([email protected])
Source: TUM
Contact: Ulrich Meyer – TUM
Image: Image credited to Neuroscience News
Original Research: Closed access. “Thalamic activation of the visual cortex at the single-synapse level” by Yang Chen, Marinus Kloos, Zsuzsanna Varga, Yonghai Zhang, Inken Piro, Tatsuo K. Sato, Bert Sakmann, Israel Nelken, and Arthur Konnerth. Science. DOI: 10.1126/science.aec9923
Abstract
Thalamic activation of the visual cortex at the single-synapse level
Understanding thalamocortical activation at individual synapses is essential for explaining how the cortex processes sensory input. This study examined thalamocortical computation underlying the emergence of orientation selectivity in the mammalian primary visual cortex (V1).
Using two-photon glutamate imaging together with optogenetic cortical silencing in vivo, the authors identified and characterized thalamocortical synapses onto layer 4 neurons in mouse V1. They report that thalamocortical-recipient spines lacked postsynaptic Ca2+ signals that were present at corticocortical-recipient spines.
These results validate the core predictions of Hubel and Wiesel’s feedforward model and reveal distinct synaptic properties that are critical for cortical computation and plasticity.