In a landmark achievement for neuroscience, researchers at the Technical University of Munich (TUM) have provided definitive proof supporting a longstanding theory of visual perception, resolving a scientific dispute that has persisted for over six decades. Using an innovative synaptic imaging method, the team visualized signal flow in the brain at the unprecedented level of individual synapses, confirming how mammals process visual information from the eyes to conscious perception.
The study, published in the prestigious journal Science on March 26, 2026, demonstrates that orientation selectivity—crucial for distinguishing edges, shapes, and movements in our visual field—emerges through precise computations within the visual cortex, rather than being pre-wired in earlier brain stages. This breakthrough not only validates foundational work by Nobel laureates David Hubel and Torsten Wiesel but also opens new avenues for understanding brain function and disorders affecting vision.
🔬 The Roots of the Visual Perception Puzzle: Hubel and Wiesel's Nobel-Winning Model
The journey to this resolution traces back to the 1960s, when neurophysiologists David Hubel and Torsten Wiesel conducted groundbreaking experiments on cats and monkeys. They discovered that neurons in the primary visual cortex (V1), the brain's first dedicated visual processing area, respond selectively to specific orientations of lines or edges. For instance, some cells fire vigorously for horizontal bars but ignore vertical ones.
Hubel and Wiesel proposed a hierarchical model: simple cells in V1 receive inputs from the lateral geniculate nucleus (LGN) in the thalamus—relay stations from the retina—and integrate them convergently to create orientation-tuned responses. Complex cells then pool these signals for motion sensitivity and larger receptive fields. This 'ice-cube' or feedforward model earned them the 1981 Nobel Prize in Physiology or Medicine and laid the groundwork for modern computational neuroscience, influencing everything from brain mapping to artificial intelligence vision systems.
However, direct evidence at the synaptic level—where neurons communicate via chemical transmitters—was lacking due to technological limitations. Early recordings measured population activity, not individual connections, leaving room for interpretation.
The 60-Year Dispute: Precise Thalamic Inputs or Cortical Invention?
The controversy ignited in the late 20th century as new data emerged, particularly from rodents. Some studies suggested that thalamic neurons already exhibited orientation selectivity, implying the cortex merely refines pre-tuned signals rather than building them from scratch. Others argued for broad, non-specific thalamic inputs sharpened by intracortical wiring.
This debate had profound implications. If selectivity originates in the thalamus, visual processing might be more hardwired; if cortical, it underscores the brain's remarkable plasticity and computational power. Resolving it required observing synaptic transmission in vivo—while the animal views stimuli—in layer 4 of V1, where thalamocortical axons first synapse with cortical dendrites.
Prior methods like electron microscopy offered static snapshots, while calcium imaging lacked synapse-specific resolution. The impasse persisted until advances in optical tools enabled the current feat.
Meet the Research Powerhouse: TUM and the Munich Excellence Cluster
Leading the charge is Prof. Arthur Konnerth, Hertie Senior Professor and TUM Emeritus of Excellence at the Institute of Neuroscience within TUM School of Medicine and Health. His team, including Dr. Yang Chen and PhD student Marinus Kloos, collaborated with experts from Kagoshima University in Japan, the Max Planck Institute for Biological Intelligence in Martinsried, and the Hebrew University of Jerusalem.
The work was bolstered by the Munich Cluster for Systems Neurology (SyNergy), a joint TUM-LMU Munich initiative funded under Germany's Excellence Strategy. This cluster fosters interdisciplinary research on brain diseases, pooling resources for cutting-edge imaging and animal models. TUM's neuroscience hub exemplifies Europe's leadership in systems biology, attracting top talent and investments exceeding €100 million since 2019.
Such collaborations highlight how German universities drive global breakthroughs, with TUM ranking among Europe's top innovators in life sciences per recent QS subject rankings.
Revolutionary Methods: Peering Inside Synapses in a Living Brain
The study's ingenuity lies in its technical prowess. Researchers employed two-photon microscopy—a laser-based technique that excites fluorescent molecules deep in tissue without scattering—to image awake, head-fixed mice viewing drifting gratings of horizontal or vertical stripes.
Key innovations:
- Glutamate and calcium sensors: Genetically encoded proteins like iGluSnFR and jGCaMP8 lit up precisely when synaptic vesicles released glutamate or triggered postsynaptic calcium influx, pinpointing active synapses at spine resolution (sub-micron scale).
