Abstract
Neuronal activity in early visual cortex depends on attention shifts but the contribution to working memory has remained unclear. Here, we examine neuronal activity in the different layers of the primary visual cortex (V1) in an attention-demanding and a working memory task. A current-source density analysis reveales top-down inputs in the superficial layers and layer 5, and an increase in neuronal firing rates most pronounced in the superficial and deep layers and weaker in input layer 4. This increased activity is strongest in the attention task but it is also highly reliable during working memory delays. A visual mask erases the V1 memory activity, but it reappeares at a later point in time. These results provide new insights in the laminar circuits involved in the top-down modulation of activity in early visual cortex in the presence and absence of visual stimuli.
Introduction
The primary visual cortex (V1) has traditionally been viewed as a passive relay for sensory information. However, the "sensory recruitment" theory suggests that early visual areas are also involved in maintaining information during working memory. While it is well-established that selective attention modulates V1 activity, whether V1 maintains robust signals during memory delays—when no stimulus is present—has been a subject of intense debate. This study investigates the laminar microcircuits of V1 to determine how top-down signals from higher brain areas influence sensory processing and memory.
The Curve-Tracing Task
To investigate these mechanisms, researchers trained rhesus macaques to perform a "mental tracing" task. In the attention condition, monkeys traced a visible curve to a target. In the working memory condition, the curves were briefly presented and then removed, forcing the monkeys to maintain a representation of the path in their working memory to successfully complete the trial.
Neural activity was recorded using high-density laminar probes (U-probes), which allowed for the simultaneous measurement of multi-unit activity (MUA) and current-source density (CSD) across all six layers of the cortex. This approach is critical for identifying the direction of information flow, distinguishing between bottom-up sensory input and top-down feedback.
Laminar Signatures of Memory and Attention
The researchers discovered a distinct laminar profile for top-down modulation. Unlike sensory signals that primarily enter through Layer 4, feedback signals for both attention and working memory targeted the superficial layers (1–3) and Layer 5.
Key Findings
- Layer-Specific Modulation: Activity increases were most pronounced in the superficial and deep layers, while the input Layer 4 showed significantly weaker modulation.
- Memory Decoding: Spiking activity in V1 was sufficient to decode which curve the monkey was remembering with high accuracy (up to 78% in one subject).
- Resilience to Interference: In experiments using a visual mask (a flickering checkerboard), V1 memory activity was temporarily abolished but reappeared shortly after the mask was removed. This suggests that the memory is not a simple "echo" in V1 but is actively maintained and re-instated by higher-order cortical areas.
- Behavioral Correlation: The strength of the V1 modulation correlated with the monkeys' performance, with weaker signals observed during error trials.
Discussion and Future Directions
These findings provide strong evidence that V1 is an active participant in working memory circuits. The shared laminar signature between attention and memory suggests that top-down feedback utilizes a canonical circuit to modulate early sensory processing.
However, several questions remain. While V1 activity correlates with memory, its causal necessity has yet to be proven through inactivation studies. Furthermore, while areas like V4, LIP, or FEF are likely sources of this feedback, simultaneous multi-area recordings are needed to map the full circuit. Understanding these layer-specific interactions is a crucial step toward a complete model of how the brain integrates perception and memory.
Summary updated