Search

In non-human primates, at the junction of the prefrontal with the premotor cortex, there is a sector designated as frontal eye field (FEF), involved in controlling oculomotor behavior and spatial attention. Evidence for at least two FEFs in humans is at the basis of the still open issue of the possible homologies between the macaque and the human frontal oculomotor system. In this review article we address this issue suggesting a new view solidly grounded on evidence from the last decade showing that, in macaques, the FEF is at the core of an oculomotor domain in which several distinct areas, including areas 45A and 45B, provide the substrate for parallel processing of different aspects of oculomotor behavior. Based on comparative considerations, we will propose a correspondence between some of the macaque and the human oculomotor fields, thus suggesting sharing of neural substrate for oculomotor control, gaze processing, and orienting attention in space. Accordingly, this article could contribute to settle some aspects of the so-called "enigma" of the human FEF anatomy.

The somatosensory system helps the human body to become aware of various stimuli and to interact with its surroundings. Humans are able to identify and to process abundant sensory information quickly due to their unique perception characteristics. As the largest sensory organ, skin has a large number of discrete receptors to sense and to transform stimuli of touch, pressure, pain, temperature, etc. into electrical signals, which are preprocessed at various levels before reaching the brain, greatly reducing the computational burden on the central nervous system. In addition, the conduction speed varies for different stimulus information, which simplifies the parallel processing of a variety of information. In this Perspective, we discuss a bioinspired design for a flexible smart sensor system by simulating the human somatosensory system. In this design, sensors with selective responses, signals separated in time sequences, and hierarchical information processing are adopted to optimize the sensing and perceiving processes, to reduce power consumption, and to improve the speed of a flexible smart sensor system.

Companion robots should perceive speech, recognize objects in the real world, and further react with speech utterances and nonverbal communicative cues. Robots should also remember the interaction history and accumulate knowledge from external text sources: news, blogs, and e-mails. We have designed a conceptual representation system for a companion robot, able to support this list of interactive tasks. The system includes a speech processing component and operates with semantic representations: sets of semantic markers, assigned to valencies within sentence predications. The reaction support system inherits a classic production architecture and consists of scripts, sensitive to rational or emotional stimuli. The general architecture is based on parallel processing of scripts, it may trigger several behavioral reactions in response to a stimulus, and after that it combines speech output and nonverbal reactions of these scripts on the robot, thus, constructing compound and rich behavior. Semantic representations and scripts are also used to index incoming utterances in a memory base.

Skilled readers may misinterpret “you that read wrong” for “you read that wrong”: a transposed-word effect. This relatively novel finding, which supports parallel word processing during sentence reading, is attributed to a combination of noisy bottom-up word position coding and top-down syntactic constraints. The present study focussed on the contribution of syntactic constraints in driving transposed-word effects. In a speeded grammatical decision experiment, two types of ungrammatical transposed-word sequences were compared, namely a transposition either across a syntactic phrase (“the have girls gone home”) or within a syntactic phrase (“the girls gone have home”). We found longer response times and lower accuracy rates for within-phrase transpositions than across-phrase transpositions, demonstrating a direct influence of syntactic structures on the transposed-word effect. We conclude that the assignment of words to positions in a sentence is guided by top-down syntactic constraints.

There is a growing body of research on ensemble perception, or our ability to form ensemble representations based on perceptual features for stimuli of varying levels of complexity, and more recently, on ensemble cognition, which refers to our ability to perceive higher-level properties of stimuli such as facial attractiveness or gaze direction. Less is known about our ability to form ensemble representations based on more abstract properties such as the semantic meaning associated with items in a scene. Previous work examining whether the meaning associated with digits can be incorporated into summary statistical representations suggests that numerical information from digit ensembles can be extracted rapidly, and likely using a parallel processing mechanism. Here, we further investigate whether participants can accurately generate summary representations of numerical value from digit sets and explore the effect of set size on their ability to do so, by comparing psychometric functions based on a numerical averaging task in which set size varied. Steeper slopes for ten- and seven-item compared to five-item digit sets provide evidence that displays with more digits yield more reliable discrimination between larger and smaller numerical averages. Additionally, consistent with previous reports, we observed a response bias such that participants were more likely to report that the numerical average was "greater than 5" for larger compared to smaller sets. Overall, our results contribute to evidence that ensemble representations for semantic attributes may be carried out via similar mechanisms as those reported for perceptual features.

