The selectivity of responses to red-green colour and achromatic contrast in the human visual cortex: an fMRI adaptation study

Increasing spatial frequency of S-cone defined gratings reduces their visibility and brain response more than for gratings defined by L-M cone contrast

Chromatic sensitivity reduces as spatial frequency increases. Here, we explore the behavioural and neuronal responses to chromatic stimuli at two spatial frequencies for which the difference in sensitivity will be greater for S-cone than L-M stimuli. Luminance artefacts were removed using the Random Luminance Modulation (RLM) technique. As expected, doubling the spatial frequency increased the detection threshold more for S-cone than for isoluminant L-M gratings. We then used fMRI to measure the cortical BOLD responses to the same two chromatic stimuli (S and L-M) at the same two spatial frequencies. Responses were measured in six visual areas (V1, V2, V3, V3a, hV4, TO1/2). We found a significant interaction between spatial frequency in V1, V2 and V4 suggesting that the behaviourally observed increase in contrast threshold for high spatial frequency S-cone stimuli is reflected in these retinotopic areas. Our measurements show that neural responses consistent with psychophysical behaviour in a colour detection task can be observed as early as primary visual cortex.

Visual mode switching: Improved general compensation for environmental color changes requires only one exposure per day

When the visual environment changes, vision adapts in order to maintain accurate perception. For repeatedly encountered environmental changes, the visual system may learn to adjust immediately, a process called "visual mode switching." For example, following experience with red glasses, participants report that the glasses' redness fades instantly when they put the glasses on. Here we tested (1) whether once-daily experience suffices for learning to switch visual modes and (2) whether effects of mode switching apply to most stimuli affected by the environmental change. In Experiment 1, 12 participants wore bright red glasses for a single 5-hr period each day for 5 days, and we tested for changes in the perception of unique yellow, which contains neither red nor green. In Experiment 2, we tested how mode switching affects larger parts of the color space. Thirteen participants donned and removed the glasses multiple times a day for 5 days, and we used a dissimilarity rating task to measure and track perception of many different colors. Across days, immediately upon donning the glasses, the world appeared less and less reddish (Experiment 1), and colors across the whole color space appeared more and more normal (Experiment 2). These results indicate that mode switching can be acquired from a once-daily experience, and it applies to most stimuli in a given environment. These findings may help to predict when and how mode switching occurs outside the laboratory.

Magnetoencephalography contrast adaptation reflects perceptual adaptation

Contrast adaptation is a fundamental visual process that has been extensively investigated and used to infer the selectivity of visual cortex. We recently reported an apparent disconnect between the effects of contrast adaptation on perception and functional magnetic resonance imaging BOLD response adaptation, in which adaptation between chromatic and achromatic stimuli measured psychophysically showed greater selectivity than adaptation measured using BOLD signals. Here we used magnetoencephalography (MEG) recordings of neural responses to the same chromatic and achromatic adaptation conditions to characterize the neural effects of contrast adaptation and to determine whether BOLD adaptation or MEG better reflect the measured perceptual effects. Participants viewed achromatic, L-M isolating, or S-cone isolating radial sinusoids before adaptation and after adaptation to each of the three contrast directions. We measured adaptation-related changes in the neural response to a range of stimulus contrast amplitudes using two measures of the MEG response: the overall response amplitude, and a novel time-resolved measure of the contrast response function, derived from a classification analysis combined with multidimensional scaling. Within-stimulus adaptation effects on the contrast response functions in each case showed a pattern of contrast-gain or a combination of contrast-gain and response-gain effects. Cross-stimulus adaptation conditions showed that adaptation effects were highly stimulus selective across early, ventral, and dorsal visual cortical areas, consistent with the perceptual effects.

Structural Differences Across Multiple Visual Cortical Regions in the Absence of Cone Function in Congenital Achromatopsia

Most individuals with congenital achromatopsia (ACHM) carry mutations that affect the retinal phototransduction pathway of cone photoreceptors, fundamental to both high acuity vision and colour perception. As the central fovea is occupied solely by cones, achromats have an absence of retinal input to the visual cortex and a small central area of blindness. Additionally, those with complete ACHM have no colour perception, and colour processing regions of the ventral cortex also lack typical chromatic signals from the cones. This study examined the cortical morphology (grey matter volume, cortical thickness, and cortical surface area) of multiple visual cortical regions in ACHM (n = 15) compared to normally sighted controls (n = 42) to determine the cortical changes that are associated with the retinal characteristics of ACHM. Surface-based morphometry was applied to T1-weighted MRI in atlas-defined early, ventral and dorsal visual regions of interest. Reduced grey matter volume in V1, V2, V3, and V4 was found in ACHM compared to controls, driven by a reduction in cortical surface area as there was no significant reduction in cortical thickness. Cortical surface area (but not thickness) was reduced in a wide range of areas (V1, V2, V3, TO1, V4, and LO1). Reduction in early visual areas with large foveal representations (V1, V2, and V3) suggests that the lack of foveal input to the visual cortex was a major driving factor in morphological changes in ACHM. However, the significant reduction in ventral area V4 coupled with the lack of difference in dorsal areas V3a and V3b suggest that deprivation of chromatic signals to visual cortex in ACHM may also contribute to changes in cortical morphology. This research shows that the congenital lack of cone input to the visual cortex can lead to widespread structural changes across multiple visual areas.