- Optogenetic silencing: Channelrhodopsin in cortical neurons allowed blue light pulses to hyperpolarize and mute intracortical circuits temporarily. Persistent activity under silencing revealed direct thalamocortical inputs; silenced activity indicated cortical relays.
- Anatomical tracing: Viral vectors labeled thalamic axons green and cortical ones magenta, distinguishing input types spatially.
This pipeline quantified over 1,000 synapses per experiment, mapping tuning curves—response strength vs. orientation—for each. Processing involved custom algorithms to align frames and extract signals amid motion artifacts.
Groundbreaking Findings: Broad Inputs, Precise Outputs
The results unequivocally support Hubel and Wiesel. Thalamocortical synapses (from LGN to V1 layer 4) responded robustly to visual stimuli but lacked orientation selectivity—their receptive fields were large and broadly tuned, akin to 'centers' without edges.
In stark contrast, corticocortical synapses (lateral connections within V1) showed sharp orientation tuning matching the postsynaptic neuron's preference. Spatial analysis confirmed convergent wiring: multiple thalamic synapses aligned to drive simple-cell selectivity.
A surprise emerged in calcium dynamics. Corticocortical synapses evoked strong postsynaptic calcium transients—hallmarks of Hebbian plasticity ('cells that fire together wire together')—while thalamocortical ones did not, even under identical glutamate release. This dichotomy suggests thalamic inputs provide stable 'feedforward' drive, while cortical synapses adapt for learning.
Statistics underscore precision: thalamic synapses averaged 45° orientation bandwidth (broad), vs. 30° for cortical (sharp); convergence ratios hit 5-10 inputs per simple cell, as theorized.
Implications for Neuroscience and Beyond
This synaptic proof cements the feedforward model, explaining how ~1 million LGN neurons compress retinal data into 300,000 V1 neurons via selective synapses. It bridges classical electrophysiology with molecular imaging, validating predictions untestable for 60 years.
For artificial intelligence, it reinforces convolutional neural networks (CNNs), inspired by Hubel-Wiesel, where layers hierarchically extract features. Biological insights could refine AI robustness to adversarial images or low-light vision.The full study in Science details these computations mathematically.
Clinically, disrupted thalamocortical wiring links to amblyopia ('lazy eye'), schizophrenia hallucinations, and Alzheimer's visual deficits. The imaging toolkit promises diagnostics, e.g., synapse loss quantification in vivo.
Synapse Types and Plasticity: A New Paradigm
The calcium asymmetry challenges uniform synapse views. Thalamocortical junctions may prioritize reliability over plasticity, acting as 'anchors' for cortical refinement. This could explain critical periods in visual development, where early deprivation rewires circuits irreversibly.
Prof. Konnerth notes: “Our results highlight how remarkably accurate and forward-looking Hubel and Wiesel’s insights were.” This finding, “unexpected,” urges re-examination of synaptic rules across brain regions.
European Leadership in Neuroscience Innovation
TUM's feat exemplifies Germany's €3.5 billion Excellence Strategy investing in clusters like SyNergy, yielding 20+ ERC grants yearly. Compared to US NIH funding biases toward disease models, Europe's strength lies in fundamental mechanisms, fostering spin-offs like AI startups modeling V1.
LMU Munich complements with synaptic electron microscopy, creating a Bavarian neuroscience powerhouse rivaling Harvard or Stanford. Student training via PhD programs benefits immensely, with alumni securing faculty posts Europe-wide.
Future Horizons: From Synapses to Therapies
The method scales to humans via minimally invasive endoscopy or fMRI analogs. Trials target glaucoma (thalamic degeneration) or stroke recovery. AI integration could simulate patient-specific circuits for personalized rehab.
Challenges remain: scaling to deeper layers or humans. Yet, as Konnerth states, “Learning from biological systems remains a powerful driver of technological innovation.” Europe's universities, via Horizon Europe (€95 billion), position to lead.TUM Excellence Strategy page
Photo by Keren Fedida on Unsplash
Stakeholder Perspectives and Broader Impact
Neurosociety leaders hail it as “synaptic Nobel redux.” Industry eyes bio-inspired cameras for autonomous vehicles. For students, it underscores optical genetics' rise—enrollments in TUM neuroscience up 25% post-2020.
In Europe, amid brain initiative funding (€1.9 billion EBRAINS), this bolsters case for sustained investment, projecting 50,000 neuroscience jobs by 2030.