When walking in our natural environment, we often solve additional cognitive tasks. This increases the demand of resources needed for both the cognitive and motor systems, resulting in Cognitive-Motor Interference (CMI). A large portion of neurophysiological investigations on CMI took place in static settings, emphasizing the experimental rigor but overshadowing the ecological validity. As a more ecologically valid alternative to treadmill and desktop-based setups to investigate CMI, we developed a dual-task walking scenario in virtual reality (VR) combined with Mobile Brain/Body Imaging (MoBI). We aimed at investigating how brain dynamics are modulated by dual-task overground walking with an additional task in the visual domain. Participants performed a visual discrimination task in VR while standing (single-task) and walking overground (dual-task). Even though walking had no impact on the performance in the visual discrimination task, a P3 amplitude reduction along with changes in power spectral densities (PSDs) were observed for discriminating visual stimuli during dual-task walking. These results reflect an impact of walking on the parallel processing of visual stimuli even when the cognitive task is particularly easy. This standardized and easy to modify VR paradigm helps to systematically study CMI, allowing researchers to control for the impact of additional task complexity of tasks in different sensory modalities. Future investigations implementing an improved virtual design with more challenging cognitive and motor tasks will have to investigate the roles of both cognition and motion, allowing for a better understanding of the functional architecture of attention reallocation between cognitive and motor systems during active behavior.

The mechanisms guiding visual attention are of great interest within cognitive and perceptual psychology. Many researchers have proposed models of these mechanisms, which serve to both formalize their theories and to guide further empirical investigations. The assumption that a number of basic features are processed in parallel early in the attentional process is common among most models of visual attention and visual search. To date, much of the evidence for parallel processing has been limited to set-size manipulations. Unfortunately, set-size manipulations have been shown to be insufficient evidence for parallel processing. We applied Systems Factorial Technology, a general nonparametric framework, to test this assumption, specifically whether color and shape are processed in parallel or in serial, in three experiments representative of feature search, conjunctive search, and odd-one-out search, respectively. Our results provide strong evidence that color and shape information guides search through parallel processes. Furthermore, we found evidence for facilitation between color and shape when the target was known in advance but performance consistent with unlimited capacity, independent parallel processing in odd-one-out search. These results confirm core assumptions about color and shape feature processing instantiated in most models of visual search and provide more detailed clues about the manner in which color and shape information is combined to guide search.

Several causes make brain cancer identification a challenging task for neurosurgeons during the surgical procedure. The surgeons' naked eye sometimes is not enough to accurately delineate the brain tumor location and extension due to its diffuse nature that infiltrates in the surrounding healthy tissue. For this reason, a support system that provides accurate cancer delimitation is essential in order to improve the surgery outcomes and hence the patient's quality of life. The brain cancer detection system developed as part of the “HypErspectraL Imaging Cancer Detection” (HELICoiD) European project meets this requirement exploiting a non-invasive technique suitable for medical diagnosis: the hyperspectral imaging (HSI). A crucial constraint that this system has to satisfy is providing a real-time response in order to not prolong the surgery. The large amount of data that characterizes the hyperspectral images, and the complex elaborations performed by the classification system make the High Performance Computing (HPC) systems essential to provide real-time processing. The most efficient implementation developed in this work, which exploits the Graphic Processing Unit (GPU) technology, is able to classify the biggest image of the database (worst case) in less than three seconds, largely satisfying the real-time constraint set to 1 minute for surgical procedures, becoming a potential solution to implement hyperspectral video processing in the near future.