Attention selectively enhances stimulus information for surround over foveal stimulus representations in occipital cortex

By attending to part of a visual scene, we can prioritize processing of the most relevant visual information and so use our limited resources effectively. Previous functional magnetic resonance imaging (fMRI) work has shown that attention can increase overall blood-oxygen-level-dependent (BOLD) signal responsiveness but also enhances the stimulus information in terms of classifier performance. Here, we investigate how these effects vary across the visual field. We compare attention-enhanced fMRI-BOLD amplitude responses and classifier accuracy in fovea and surrounding stimulus regions using a set of four simple stimuli subdivided into a foveal region (1.4° diameter) and a surround region (15° diameter). We found dissociations between the effects of attention on average response and in enhancing stimulus information. In early visual cortex, we found that attention increased the amplitude of responses to both foveal and surround parts of the stimuli and increased classifier performance only for the surround stimulus. Conversely, ventral visual areas showed less change in average response but greater changes in decoding. Unlike for early visual cortex, in the ventral visual cortex attention produced similar changes in decoding for center and surround stimuli.

Color for object recognition: Hue and chroma sensitivity in the deep features of convolutional neural networks

In this work, we examined the color tuning of units in the hidden layers of AlexNet, VGG-16 and VGG-19 convolutional neural networks and their relevance for the successful recognition of an object. We first selected the patches for which the units are maximally responsive among the 1.2 M images of the ImageNet training dataset. We segmented these patches using a k-means clustering algorithm on their chromatic distribution. Then we independently varied the color of these segments, both in hue and chroma, to measure the unit's chromatic tuning. The models exhibited properties at times similar or opposed to the known chromatic processing of biological system. We found that, similarly to the most anterior occipital visual areas in primates, the last convolutional layer exhibited high color sensitivity. We also found the gradual emergence of single to double opponent kernels. Contrary to cells in the visual system, however, these kernels were selective for hues that gradually transit from being broadly distributed in early layers, to mainly falling along the blue-orange axis in late layers. In addition, we found that the classification performance of our models varies as we change the color of our stimuli following the models' kernels properties. Performance was highest for colors the kernels maximally responded to, and images responsible for the activation of color sensitive kernels were more likely to be mis-classified as we changed their color. These observations were shared by all three networks, thus suggesting that they are general properties of current convolutional neural networks trained for object recognition.

Color Compensation in Anomalous Trichromats Assessed with fMRI

Anomalous trichromacy is a common form of congenital color deficiency resulting from a genetic alteration in the photopigments of the eye's light receptors. The changes reduce sensitivity to reddish and greenish hues, yet previous work suggests that these observers may experience the world to be more colorful than their altered receptor sensitivities would predict, potentially indicating an amplification of post-receptoral signals. However, past evidence suggesting such a gain adjustment rests on subjective measures of color appearance or salience. We directly tested for neural amplification by using fMRI to measure cortical responses in color-anomalous and normal control observers. Color contrast response functions were measured in two experiments with different tasks to control for attentional factors. Both experiments showed a predictable reduction in chromatic responses for anomalous trichromats in primary visual cortex. However, in later areas V2v and V3v, chromatic responses in the two groups were indistinguishable. Our results provide direct evidence for neural plasticity that compensates for the deficiency in the initial receptor color signals and suggest that the site of this compensation is in early visual cortex.

fMRI representational similarity analysis reveals graded preferences for chromatic and achromatic stimulus contrast across human visual cortex

Human visual cortex is partitioned into different functional areas that, from lower to higher, become increasingly selective and responsive to complex feature dimensions. Here we use a Representational Similarity Analysis (RSA) of fMRI-BOLD signals to make quantitative comparisons across LGN and multiple visual areas of the low-level stimulus information encoded in the patterns of voxel responses. Our stimulus set was picked to target the four functionally distinct subcortical channels that input visual cortex from the LGN: two achromatic sinewave stimuli that favor the responses of the high-temporal magnocellular and high-spatial parvocellular pathways, respectively, and two chromatic stimuli isolating the L/M-cone opponent and S-cone opponent pathways, respectively. Each stimulus type had three spatial extents to sample both foveal and para-central visual field. With the RSA, we compare quantitatively the response specializations for individual stimuli and combinations of stimuli in each area and how these change across visual cortex. First, our results replicate the known response preferences for motion/flicker in the dorsal visual areas. In addition, we identify two distinct gradients along the ventral visual stream. In the early visual areas (V1-V3), the strongest differential representation is for the achromatic high spatial frequency stimuli, suitable for form vision, and a very weak differentiation of chromatic versus achromatic contrast. Emerging in ventral occipital areas (V4, VO1 and VO2), however, is an increasingly strong separation of the responses to chromatic versus achromatic contrast and a decline in the high spatial frequency representation. These gradients provide new insight into how visual information is transformed across the visual cortex.

