Glass pattern responses in macaque V2 neurons

Matthew A. Smith, Adam Kohn, J. Anthony MovshonJournal of Vision 2007 · DOI

TL;DR

This study investigates how neurons in macaque V2, a visual processing area, respond to "Glass patterns" – stimuli that create a global form perception from sparse local cues. The researchers found that V2 neurons process these patterns similarly to V1 neurons, primarily responding to local features within their receptive fields, and are not significantly influenced by global form cues outside their receptive fields. This suggests that the integration of complex global structure in Glass patterns likely occurs in later stages of visual processing.

What Problem Does This Paper Address?

Understanding how the brain constructs a coherent visual perception from the vast amount of sensory information it receives is a fundamental challenge in neuroscience. Visual cortex is organized hierarchically, with early areas like V1 processing basic features and later areas integrating these features into more complex representations. "Glass patterns" are particularly interesting stimuli because they evoke a strong perception of global form (e.g., a circle, a square) from a collection of randomly placed dots, where the global structure is defined by the relationship between pairs of dots rather than the dots themselves.

Previous research had shown that neurons in V1 respond to these patterns in a way that can be predicted by their responses to simpler stimuli like gratings, suggesting V1 primarily processes the local cues. However, it was unclear whether V2, a downstream area known for its larger receptive fields and potential for integrating information, plays a role in processing the global structure of Glass patterns, either within its receptive field or through its surround. This paper aims to clarify V2's contribution to Glass pattern perception, specifically by comparing its responses to V1 and investigating the influence of stimuli outside the classical receptive field (CRF).

Methods

The researchers employed electrophysiology to record the activity of individual neurons in the V1 and V2 areas of anesthetized and paralyzed macaque monkeys.

  1. Stimulus Presentation: They used a variety of visual stimuli, including:
  2. Drifting sinusoidal gratings: Used to characterize basic neuronal properties like orientation and spatial frequency tuning.
  3. Translational Glass patterns: These were the primary stimuli, consisting of dot pairs with a specific orientation and separation, presented within the neuron's CRF.
  4. Concentric and radial Glass patterns: These were used to probe responses to global form cues outside the CRF. These patterns were presented in various positions relative to the CRF.
  5. Random dot stimuli: Used as a control.
  6. Data Analysis:
    • Orientation Tuning: Neuronal responses to gratings and Glass patterns at different orientations were measured to determine preferred orientation and selectivity. A "tuning bias vector" was used to quantify selectivity.
  7. Dot Separation Tuning: For Glass patterns, the effect of varying the distance between dot pairs (r) relative to the grating's spatial period (λ) was analyzed to find the optimal separation for maximal response and tuning.
  8. Surround Suppression: The response to gratings of increasing size was measured to quantify how stimuli outside the CRF affected neuronal activity. A "suppression index" (SI) was calculated.
  9. Response Modulation: The difference between the maximum and minimum response in an orientation tuning curve was used to quantify the strength of the response.
  10. Modeling: A quasi-linear model, based on Gabor filters, was used to predict V1 and V2 responses to translational Glass patterns. This model assumes that the response is a linear combination of responses to local features.

Datasets

The study did not introduce new datasets or benchmarks. Instead, it used well-established visual stimuli (gratings, Glass patterns) to record from neurons in macaque V1 and V2. The data collected from these recordings served as the "dataset" for their analyses.

Key Insights and Ideas

  • V2 as a "Linear Extender" of V1 for Glass Patterns: The most striking finding is the remarkable similarity between V1 and V2 responses to translational Glass patterns within the CRF. This suggests that V2, at this stage, doesn't add significant new computations for processing the local cues of these patterns, but rather extends the V1-like processing.
  • Surround Suppression is Present but Not Form-Selective: V2 neurons do exhibit surround suppression to gratings, similar to V1. However, when presented with global forms (concentric, radial Glass patterns) outside the CRF, V2 neurons did not show selective responses to these global cues. This implies that the surround mechanisms, while present, are not specifically tuned to detect the global structure of Glass patterns.
  • Global Form Integration Happens Later: The lack of global form sensitivity in V2's surround suggests that the complex integration required to perceive the global structure of Glass patterns likely occurs in higher visual areas (e.g., V4 or IT cortex), which have larger receptive fields and are known to process more complex shapes.
  • The Power of Simple Stimuli: The study reinforces the utility of Glass patterns as a tool for dissecting visual processing, allowing researchers to isolate the contributions of local cues versus global integration.

Key Results

  • Similar Tuning: V2 neurons showed orientation tuning and optimal dot separation for translational Glass patterns that were very similar to V1 neurons. A quasi-linear model accurately predicted these responses.
  • Response Strength Differences: V2 neurons generally showed less response modulation to both gratings and Glass patterns compared to V1 neurons. The ratio of grating response to Glass pattern response was higher in V1.
  • No Global Form Influence from Surround: When presented with concentric or radial Glass patterns outside the CRF, V2 neurons did not show significant selectivity for the global form or position of these stimuli. Responses were largely consistent with a linear model of the CRF.
  • Surround Suppression Magnitude: V2 neurons exhibited significant iso-orientation surround suppression for gratings, with an average suppression index (SI) of 0.38, comparable to V1.

Limitations and Future Work

  • Stimulus Limitations: The authors acknowledge that it's possible the specific Glass pattern stimuli used might not have been sufficient to drive V2 circuits capable of detecting global form outside the CRF.
  • Attention and Higher Areas: They suggest that attention might play a role in neuronal responses to spatial configurations, and that the computation of global form might occur in higher visual areas with larger receptive fields (like V4 or IT cortex).
  • Future Directions: The paper strongly implies that future research should focus on investigating V4 and IT cortex to understand where and how global form integration for Glass patterns truly takes place. Further studies could also explore the role of attention in V2's processing of these stimuli.

Summary updated

Abstract

Area V2 of macaque visual cortex is known to respond well to conventional oriented bar and grating stimuli, but some recent physiological data have shown that it may play an important role in coding more complicated patterns. Most of these data come from testing done with stimuli presented within the classical receptive field (CRF), whereas relatively little attention has been paid to the role played by the extra-classical surround. We have previously shown that neurons in primary visual cortex (V1) respond to translational Glass patterns in a manner that is predictable from their responses to grating stimuli. In this article, we first extend our experiments and modeling of Glass pattern responses in V1 to include V2. We explored the sensitivity of V2 cells to global form cues in Glass patterns confined to the CRF. Our results indicate that V2 neurons respond to the local signals in Glass patterns in a manner similar to V1 and that those responses are not influenced by global form present in the surround. It appears that the coding of the more complicated global structure in Glass patterns takes place further downstream in the visual system.
BibTeX
@article{smith2007glass,
  author = {Matthew A. Smith and Adam Kohn and J. Anthony Movshon},
  title = {{Glass pattern responses in macaque V2 neurons}},
  journal = {Journal of Vision},
  year = {2007},
  doi = {10.1167/7.3.5},
}
Figure