Prioritization of emotional faces is not driven by emotional content.

Emotional faces have prioritized access to visual awareness. However, studies concerned with what expressions are prioritized most are inconsistent and the source of prioritization remains elusive. Here we tested the predictive value of spatial frequency-based image-features and emotional content, the sub-part of the image content that signals the emotional expression of the actor in the image as opposed to the image content irrelevant for the emotional expression, for prioritization for awareness. Participants reported which of two faces (displaying a combination of angry, happy, and neutral expressions), that were temporarily suppressed from awareness, was perceived first. Even though the results show that happy expressions were prioritized for awareness, this prioritization was driven by the contrast energy of the images. In fact, emotional content could not predict prioritization at all. Our findings show that the source of prioritization for awareness is not the information carrying the emotional content. We argue that the methods used here, or similar approaches, should become standard practice to break the chain of inconsistent findings regarding emotional superiority effects that have been part of the field for decades.

Forecasting transitions in the state of food security with machine learning using transferable features.

Food insecurity is a growing concern due to man-made conflicts, climate change, and economic downturns. Forecasting the state of food insecurity is essential to be able to trigger early actions, for example, by humanitarian actors. To measure the actual state of food insecurity, expert and consensus-based approaches and surveys are currently used. Both require substantial manpower, time, and budget. This paper introduces an extreme gradient-boosting machine learning model to forecast monthly transitions in the state of food security in Ethiopia, at a spatial granularity of livelihood zones, and for lead times of one to 12 months, using open-source data. The transition in the state of food security, hereafter referred to as predictand, is represented by the Integrated Food Security Phase Classification Data. From 19 categories of datasets, 130 variables were derived and used as predictors of the transition in the state of food security. The predictors represent changes in climate and land, market, conflict, infrastructure, demographics and livelihood zone characteristics. The most relevant predictors are found to be food security history and surface soil moisture. Overall, the model performs best for forecasting Deteriorations and Improvements in the state of food security compared to the baselines. The proposed method performs (F1 macro score) at least twice as well as the best baseline (a dummy classifier) for a Deterioration. The model performs better when forecasting long-term (7 months; F1 macro average = 0.61) compared to short-term (3 months; F1 macro average = 0.51). Combining machine learning, Integrated Phase Classification (IPC) ratings from monitoring systems, and open data can add value to existing consensus-based forecasting approaches as this combination provides longer lead times and more regular updates. Our approach can also be transferred to other countries as most of the data on the predictors are openly available from global data repositories.

Grouping of optic flow stimuli during binocular rivalry is driven by monocular information.

During binocular rivalry, perception alternates between two dissimilar images, presented dichoptically. Although binocular rivalry is thought to result from competition at a local level, neighboring image parts with similar features tend to be perceived together for longer durations than image parts with dissimilar features. This simultaneous dominance of two image parts is called grouping during rivalry. Previous studies have shown that this grouping depends on a shared eye-of-origin to a much larger extent than on image content, irrespective of the complexity of a static image. In the current study, we examine whether grouping of dynamic optic flow patterns is also primarily driven by monocular (eye-of-origin) information. In addition, we examine whether image parameters, such as optic flow direction, and partial versus full visibility of the optic flow pattern, affect grouping durations during rivalry. The results show that grouping of optic flow is, as is known for static images, primarily affected by its eye-of-origin. Furthermore, global motion can affect grouping durations, but only under specific conditions. Namely, only when the two full optic flow patterns were presented locally. These results suggest that grouping during rivalry is primarily driven by monocular information even for motion stimuli thought to rely on higher-level motion areas.

Visual directional anisotropy does not mirror the directional anisotropy apparent in postural sway.

Presenting a large optic flow pattern to observers is likely to cause postural sway. However, directional anisotropies have been reported, in that contracting optic flow induces more postural sway than expanding optic flow. Recently, we showed that the biomechanics of the lower leg cannot account for this anisotropy (Holten, Donker, Verstraten, & van der Smagt, 2013, Experimental Brain Research, 228, 117-129). The question we address in the current study is whether differences in visual processing of optic flow directions, in particular the perceptual strength of these directions, mirrors the anisotropy apparent in postural sway. That is, can contracting optic flow be considered to be a perceptually stronger visual stimulus than expanding optic flow? In the current study we use a breaking continuous flash suppression paradigm where we assume that perceptually stronger visual stimuli will break the flash suppression earlier, making the suppressed optic flow stimulus visible sooner. Surprisingly, our results show the opposite, in that expanding optic flow is detected earlier than contracting optic flow.

Image-based grouping during binocular rivalry is dictated by eye-of-origin.