Hue selectivity of collinear facilitation

Collinear facilitation (CF) is the improvement of the detection sensitivity of the target when two high-contrast flanking stimuli (flankers) have the same visual properties. While it is known that CF does not occur between achromatic flanking stimuli and chromatic targets, or vice versa, it remains unclear whether CF occurs when the hue of the target and flankers are different. We measured CF for Gabor stimuli defined in an isoluminant plane using stimuli defined by isoluminant colors along isolated cone-opponent axes and in two diagonal directions. The measured CF varied with the difference in hue between the target and flankers. Moreover, increased thresholds were also observed. These results suggest that CF exhibits hue selectivity and involves a suppression as well as a facilitation component. The hue selectivity profile of these factors infer that the CF cannot be simply explained by the assumption of two independent cone opponent mechanisms.

What Cognitive Neurology Teaches Us about Our Experience of Color

Color provides valuable information about the environment, yet the exact mechanisms explaining how colors appear to us remain poorly understood. Retinal signals are processed in the visual cortex through high-level mechanisms that link color perception with top-down expectations and knowledge. Here, we review the neuroimaging evidence about color processing in the brain, and how it is affected by acquired brain lesions in humans. Evidence from patients with brain-damage suggests that high-level color processing may be divided into at least three modules: perceptual color experience, color naming, and color knowledge. These modules appear to be functionally independent but richly interconnected, and serve as cortical relays linking sensory and semantic information, with the final goal of directing object-related behavior. We argue that the relations between colors and their objects are key mechanisms to understand high-level color processing.

Color contrast adaptation: fMRI fails to predict behavioral adaptation

fMRI-adaptation is a valuable tool for inferring the selectivity of neural responses. Here we use it in human color vision to test the selectivity of responses to S-cone opponent (blue-yellow), L/M-cone opponent (red-green), and achromatic (Ach) contrast across nine regions of interest in visual cortex. We measure psychophysical adaptation, using comparable stimuli to the fMRI-adaptation, and find significant selective adaptation for all three stimulus types, implying separable visual responses to each. For fMRI-adaptation, we find robust adaptation but, surprisingly, much less selectivity due to high levels of cross-stimulus adaptation in all conditions. For all BY and Ach test/adaptor pairs, selectivity is absent across all ROIs. For RG/Ach stimulus pairs, this paradigm has previously shown selectivity for RG in ventral areas and for Ach in dorsal areas. For chromatic stimulus pairs (RG/BY), we find a trend for selectivity in ventral areas. In conclusion, we find an overall lack of correspondence between BOLD and behavioral adaptation suggesting they reflect different aspects of the underlying neural processes. For example, raised cross-stimulus adaptation in fMRI may reflect adaptation of the broadly-tuned normalization pool. Finally, we also identify a longer-timescale adaptation (1h) in both BOLD and behavioral data. This is greater for chromatic than achromatic contrast. The longer-timescale BOLD effect was more evident in the higher ventral areas than in V1, consistent with increasing windows of temporal integration for higher-order areas.

Processing of chromatic information in a deep convolutional neural network

Deep convolutional neural networks are a class of machine-learning algorithms capable of solving non-trivial tasks, such as object recognition, with human-like performance. Little is known about the exact computations that deep neural networks learn, and to what extent these computations are similar to the ones performed by the primate brain. Here, we investigate how color information is processed in the different layers of the AlexNet deep neural network, originally trained on object classification of over 1.2M images of objects in their natural contexts. We found that the color-responsive units in the first layer of AlexNet learned linear features and were broadly tuned to two directions in color space, analogously to what is known of color responsive cells in the primate thalamus. Moreover, these directions are decorrelated and lead to statistically efficient representations, similar to the cardinal directions of the second-stage color mechanisms in primates. We also found, in analogy to the early stages of the primate visual system, that chromatic and achromatic information were segregated in the early layers of the network. Units in the higher layers of AlexNet exhibit on average a lower responsivity for color than units at earlier stages.

