Summary: The brain’s visual system operates both bottom-up and top-down. New research shows that the top-down “return lane” — feedback connections from higher visual areas back to primary visual cortex — is not a generic, pre-set wiring diagram. Instead, these feedback pathways are sculpted by early visual experience and carry content-specific information shaped by the visual statistics an individual encounters.
In an elegant experiment, researchers raised mice wearing miniature goggles that restricted their view to edges of a particular orientation. Using advanced dual-color two-photon imaging, they demonstrated that feedback inputs from higher-order visual regions reorganize according to those experienced orientations. In short, the feedback system becomes a learned bridge between broad scene representations and the detailed, local features those early visual areas encode.
Key Facts
- The “Return Lane”: Feedback connections convey abstract, context-rich information from high-level visual areas back to lower-level regions that detect edges, orientations and simple features.
- Instructive Wiring: The study provides evidence that feedback wiring is instructive — its structure and tuning reflect specific visual experiences rather than forming through a generic developmental program.
- The Goggle Experiment: Mice reared with goggles that biased their visual world toward a narrow set of orientations developed feedback connections whose tuning and spatial organization matched those experienced orientations.
- Dual-Color Imaging: Researchers applied dual-color two-photon imaging to measure the tuning properties and retinotopic layout of feedback inputs from a higher visual area (LM) into layer 1 of primary visual cortex (V1).
- Contextual Integration: These learned feedback pathways enable higher areas to provide context — for example, identifying an object — so lower areas can interpret local features more accurately and efficiently.
Source: Champalimaud Foundation
The visual system is organized hierarchically. Neurons in lower visual areas sample small portions of the visual field and respond to simple features such as edges and their orientation. Higher visual regions represent larger portions of the scene and encode more abstract content such as objects and faces.
Importantly, information flows upward to build these abstract representations and downward via feedback connections to inform local processing. Feedback is thought to supply contextual knowledge — the “big picture” — so that lower areas can interpret local features in a broader scene context. For example, a higher-area neuron that signals the presence of a table can send feedback to lower-area neurons representing a table leg, helping those neurons interpret ambiguous local inputs.
Previous work by Leopoldo Petreanu’s team at the Champalimaud Foundation showed that visual experience is required for normal feedback organization: mice raised in darkness exhibited disrupted feedback connectivity. That finding established that experience matters, but it left open whether the effect was passive (any visual input triggers a default organization) or instructive (different experiences produce different, content-specific wiring).
The new study, published in Current Biology, directly tests that question. The researchers reared separate groups of mice with goggles that biased the animals’ visual input toward a single orientation or toward a different orientation. If feedback wiring were generic, both groups should show similar feedback organization; if instructive, each group should develop distinct feedback patterns matched to their experienced orientations.
Results were clear: the orientation tuning and retinotopic arrangement of LM-to-V1 feedback inputs shifted toward the orientations the mice had experienced. In other words, feedback connections encoded the statistical regularities of each animal’s visual environment. The authors propose that ascending and descending pathways may follow different plasticity rules — a model combining Hebbian changes for V1-to-LM feedforward synapses with anti-Hebbian shaping of feedback weights reproduced the experimental effects.
These findings support an instructive role for visual experience in shaping both the functional properties and spatial organization of feedback inputs. They provide a mechanism by which cortical circuits can learn and encode regularities of the environment so that context-dependent modulation of sensory responses aligns with the statistics of what the animal has actually seen.
Key Questions Answered:
A: The study suggests that early and sustained differences in visual experience can produce measurable differences in neural wiring. If an individual’s visual world were dominated by a particular set of features (for example, mostly vertical lines), feedback circuits would adapt so that the cortical map reflects those experienced features.
A: Feedback provides contextual hypotheses about the scene, which biases local circuits toward interpretations consistent with that context. This improves recognition and speeds accurate processing of ambiguous or noisy local inputs.
A: The brain is most plastic during early development when these instructive connections form, but understanding these mechanisms clarifies how ongoing experience can continue to reshape cortical circuits throughout life.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The full journal paper was reviewed.
- Additional context was added by editorial staff.
About this visual neuroscience research news
Author: Afonso Vaz Pinto
Source: Champalimaud Foundation
Contact: Afonso Vaz Pinto – Champalimaud Foundation
Image: Image credit: Neuroscience News
Original Research: Open access.
“Visual experience exerts an instructive role on cortical feedback inputs to the primary visual cortex” by Radhika Rajan, Rodrigo F. Dias, Nikos Malakasis, Margarida Baeta, Xinyun Zhang, Julijana Gjorgjieva, and Leopoldo Petreanu. Current Biology
DOI: 10.1016/j.cub.2026.01.031
Abstract
Visual experience exerts an instructive role on cortical feedback inputs to the primary visual cortex
Contextual modulation in primary visual cortex (V1) depends on feedback (FB) connections from higher-order cortical areas. These FB projections can reflect the statistical regularities of natural images through their connectivity and tuning, enabling context-dependent modulation of cortical activity.
To test whether FB properties reflect experienced visual statistics, the authors used dual-color two-photon imaging to record tuning and retinotopic organization of inputs from the lateromedial (LM) area into layer 1 of V1. Mice reared with orientation-restricting goggles showed biased orientation tuning and receptive field axes in LM inputs toward the experienced orientations. The retinotopic specificity of FB inputs differed across groups exposed to different orientations.
A computational model combining Hebbian plasticity for feedforward connections and anti-Hebbian rules for feedback reproduced the effects of goggle rearing. These results support an instructive role for visual experience in shaping both the functional properties and organization of cortical feedback, and they suggest distinct plasticity rules across ascending and descending cortical pathways.
Overall, the findings identify circuit mechanisms by which cortical networks can learn and encode the statistical structure of the environment, allowing context-sensitive processing that mirrors an observer’s visual history.