Despite the possible social implications of the coronavirus disease 2019 (COVID-19), previous studies of the extended parallel processing model (EPPM) in the context of COVID-19 overlooked the emotional aspects when processing fear-inducing COVID-19-related messages. Drawing upon the moral foundation theory (MFT), this study aimed to (a) apply EPPM in the Korean COVID-19 context, (b) introduce MFT and explain why moral intuitions can be related to the processing of COVID-19 messages, and (c) examine the moderating role of moral intuitions in the EPPM model. Based on the theoretical backgrounds, this study tested EPPM hypotheses and also tested whether moral intuition can moderate the relationship between perceived self-efficacy, perceived threat, fear of COVID-19, and health compliance behavioral intention. This study conducted an online survey using measurements of perceived self-efficacy, perceived threat, MFQ-20, fear of COVID, and health compliance. Our study showed three main findings. First, our study found the main effects of (a) self-efficacy on health compliance behavioral intention and (b) perceived threat on health compliance behavioral intention. Second, our study found that morality moderated the main effects of self-efficacy or perceived threat and also moderated EPPM interaction on fear of COVID. Third, the moderation of morality in the relationship between self-efficacy and health compliance behavioral intention showed that health compliance intention decreased as morality increased. Our findings suggest that people can consider COVID-19 as a social and moral issue that involves protecting others.

The extent to which higher-order representations can be extracted from more than one word in parallel remains an unresolved issue with theoretical import. Here, we used ERPs to investigate the timing with which semantic information is extracted from parafoveal words. Participants saw animal and non-animal targets paired with response congruent or incongruent flankers in a semantic categorization task. Animal targets elicited smaller amplitude negativities when they were paired with semantically related and response congruent animal flankers (e.g., wolf coyote wolf) compared to unrelated and response incongruent flankers (e.g., sock coyote sock) in the N400 window and a post-N400 window. We interpret the N400 effect in terms of facilitated processing from the joint activation of shared semantic features (e.g., animal, furry) across target and flanker words and the later effect in terms of post-lexical decision-making. Thus, semantic information can be extracted from flankers in parallel and impacts various stages of processing. ispartof: BRAIN AND LANGUAGE vol:219 ispartof: location:Netherlands status: published

The modern advances of social media platforms and content sharing websites led to the popularization of Internet memes, and today's Internet landscape contains websites that are predominantly dedicated to meme sharing. While at their inception memes were mostly humorous, this concept evolved and nowadays memes cover a wide variety of subjects, including political and social commentaries. Considering the widespread use of memes and their power of conveying distilled messages, they became an important method for spreading hate speech against individuals or targeted groups.Given the multimodal nature of Internet memes, our proposed approach is also a multimodal one, consisting of two parallel processing branches, one textual and one visual, that are joined in a final classification step, providing prediction results for the samples. We test our approach on the publicly available Memotion 7k dataset and compare our results with the baseline approach developed for the dataset.

Speech processing is highly incremental. It is widely accepted that human listeners continuously use the linguistic context to anticipate upcoming concepts, words, and phonemes. However, previous evidence supports two seemingly contradictory models of how a predictive context is integrated with the bottom-up sensory input: Classic psycholinguistic paradigms suggest a two-stage process, in which acoustic input initially leads to local, context-independent representations, which are then quickly integrated with contextual constraints. This contrasts with the view that the brain constructs a single coherent, unified interpretation of the input, which fully integrates available information across representational hierarchies, and thus uses contextual constraints to modulate even the earliest sensory representations. To distinguish these hypotheses, we tested magnetoencephalography responses to continuous narrative speech for signatures of local and unified predictive models. Results provide evidence that listeners employ both types of models in parallel. Two local context models uniquely predict some part of early neural responses, one based on sublexical phoneme sequences, and one based on the phonemes in the current word alone; at the same time, even early responses to phonemes also reflect a unified model that incorporates sentence level constraints to predict upcoming phonemes. Neural source localization places the anatomical origins of the different predictive models in non-identical parts of the superior temporal lobes bilaterally, with the right hemisphere showing a relative preference for more local models. These results suggest that speech processing recruits both local and unified predictive models in parallel, reconciling previous disparate findings. Parallel models might make the perceptual system more robust, facilitate processing of unexpected inputs, and serve a function in language acquisition.Impact statementSpeech processing engages multiple predictive models, using sublexical, word- and sentence contexts in parallel to anticipate upcoming phonemes.

