Abstract
Spatial attention influences representations in visual cortical areas as well as perception. This study investigated how top-down signals from the frontal eye field (FEF) modulate the contrast response function (CRF) throughout the visual system. Using fMRI in awake, fixating macaques, researchers measured responses to stimuli of varying luminance contrast while simultaneously microstimulating subregions of the FEF. The results indicate that FEF microstimulation produces a non-proportional scaling of the CRF, consistent with a contrast gain model: low-contrast stimuli are enhanced, while high-contrast stimuli are unaffected or suppressed.
Introduction
A central question in visual neuroscience is how spatial attention—the ability to focus on specific locations—alters the sensitivity of visual neurons. Two primary models have been proposed: response gain, where attention acts as a proportional multiplier across all contrast levels, and contrast gain, where attention shifts the system's sensitivity, providing the greatest boost to faint, low-contrast stimuli. While single-cell studies have yielded conflicting results, this study utilizes functional MRI (fMRI) to examine these effects at a systems level across the macaque visual cortex.
Methods
The researchers employed a simultaneous microstimulation and imaging paradigm in two awake rhesus monkeys. Sub-threshold electrical microstimulation (FEF-EM) was applied to the Frontal Eye Field at 50% of the threshold required to evoke saccades, mimicking covert spatial attention.
Visual stimuli consisted of gratings presented at five contrast levels (0%, 3%, 6%, 12%, and 50%). The FEF was stimulated in subregions whose movement fields overlapped with the visual stimulus locations. To achieve high spatial resolution and sensitivity, the team used fMRI with a MION (monocrystalline iron oxide nanoparticle) contrast agent, which measures changes in cerebral blood volume (CBV).
Results
The primary finding was that FEF microstimulation modulated the CRF in a non-linear fashion. In early visual areas like V1, stimulation significantly enhanced the response to low-contrast (3%) stimuli, often doubling the activation. However, as contrast increased, this enhancement diminished. At the highest contrast level (50%), the response was actually suppressed by approximately 15%.
This pattern of low-contrast enhancement and high-contrast suppression was remarkably consistent across the 12 visual areas mapped, including V1, V2, V3, V4, and MT. The data for the visual-only conditions closely followed the standard Naka-Rushton equation, but the FEF-stimulation data required additional parameters to account for the observed baseline shifts and contrast-dependent scaling.
Discussion
The observed crossover from enhancement to suppression strongly supports a contrast gain model of attention. Rather than simply "turning up the volume" on all visual inputs, top-down signals from the FEF appear to reallocate neural resources.
The authors propose a center-surround schematic to explain these results. At low contrast, FEF input enhances the "center" representation of the stimulus. As contrast increases, the system may engage inhibitory surround mechanisms to sharpen the population's tuning and improve discriminability, leading to the observed suppression at high contrast levels. This suggests that the brain's attentional system is optimized for detecting faint signals while maintaining high-fidelity representations of salient objects.
Conclusion
By combining microstimulation with whole-brain imaging, this study provides a comprehensive view of how spatial attention modulates visual processing. The evidence for contrast gain across the visual hierarchy suggests a sophisticated mechanism for dynamic sensitivity adjustment, prioritizing the enhancement of weak visual signals to guide behavior.
Summary updated