Late development of contrast normalization mechanisms

Francesca Pei, Stefano Baldassi and Anthony Norcia

Contrast sensitivity is regulated by gain control mechanisms that normalize sensory responses. We studied gain control mechanisms in school-age children (6-16 years) and adults using a visual masking paradigm in which a spatially random 2-D noise test pattern was masked by the presence of a superimposed independent noise pattern presented at 0, 12 and 40% contrast. Frequency-tagged Steady State Visual Evoked Potentials were used to separately record responses to the test (5.14Hz) and the mask (7.2 Hz). By incrementally increasing the test contrast we measured contrast response functions for each mask contrast. The unmasked contrast response functions were largely similar in shape but peak amplitude was higher in the children. Masking shifted the adult contrast response function rightward on the contrast axis, consistent with contrast gain control. However, masking impacted children’s response functions differently. The main effect of the masker was to reduce response slope, rather than to shift the response function rightward, suggesting that children exhibit more response gain than contrast gain. These results demonstrate that contrast normalization processes are still immature in school-age children.

Representation of category-specific and non-specific information in human inferior temporal cortex

Bingbing Guo and Ming Meng

A classical hierarchical model would assume that visual processing progresses from simple features to complex objects. Indeed, recent neuroimaging and neurophysiology evidence suggests that areas in inferior temporal (IT) cortex are selective to a certain semantic category (e.g., FFA, fusiform face area; PPA, parahippocampal place area). However, low-level stimulus features modulate FFA activation (e.g., Yue et al, 2011, Cerebral Cortex, 21(1), 35-47). Relationship between feature processing and categorical object representation in these visual areas remains unclear. Here we examine both average BOLD and multivariate patterns of fMRI data regarding how category-specific (face versus house) and non-specific (stimulus contrast) information is represented in human IT. The average BOLD in the FFA was modulated by the contrast of faces, but not houses, while the PPA was modulated by contrast of both stimuli. Interestingly, activation patterns in the FFA and PPA were barely affected by contrast at statistically near chance level, whereas the categorical information can be decoded from the patterns at significantly above chance level. Our results affirm that both category-specific and non-specific information is represented in IT. However, stimulus contrast and object category are encoded in qualitatively disassociated ways, suggesting non-hierarchical and separate neural mechanisms underlying feature processing and object representation.

Masking asymmetries reveal a new specificity for face coding

Stefano Baldassi, Francesca Pei and Anthony Norcia

There is wide agreement that faces are efficiently processed by specialized, independent pathways that treat them as global configurations of particular relevance. Here we studied whether face-specific effects occur as early as the level of visual detection and contrast masking. Stimuli were faces or cars (objects), with both regular and inverted contrast polarity, and noise stimuli obtained by scrambling the phases of each specific image, so that faces/cars and their respective noise images had identical spectra. Contrast discrimination functions were the same for objects masked by noise and for inverted polarity faces masked by noise. There was an asymmetry of masking for positive contrast faces: positive contrast faces elevated noise-contrast discrimination thresholds more than noise masker were able to elevate face-contrast discrimination thresholds, despite the identical spectral content of the two stimuli. These data reveal a novel form of masking asymmetry that is not simply attributable to phase metrics such as phase congruency, and may be linked to efficient coding of visual structures of particular relevance, such as faces.

Contrast thresholds for letter identification as a function of size

Andrew Watson and Albert Ahumada

As part of our effort to develop practical models for image identification [Watson and Ahumada, 2005, Journal of Vision, 5(9), 717-740; Watson and Ahumada, 2008, Journal of Vision, 8(4), 1-19], we have reviewed the data for contrast thresholds for letter identification as a function of size. All eight studies that we have found, examining letter sizes from 0.04 to 60 degrees, and employing a variety of stimulus materials and conditions, agree that sensitivity increases rapidly with size from the acuity limit, but then saturates, and may decline slightly, for letter sizes above about 1 degree. None of these studies has offered a comprehensive account for this pattern of results. We have explored the performance of an ideal observer limited by optics, eccentricity-dependent retinal filtering and sampling, neural noise, and an efficiency that varies with pattern size. We find that all of these components are required to account for the pattern of results. We consider various explanations for the decline in efficiency with size, and discuss alternative explanations, such as a decline in efficiency with eccentricity.

Long-range template matching follows spatial inhomogeneity, short-range spatial filtering, square-law contrast transduction and the addition of internal noise

Tim Meese and Robert Summers

Since the 1970s and 1980s, the standard view of spatial vision has been that contrast detection involves probability summation over multiple short-range filter-elements, typical of those found in the primary visual cortex. However, more recent work has favoured a model involving physiological integration of luminance contrast over substantial parts of the image, concluding that in previous studies, the effects of a square-law transducer followed by linear summation of noise with the signal have combined to masquerade as probability summation. Here we addressed the details of the contrast integration process by comparing it to an ideal observer model, which was progressively adjusted until it fitted our area summation results (1 to 32 cycles square). We argue that observers use spatially extensive templates (with diameters of several cycles) that operate on the luminance contrast image, but only after retinal inhomogeneity, short-range filtering, square-law transduction and the addition of internal noise. Ideal templates that are matched to the image after the effects of the three processing stages above, provide the modeler with a convenient implementation but might imply a greater level of sophistication than is needed to explain the results.

Contrast Energy Discrimination Thresholds for Blur and Contrast Compared

Michael Morgan

Watson & Ahumada (Journal of Vision, 2011) predicted blur discrimination thresholds from contrast energy differences weighted by the appropriate human contrast sensitivity function. Many aspects of previously published results were accurately predicted but thresholds for blur tended to be lower than predicted. To compare blur discrimination and contrast discrimination thresholds with the same stimuli, we used a standard chess board stimulus with contrast 0.5 and blur 5 arcmin, and also measured the ability of observers to discriminate changes in blur and contrast using a 4-key response. Contrast Energy Thresholds were higher (~ 1.5 x) for contrast than for blur discrimination, and blur and contrast changes could be discriminated at their respective energy thresholds in a manner consistent with two channels, inheriting the same contrast energy noise, but with independent late noise.