Visual attention span (VAS) has been shown to make a unique contribution to reading skills over and above phonological awareness and rapid automatized naming (RAN). In the current study, we examined the nature of this unique relationship. In particular, we tested whether VAS reflects the retrieval of a verbal code, serial processing, or parallel multi-element processing. To this end, we presented 180 third graders with tasks for VAS, discrete RAN, and serial RAN as well as serial and discrete reading of short words, pseudowords, and long words. VAS was found to correlate with serial RAN but not with discrete RAN. More important, similar relations were found for VAS with serial and discrete reading, which clearly differed from the format-specific relations between RAN and reading. Together, these findings suggest that VAS and serial RAN are related but are associated with reading for different reasons. Serial RAN appears to reflect serial interword reading processes, whereas the unique contribution of VAS mainly involves the parallel processing of orthographic units within words.

The paper describes a theoretical model for simulating electrical signals in specialized cortical areas using artificial neuron connections. The study focuses on the structure of the visual cortex and the development of an elementary neural model of the primary visual cortex (V1) responsible for processing contrast visual stimuli type. It’s proposed a basic artificial neural structure composed by connecting representative neurons in V1 capable of generating visual evoked potentials (VEP) responses. The reference response model utilized is the P100 wave that appears unconditionally at people with normal vision to visual stimuli such as flashes or contrast reversal. It was considered that the electrical activity of the primary visual cortex is due to the functioning of a set of excitatory and inhibitory neurons and so, it was simulated the VEP signal with simple type neurons.

A general problem of sensory systems is how to simultaneously encode prevailing input as well as deviations from this baseline. A new study shows how the fly visual system has solved this by using parallel processing.

Reading research has long endorsed the view that words are processed strictly one by one. The primary empirical test of this notion is the search for effects from upcoming words on readers' eye movements during sentence reading. Here we argue that no conclusions can be drawn from the absence of such effects, and that the serial versus parallel processing debate cannot be resolved without treading beyond the methodological scope of tracking eye movements. Recent considerations of how the brain organizes linguistic input have sparked key predictions in- and outside the realm of text reading, with ensuing research revealing phenomena that complicate the serial processing perspective. A case is made for parallelism, along with new methods to infer the cognitive architecture driving reading.

During reading, word recognition speed is influenced by the amount of orthographic overlap with surrounding words. The nature of this phenomenon is not understood: some theories attribute it to low-level visual operations (i.e., parafoveal feature detectors influencing foveal letter detectors), whereas other theories assume that orthographic processing (i.e., letter position coding and word activation) occurs across multiple words in parallel. To arbitrate between these theories, we used electroencephalography to reveal the time course of orthographic spatial integration in a lexical decision task. Foveal target words were flanked on each side by parafoveal words, manipulated across three conditions: repetition flankers (e.g. rock rock rock), unrelated flankers (step rock step) and a no-flanker condition. Linear mixed-effect models were constructed to analyze EEG data on a trial-by-trial basis. Word recognition was worse in the unrelated flanker condition than in the repetition and no-flanker conditions. This behavioral pattern was accompanied by increased negativity in the N250 and N400 windows, associated with the activation of sub-lexical and lexico-semantic representations, respectively. Crucially, the absence of effects prior to 200 ms post-stimulus onset provides evidence against the involvement of low-level visual processes. We conclude that orthographic spatial integration is driven by parallel processing of multiple words, which leads to the activation of a larger set of sub-lexical nodes and more difficult processing at the lexical level when those words are orthographically unrelated., Highlights • Words are recognized faster when they share many letters with surrounding words. • EEG measurements suggest that this effect unfolds 200 ms after stimulus onset. • The locus of effects contradicts the involvement of low-level visual processing. • Instead, the phenomenon appears to be driven by parallel orthographic processing.

In the current era of the rapid development of big data, it has become increasingly critical and practical to study recommender systems with streaming data input. However, the recommender system is often faced with the challenge that the history records of new users or anonymous users are not available. Specifically, session-based recommendation, which aims to predict a user’s next actions, is a typical task to overcome the challenge. To capture a user’s long-term preference in session-based recommendations, recurrent neural networks (RNN)-based models have been widely applied with impressive results, but the inherent sequential nature of RNNs prevents parallelism within training examples, which is critical in long sessions because memory constraints limit batching across instances. In this paper, we propose a novel method, i.e., self-attention network for session-based recommendation (SANSR), which is based on only attention mechanisms, dispensing with recurrence, and supports parallelism in the session. The proposed model attempts to find items that are relevant based on previous time steps in the ongoing session and to assign them different weights to predict the next item. The extensive experiments are conducted on two real-world datasets, and the experimental results show that our proposed model is superior to the state-of-the-art methods.

