Temporal dynamics of the neural representation of hue and luminance polarity

Abstract

Hue and luminance contrast are the most basic visual computations, and are reflected in the earliest layers of convolutional neural networks, yet the extent to which they are extracted by the same or separate circuits in the brain, and the timing of these neural computations, is unknown. Here we answer these questions using multivariate analyses of human brain responses measured with magnetoencephalography. We report three discoveries. First, hue and luminance contrast could be decoded independently, indicating these computations are somewhat separable. Hue was computed about 15-24ms after luminance contrast. Second, representations of hue showed relatively greater generalization across time and were more sustained, providing the first neural correlate of the perceptual preeminence of hue over luminance contrast in grouping objects. Finally, luminance contrast could be decoded less well for hues associated with daylight (orange and blue), suggesting that colorconstancy mechanisms are adapted to natural lighting.

Uncovering the geometry of color space with magnetoencephalography (MEG)

“…The color percept associated with a stimulus can be decoded from functional magnetic resonance imaging activity [10], MEG and EEG in humans and monkeys [11][12][13][14], or responses of single neurons in monkeys [15,16]. Moreover, as many of these studies show, the extent to which activity patterns reflect the sequence of hues in the color wheel can be used to identify neurons and brain areas that are likely involved in encoding color.…”

Decoding visual colour from scalp electroencephalography measurements

“…For orientation, the cardinal orientations (0, 90, and 180 degrees) showed higher decoding than the off-cardinal decoding. Interestingly, due to some nonlinearities in the circular space, we observe some colours (pink, red, orange) in one part of the representational space and other colours (purple, blue, green) on the other (see also Hermann et al 2020).…”

“…We investigated the ability to decode visual colours from scalp EEG measurements. We built on other recent studies that have employed colour decoding in scalp EEG (Bocincova and Johnson 2018; Sandhaeger et al 2019) and MEG (Hermann et al 2020;Rosenthal et al in press. ;Sandhaeger et al 2019;Teichmann et al 2019Teichmann et al , 2020 and extended these in several ways.…”

“…Therefore, finding that colour-related decoding follows parametric variations in stimuli (as opposed to binary or categorical boundaries) in a manner akin to decoding of stimulus orientation is an important step toward ensuring that decoding can reveal colour-specific neural information (in line with Rosenthal et al in press). Further reassurance would come from the involvement of electrodes sensitive to visual processing to the decoding, at posterior sites posterior sites contralateral to the decoded stimulus (akin to recent MEG research; Hermann et al 2020). Finally, the ability to decode the colours of multiple stimuli presented concurrently (as opposed to a single central stimulus in isolation, as in most prior colour decoding studies) would greatly add to the flexibility and analytical power of colour decoding as a tool in future studies.…”

“…The demonstration of colour decoding from scalp EEG should therefore have important implications -affording novel ways to track perceptual representations in various contexts. To date, a number of MEG (Hermann et al 2020;Rosenthal et al in press;Sandhaeger et al 2019;Teichmann et al 2019Teichmann et al , 2020) and, to a lesser extent, EEG studies (Bocincova and Johnson 2018; Sandhaeger et al 2019) have applied decoding approaches to distinguish among colour values upon visual presentation. We have built on these initial studies to further establish and characterise the properties of colour decoding from scalp EEG measurements.…”