TL;DR
Our visual perception remains remarkably stable despite constant eye movements. This study reveals that while early visual areas are highly sensitive to saccade-induced visual shifts, higher-level visual regions actively suppress these signals, contributing to a stable perceptual experience. This suppression mechanism appears to be shared with the suppression of eye blink-related visual disruptions.
What Problem Does This Paper Address?
We experience a stable visual world, yet our eyes are constantly moving, primarily through rapid "saccades." This creates a significant mismatch between the stable perception and the rapidly shifting retinal image. How the brain reconciles this discrepancy, particularly how it distinguishes between self-generated eye movements and external visual changes, is a fundamental question in neuroscience. Previous work showed that eye blinks, which also disrupt vision, are suppressed in higher visual areas. This paper investigates whether a similar suppression mechanism exists for saccades and how it unfolds across the visual hierarchy.
Methods
The researchers used intracranial electroencephalography (iEEG) from 12 neurosurgical patients who were undergoing monitoring for clinical reasons. This technique provides high spatial and temporal resolution of brain activity. They presented participants with visual stimuli and recorded their brain responses while simultaneously tracking their eye movements with a high-precision eye tracker.
The core of their method involved comparing the brain's high-frequency broadband (HFB) activity – a proxy for neural firing rate – in response to two conditions:
- External Stimulus Displacements: The visual image on the screen was shifted.
- Saccades: Participants naturally made small saccades (1.3° to 3.7°) while viewing the stimuli.
They used a Finite Impulse Response (FIR) General Linear Model to deconvolve and estimate the unique contribution of each event (saccade or displacement) to the HFB signal. Statistical analyses, including random permutation tests, were employed to determine significant differences in brain activity between the two conditions across various visual regions of interest.
Datasets
The study utilized data from 12 neurosurgical patients who underwent iEEG monitoring. These patients viewed still images of faces and non-face control stimuli. The data from these patients were collected as part of experimental sessions previously reported in Golan et al. (2016), with some exclusions and one additional subject.
Key Insights and Ideas
- Early Visual Cortex is "Unstable": Unlike higher visual areas, early visual areas (V1-V3) show similar excitatory responses to both external visual shifts and saccade-induced shifts. This suggests that these early regions are highly sensitive to retinal input, regardless of its origin.
- Higher Visual Areas Actively Suppress Saccades: The most striking finding is that in higher-level visual regions (V4, VO, and face-selective areas), the brain significantly suppresses the neural response to saccades, while still responding robustly to external displacements. This indicates an active, extra-retinal mechanism at play.
- Hierarchical Suppression: The suppression of saccade-related activity is not uniform; it increases as you move up the visual hierarchy. This suggests a sophisticated processing stream that filters out self-generated visual noise.
- Shared Suppression Mechanism: The study found a strong overlap between the suppression of saccade-related transients and the previously reported suppression of eye blink-related transients. This hints at a shared or overlapping neural pathway for filtering out visual disruptions caused by different motor commands.
- Ventral Stream Involvement: The suppression effect extends to the ventral visual stream, including face-selective areas, challenging the notion that these areas are purely invariant to position and motion.
Key Results
- In early visual areas (V1-V3), both external displacements and saccades elicited similar increases in HFB activity.
- In higher-order visual areas (V4, VO, face-selective, and non-face-selective regions), external displacements triggered a significant HFB increase, whereas saccades elicited almost no HFB response modulation.
- Statistical analysis revealed that 36 out of 115 electrodes showed significantly greater HFB contribution from external displacements compared to saccades, with a higher prevalence in more anterior visual regions.
- No electrodes showed a significantly greater HFB increase for saccades compared to external displacements.
- The suppression of saccade-related transients showed a strong correlation with the suppression of blink-related transients, suggesting a common underlying mechanism.
Limitations and Future Work
- Saccade Magnitude: The study focused on relatively small saccades (1.3°-3.7°). Larger saccades, which bring new objects into the fovea, might elicit different responses. Future work could explore the effect of larger saccades.
- Retinal vs. Extra-retinal Contributions: While the findings strongly suggest an extra-retinal suppression mechanism, the exact nature of this signal and its precise contribution to early visual cortex activity remain somewhat unclear. The authors suggest that future studies using more sophisticated methods to simulate the retinal impact of saccades (e.g., rotating mirrors) are needed to fully disentangle these effects.
- Ventral Stream Mechanisms: The precise mechanism underlying saccade suppression in the ventral stream is still open to investigation. It could be a general gating mechanism or a more specific offsetting process. Future experiments are needed to differentiate these possibilities.
- HFB as a Proxy: The study relies on HFB as a proxy for neural firing rate. While a good correlate, it's possible that single-unit activity or slower iEEG frequencies might reveal different patterns.
Summary updated