Using eye-tracking, we examined whether overspecification hinders or facilitates referent selection and the extent to which this depends on the properties of the attribute mentioned in the referring expressions and the underpinning processing mode. Following spoken instructions, participants selected the referent in a visual display while their eye movements were monitored. The referring expressions were presented either simultaneously with the displays, so the attributes could be incrementally processed in sequence, or before the display presentation, so the attributes could be processed in parallel from the outset of search. Experiment 1 showed that when the attributes were processed incrementally, how quickly an earlier-mentioned attribute discriminated determined whether a late-mentioned, overspecified attribute contributed to discrimination: When color was mentioned first and was fully discriminating, the referent was selected fast regardless of the second-mentioned pattern, whereas when pattern was mentioned first and fully discriminating, the second-mentioned color facilitated discrimination. Experiment 2 found that under incremental processing, color mention after a fully discriminating pattern increased fixations but delayed referent selection relative to a pattern-only description; under parallel processing, however, color mention immediately eliminated alternatives and sped up referent selection. Experiment 3 showed that pattern mention after a fully discriminating color delayed referent selection and tended to reduce fixations relative to a color-only description in both processing modes. Hence, additional attributes can speed up referent selection but only when they can discriminate much faster than alternative attributes mentioned in a more concise description and, critically, when they can be used early for referent search. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Reading words aloud is a fundamental aspect of literacy. The rapid rate at which multiple distributed neural substrates are engaged in this process can only be probed via techniques with high spatiotemporal resolution. We probed this with direct intracranial recordings covering most of the left hemisphere in 46 humans as they read aloud regular, exception and pseudo-words. We used this to create a spatiotemporal map of word processing and to derive how broadband gamma activity varies with multiple word attributes critical to reading speed: lexicality, word frequency and orthographic neighborhood. We found that lexicality is encoded earliest in mid-fusiform (mFus) cortex and precentral sulcus, and is represented reliably enough to allow single-trial lexicality decoding. Word frequency is first represented in mFus and later in the inferior frontal gyrus (IFG) and inferior parietal sulcus (IPS), while orthographic neighborhood sensitivity resides solely in IPS. We thus isolate the neural correlates of the distributed reading network involving mFus, IFG, IPS, precentral sulcus and motor cortex and provide direct evidence for dual-route models of reading, with parallel processes via the lexical route from mFus to IFG, and the sub-lexical route from IPS and precentral sulcus to anterior IFG.

We report that airflow produces a complex activation pattern in the antennal lobes of the honeybee Apis mellifera. Glomerular response maps provide a stereotypical code for the intensity and the dynamics of mechanical stimuli that is superimposed on the olfactory code. We show responses to modulated stimuli suggesting that this combinatorial code could provide information about the intensity, direction, and dynamics of the airflow during flight and waggle dance communication.

My research is on Human-Robot Interaction (HRI) from a User Centered Design (UCD) perspective and values the user’s qualitative assessment of their interaction under the concept of User Experience (UX). Traditionally, UX is retrieved in a written questionnaire as a satisfaction survey, nevertheless, with the hike in electronic devices equipped with a diversity of sensors, actuators and processing capabilities (turning everyday appliances into robots) comes an intention to migrate from qualitative to quantitative assessment of the UX; for example, visual recognition of emotional facial expressions. In this regard, what I present as a novelty in the research area is the study of interaction data generated through the robot’s interface during its intended use (everyday use), while providing an anonymous experience (no cameras, no mics) to the user. For the experimental setup, I designed and built a robotic Desk-Lamp with 5 degrees of freedom (height, brightness, projection angle, sensitivity of the interface and ambient lighting) as test bench, equipped with: an interface instrumented to sense the force delivered by the user while “pushing buttons”; 18 variable data-logging system at 20 SPS (Sample Per Second); high precision and high repeatability electronics and mechanisms; real-time sensing of the user and low-latency operation due to parallel processing. The test bench is designed to be used in multiple experiments with users, varying the robot behavior and interaction procedures to compare the interaction data generated in the interface with a written UX questionnaire.