Prolonged viewing of dichoptically presented images with different content results in perceptual alternations known as binocular rivalry. This phenomenon is thought to be the result of competition at a local level, where local rivalry zones interact to give rise to a single, global dominant percept. Certain perceived combinations that result from this local competition are known to last longer than others, which is referred to as grouping during binocular rivalry. In recent years, the phenomenon has been suggested to be the result of competition at both eye- and image-based processing levels, although the exact contribution from each level remains elusive. Here we use a paradigm designed specifically to quantify the contribution of eye- and image-based processing to grouping during rivalry. In this paradigm we used sine-wave gratings as well as upright and inverted faces, with and without binocular disparity-based occlusion. These stimuli and conditions were used because they are known to result in processing at different stages throughout the visual processing hierarchy. Specifically, more complex images were included in order to maximize the potential contribution of image-based grouping. In spite of this, our results show that increasing image complexity did not lead to an increase in the contribution of image-based processing to grouping during rivalry. In fact, the results show that grouping was primarily affected by the eye-of-origin of the image parts, irrespective of stimulus type. We suggest that image content affects grouping during binocular rivalry at low-level processing stages, where it is intertwined with eye-of-origin information.

What is Grouping during Binocular Rivalry?

During binocular rivalry, perception alternates between dissimilar images presented dichoptically. Although perception during rivalry is believed to originate from competition at a local level, different rivalry zones are not independent: rival targets that are spaced apart but have similar features tend to be dominant at the same time. We investigated grouping of spatially separated rival targets presented to the same or to different eyes and presented in the same or in different hemifields. We found eye-of-origin to be the strongest cue for grouping during binocular rivalry. Grouping was additionally affected by orientation: identical orientations were grouped longer than dissimilar orientations, even when presented to different eyes. Our results suggest that eye-based and orientation-based grouping is independent and additive in nature. Grouping effects were further modulated by the distribution of the targets across the visual field. That is, grouping within the same hemifield can be stronger or weaker than between hemifields, depending on the eye-of-origin of the grouped targets. We also quantified the contribution of the previous cues to grouping of two images during binocular rivalry. These quantifications can be successfully used to predict the dominance durations of different studies. Incorporating the relative contribution of different cues to grouping, and the dependency on hemifield, into future models of binocular rivalry will prove useful in our understanding of the functional and anatomical basis of the phenomenon of binocular rivalry.

Suppressed images selectively affect the dominant percept during binocular rivalry.

During binocular rivalry, perception alternates between dissimilar images that are presented dichoptically. It has been argued that perception during the dominance phase of rivalry is unaffected by the suppressed image. Recent evidence suggests, however, that the suppressed image does affect perception of the dominant image, yet the extent and nature of this interaction remain elusive. We hypothesize that this interaction depends on the difference in feature content between the rivaling images. Here, we investigate how sensitivity to probes presented in the image that is currently dominant in perception is affected by the suppressed image. Observers performed a 2AFC discrimination task on oriented probes (Experiment 1) or probes with different motion directions (Experiment 2). Our results show that performance on both orientation and motion direction discrimination was affected by the content of the suppressed image. The strength of interference depended specifically on the difference in feature content (e.g., the difference in orientation) between the probe and the suppressed image. Moreover, the pattern of interference by the suppressed image is qualitatively similar to the situation where this image and the probe are simultaneously visible. We conclude that perception during the dominance phase of rivalry is affected by a suppressed image as if it were visible.

Saliency in a suppressed image affects the spatial origin of perceptual alternations during binocular rivalry.

During binocular rivalry, perception alternates between dichoptically presented incompatible images. With larger images, such perceptual alternations will typically start locally and then gradually spread across the image, known as traveling waves of perceptual dominance. Several image-features (such as local contrast) are known to determine where in the image a traveling wave originates. Here we investigate whether orientation contrast in the suppressed image affects these spatial origin(s) of perceptual alternations. The results show that the origins are increasingly biased towards locations of increasing orientation contrast in the suppressed image. This increase in bias is related to the efficiency of visual search for the orientation contrast, tested offline: we find large biases towards orientation contrast when visual search for it is efficient, and small biases when search for it is inefficient. Our results imply that rivalry suppression is not homogenous across the suppressed image, but is dependent on local image-features in the suppressed image. The relation between spatial bias and visual search performance suggests that spatial origins of perceptual alternations are biased to salient locations in the suppressed image. Moreover, the finding that saliency affects the spatial origin of a perceptual alternation is in agreement with the idea that saliency is represented at a monocular, unconscious level of visual processing.

Orientation-tuned suppression in binocular rivalry reveals general and specific components of rivalry suppression.

During binocular rivalry (BR), conflicting monocular images are alternately suppressed from awareness. During suppression of an image, contrast sensitivity for probes is reduced by approximately 0.3-0.5 log units relative to when the image is in perceptual dominance. Previous studies on rivalry suppression have led to controversies concerning the nature and extent of suppression during BR. We tested for feature-specific suppression using orthogonal rivaling gratings and measuring contrast sensitivity to small grating probes at a range of orientations in a 2AFC orientation discrimination task. Results indicate that suppression is not uniform across orientations: suppression was much greater for orientations close to that of the suppressed grating. The higher suppression was specific to a narrow range around the suppressed rival grating, with a tuning similar to V1 orientation bandwidths. A similar experiment tested for spatial frequency tuning and found that suppression was stronger for frequencies close to that of the suppressed grating. Interestingly, no tuned suppression was observed when a flicker-and-swap paradigm was used, suggesting that tuned suppression occurs only for lower-level, interocular rivalry. Together, the results suggest there are two components to rivalry suppression: a general feature-invariant component and an additional component specifically tuned to the rivaling features.