Detailed spatiotemporal brain mapping of chromatic vision combining high-resolution VEP with fMRI and retinotopy

Neuroimaging studies have identified so far, several color-sensitive visual areas in the human brain, and the temporal dynamics of these activities have been separately investigated using the visual-evoked potentials (VEPs). In the present study, we combined electrophysiological and neuroimaging methods to determine a detailed spatiotemporal profile of chromatic VEP and to localize its neural generators. The accuracy of the present co-registration study was obtained by combining standard fMRI data with retinotopic and motion mapping data at the individual level. We found a sequence of occipito activities more complex than that typically reported for chromatic VEPs, including feed-forward and reentrant feedback. Results showed that chromatic human perception arises by the combined activity of at the least five parieto-occipital areas including V1, LOC, V8/VO, and the motion-sensitive dorsal region MT+. However, the contribution of V1 and V8/VO seems dominant because the re-entrant activity in these areas was present more than once (twice in V8/VO and thrice in V1). This feedforward and feedback chromatic processing appears delayed compared with the luminance processing. Associating VEPs and neuroimaging measures, we showed for the first time a complex spatiotemporal pattern of activity, confirming that chromatic stimuli produce intricate interactions of many different brain dorsal and ventral areas.

The spectral, spatial and contrast sensitivity of human polarization pattern perception

It is generally believed that humans perceive linear polarized light following its conversion into a luminance signal by diattenuating macular structures. Measures of polarization sensitivity may therefore allow a targeted assessment of macular function. Our aim here was to quantify psychophysical characteristics of human polarization perception using grating and optotype stimuli defined solely by their state of linear polarization. We show: (i) sensitivity to polarization patterns follows the spectral sensitivity of macular pigment; (ii) the change in sensitivity across the central field follows macular pigment density; (iii) polarization patterns are identifiable across a range of contrasts and scales, and can be resolved with an acuity of 15.4 cycles/degree (0.29 logMAR); and (iv) the human eye can discriminate between areas of linear polarization differing in electric field vector orientation by as little as 4.4°. These findings, which support the macular diattenuator model of polarization sensitivity, are unique for vertebrates and approach those of some invertebrates with a well-developed polarization sense. We conclude that this sensory modality extends beyond Haidinger's brushes to the recognition of quantifiable spatial polarization-modulated patterns. Furthermore, the macular origin and sensitivity of human polarization pattern perception makes it potentially suitable for the detection and quantification of macular dysfunction.

Visual cortical activity before and after cochlear implantation: A follow up ERP prospective study in deaf children

ERPs were recorded in response to presentation of static colored patterned stimuli in 25 children (19 to 80months of age at cochlear implantation, CI) with very early prelingual profound deafness (PreLD), 21 postlingual profoundly deaf children (PostLD) (34 to 180months of age at CI) and gender- and age-matched control hearing children. Recording sessions were performed before CI, then 6 and 24months after CI. Results showed that prelingual and, at a lesser degree, postlingual auditory deprivation altered cortical visual neural activity associated to colored shapes from both P1 and N1 cortical processing stages. The P1 and N1 amplitude modifications vanished about 24months after CI in both PreLD and PostLD deaf children. In PreLD the visual processing pattern becomes similar to the typical one essentially by an amplitude decrease of P1 on the left hemisphere together with an amplitude increase of the N1 on the right hemisphere. Finally, in PreLD, increased LH advantage over the RH in N1 amplitude on the cerebellar-occipito-parietal region before CI showed a significant inverse relationship with speech perception outcomes 3years after CI. Investigating early visual processing development and its neural substrates in deaf children would help to understand the variability of CI outcome, because their cortical visual organization diverged from the one of typically developing hearing children, and cannot be predicted from what is observed in deaf adults.

Color responses and their adaptation in human superior colliculus and lateral geniculate nucleus

We use an fMRI adaptation paradigm to explore the selectivity of human responses in the lateral geniculate nucleus (LGN) and superior colliculus (SC) to red-green color and achromatic contrast. We measured responses to red-green (RG) and achromatic (ACH) high contrast sinewave counter-phasing rings with and without adaptation, within a block design. The signal for the RG test stimulus was reduced following both RG and ACH adaptation, whereas the signal for the ACH test was unaffected by either adaptor. These results provide compelling evidence that the human LGN and SC have significant capacity for color adaptation. Since in the LGN red-green responses are mediated by P cells, these findings are in contrast to earlier neurophysiological data from non-human primates that have shown weak or no contrast adaptation in the P pathway. Cross-adaptation of the red-green color response by achromatic contrast suggests unselective response adaptation and points to a dual role for P cells in responding to both color and achromatic contrast. We further show that subcortical adaptation is not restricted to the geniculostriate system, but is also present in the superior colliculus (SC), an oculomotor region that until recently, has been thought to be color-blind. Our data show that the human SC not only responds to red-green color contrast, but like the LGN, shows reliable but unselective adaptation.