Summary: It has been proposed that sound information is separately streamed into onset and offset pathways for parallel processing. However, how offset responses contribute to auditory perception remains unclear. Here, loose-patch and whole-cell recordings in awake mouse primary auditory cortex (A1) reveal that a subset of pyramidal neurons exhibit a transient “Off” response, with its onset tightly time-locked to the sound termination and its frequency tuning similar to that of the transient “On” response. Both responses are characterized by excitation briefly followed by inhibition, with the latter mediated by parvalbumin (PV) inhibitory neurons. Optogenetically manipulating sound-evoked A1 responses at different temporal phases or artificially creating phantom sounds in A1 further reveals that the A1 phasic On and Off responses are critical for perceptual discrimination of sound duration. Our results suggest that perception of sound duration is dependent on precisely encoding its onset and offset timings by phasic On and Off responses.

Humans have a remarkable ability to efficiently explore visual scenes and text using eye movements. Humans typically make eye movements (saccades) every ~250 ms. Since saccade initiation and execution take 100 ms, this leaves only ~150 ms to recognize the fixated object (or word) while simultaneously previewing candidates for the next saccade goal. We propose a pipelining mechanism where serial processing occurs within a specific brain region, whereas parallel processing occurs across different brain regions. The mechanism is timed by alpha oscillations that coordinate the saccades, visual recognition, and previewing in the cortical hierarchy. Consequently, the neuronal mechanism supporting natural vision and saccades must be studied in unison to uncover the brain mechanisms supporting visual exploration and reading.

Active vision, the ability of the visual system to actively sample and select relevant information out of a visual scene through eye and head movements, has been explored in a variety of animal species. Small-brained animals such as insects might rely even more on sequential acquisition of pattern features since there might be less parallel processing capacity in their brains than in vertebrates. To investigate how active vision strategies enable bees to solve visual tasks, here, we employed a simple visual discrimination task in which individual bees were presented with a multiplication symbol and a 45° rotated version of the same pattern (“plus sign”). High-speed videography of unrewarded tests and analysis of the bees’ flight paths shows that only a small region of the pattern is inspected before successfully accepting a target or rejecting a distractor. The bees’ scanning behaviour of the stimuli differed for plus signs and multiplication signs, but for each of these, the flight behaviour was consistent irrespective of whether the pattern was rewarding or unrewarding. Bees typically oriented themselves at ~±30° to the patterns such that only one eye had an unobscured view of stimuli. There was a significant preference for initially scanning the left side of the stimuli. Our results suggest that the bees’ movement may be an integral part of a strategy to efficiently analyse and encode their environment.Summary statementAutomated video tracking and flight analysis is proposed as the next milestone in understanding mechanisms underpinning active vision and cognitive visual abilities of bees.

The visualization of the whole eye fundus with enough resolution to discriminate single photoreceptors would be of an enormous interest for understanding retinal diseases and distrophies. In this work, we present a versatile and flexible SLO device that is able to provide high quality images in real time either in large field of view (40ox30o) or small field of view but with high-resolution (4ox3o). The combination of an efficient electronics design and the optical system with adaptive optics provides a large set of customization parameters.

Adaptive optics optical coherence tomography technology enables non-invasive high-resolution retinal imaging and promises earlier detection of ocular disease. However, the images are corrupted by eye-movement artifacts that must be corrected to permit proper image analysis. We developed a method for efficiently removing eye-movement artifacts of A-lines using a multiple-reference global coordinate system. It corrects 3D translational eye movements, torsional eye movements, and image scaling, minimizing image distortion and substantially improving both regularity of the cone photoreceptor mosaic and clarity of individual cones.

Cerebellar granule cells (GrCs) are usually regarded as a uniform cell type that collectively expands the coding space of the cerebellum by integrating diverse combinations of mossy fiber inputs. Accordingly, stable molecularly or physiologically defined GrC subtypes within a single cerebellar region have not been reported. The only known cellular properties that distinguishes otherwise homogeneous GrCs is the correspondence between GrC birthtime and the depth of the molecular layer to which their axons (parallel fibers) project. To determine the role birth timing plays in GrC wiring and function, we developed genetic strategies to access early- and late-born GrCs. We initiated retrograde monosynaptic rabies virus tracing from control, early-born, and late-born GrCs, revealing the different patterns of mossy fiber input to GrCs in vermis lobule 6 and simplex, as well as to early- and late-born GrCs of vermis lobule 6: sensory and motor nuclei provide more input to early-born GrCs, while basal pontine and cerebellar nuclei provide more input to late-born GrCs.In vivomulti-depth 2-photon Ca2+imaging of parallel fibers of early- and late-born GrCs revealed representations of diverse task variables and stimuli by both populations, with differences in the proportions of parallel fibers encoding movement, reward anticipation, and reward consumption. Our results suggest neither organized parallel processing nor completely random organization of mossy fiber→GrC circuitry, but instead a moderate influence of birth timing on GrC wiring and encoding. Our imaging data also suggest that GrCs can represent general aversiveness, in addition to recently described reward representations.Significance StatementCerebellar granule cells (GrCs) comprise the majority of all neurons in the mammalian brain and are usually regarded as a uniform cell type. However, the birth timing of an individual GrC dictates where its axon projects. Using viral-genetic techniques, we find that early- and late-born GrCs receive different proportions of inputs from the same set of input regions. Usingin vivomulti-depth 2-photon Ca2+imaging of axons of early- and late-born GrCs, we found that both populations represent diverse task variables and stimuli, with differences in the proportions of axons in encoding of a subset of movement and reward parameters. These results indicate that birth timing contributes to the input selection and physiological response properties of GrCs.

It is unclear how different types of cortical projection neurons work together to support diverse cortical functions. We examined the discharge characteristics and inactivation effects of intratelencephalic (IT) and pyramidal tract (PT) neurons—two major types of cortical excitatory neurons that project to cortical and subcortical structures, respectively—in the medial prefrontal cortex of mice performing a delayed response task. We found that IT neurons, but not PT neurons, convey significant working memory-related signals. We also found that the inactivation of IT neurons, but not PT neurons, impairs behavioral performance. In contrast, PT neurons convey far more temporal information than IT neurons during the delay period. Our results indicate a division of labor between IT and PT projection neurons in the prefrontal cortex for the maintenance of working memory and for tracking the passage of time, respectively.

Recently, with the development of technology, computer vision research based on the human visual system has been actively conducted. Saliency maps have been used to highlight areas that are visually interesting within the image, but they can suffer from low performance due to external factors, such as an indistinct background or light source. In this study, existing color, brightness, and contrast feature maps are subjected to multiple shape and orientation filters and then connected to a fully connected layer to determine pixel intensities within the image based on location-based weights. The proposed method demonstrates better performance in separating the background from the area of interest in terms of color and brightness in the presence of external elements and noise. Location-based weight normalization is also effective in removing pixels with high intensity that are outside of the image or in non-interest regions. Our proposed method also demonstrates that multi-filter normalization can be processed faster using parallel processing.

The active battle sonar abroad adopts low frequency, high power, large aperture array and adaptive processing technology, which integrates active/passive, multiband and multi-function, and presents the development trend of fully adaptive intelligent cognition. Therefore, the research on how to develop effective countermeasures should be carried out. Firstly, the basic structure of cognitive sonar and its challenges to confrontation are introduced on this paper. Secondly, the artificial intelligence technology is applied to the field of countermeasure, and the architecture of cognitive sonar intelligent countermeasure system is proposed. Finally, the key technologies of cognitive sonar intelligent countermeasure are analyzed, which lays a foundation for the research of cognitive sonar intelligent countermeasure technology.

Moving object detection in a video plays an important role in many vision applications. Recently, moving object detection using appearance modeling based on a convolutional neural network has been actively developed. However, the CNN-based methods usually require the user's supervision of the first frame so that it becomes highly dependent on the training dataset. In contrast, the method of finding a foreground, which models a background occupying a large proportion in an image, can detect a moving object efficiently in an unsupervised manner. However, existing methods based on background modeling in a pan-tilt-zoom (PTZ) camera suffer many false positives or loss of moving objects due to the estimation error of camera motion. To overcome the aforementioned limitations, we propose a moving object detection method for a PTZ camera through two background models. In an unsupervised way, our method builds the two background models that have different roles: 1) a coarse background model for detecting large changes, and 2) a fine background model for detecting small changes. In more detail, the coarse background model builds a block-based Gaussian model, and the fine model builds a sample consensus model. Both models are adaptively updated according to the estimated camera motion in the video recorded by a PTZ camera. Then, each foreground result from two background models is incorporated to fill the moving object region. Through experiments, the proposed method achieves better performance than the state-of-the-art methods and operates in real-time without parallel processing. In addition, we showed the effectiveness of the proposed model through improved results of moving object detection through combination with the latest supervised method.

Vision is the principal source of information of the surrounding world. It facilitates our movement and development of everyday activities. In this sense, blind people have great difficulty for moving, especially in unknown environments, which reduces their autonomy and puts them at risk of suffering an accident. Electronic Travel Aids (ETAs) have emerged and provided outstanding navigation assistance for blind people. In this work, we present the methodology followed for implementing a stereo vision-based system that assists blind people to wander unknown environments in a safe way, by sensing the world, segmenting the floor in 3D, fusing local 2D grids considering the camera tracking, creating a global occupancy 2D grid, reacting to close obstacles, and generating vibration patterns with an haptic belt. For segmenting the floor in 3D, we evaluate normal vectors and orientation of the camera obtained from depth and inertial data, respectively. Next, we apply RANSAC for computing efficiently the equation of the supporting plane (floor). The local grids are fused, obtaining a global map with data of free and occupied areas along the whole trajectory. For parallel processing of dense data, we leverage the capacity of the Jetson TX2, achieving high performance, low power consumption, and portability. Finally, we present experimental results obtained with ten (10) participants, in different conditions, with obstacles of different height, hanging obstacles, and dynamic obstacles. These results show high performance and acceptance by the participants, highlighting the easiness to follow instructions and the short period of training.

This study investigates the impact of source-text metaphors on the cognitive patterns when translating in both directions of a language pair and whether this impact is affected by directionality. Thirty-eight Chinese trainee translators performed a series of translation tasks from Chinese to English and from English to Chinese, and their performances were recorded by eye-tracking and keylogging and all participants gave cue-based retrospective reports. Four eye-key combination indicators were calculated using generalized linear models to demonstrate translators’ allocation of cognitive resources: Total Attentional Duration, Attention Unit Count, Attention Unit Duration, and pupil dilation. The eye-key data were compared with the participants’ introspections. The overall findings suggest that translation directionality can significantly affect the relationship between processing types (ST processing, TT processing and parallel processing), metaphor-related text types and attention distribution patterns. Detailed eye-key descriptions show that this impact is inconsistent.

We present an ensemble learning methodology that combines multiple existing robotic grasp synthesis algorithms and obtain a success rate that is significantly better than the individual algorithms. The methodology treats the grasping algorithms as "experts" providing grasp "opinions". An Ensemble Convolutional Neural Network (ECNN) is trained using a Mixture of Experts (MOE) model that integrates these opinions and determines the final grasping decision. The ECNN introduces minimal computational cost overhead, and the network can virtually run as fast as the slowest expert. We test this architecture using open-source algorithms in the literature by adopting GQCNN 4.0, GGCNN and a custom variation of GGCNN as experts and obtained a 6% increase in the grasp success on the Cornell Dataset compared to the best-performing individual algorithm. The performance of the method is also demonstrated using a Franka Emika Panda arm., Comment: Accepted in: 2021 IEEE International Conference on Robotics and Automation (ICRA)

Speech synthesis technology is evolving from the concept of converting texts into voices to directly learning and imitating user voices. In the past, the voice was generated through several stages of recording frequently used sentences and synthesizing text into sounds that wanted to be converted. However, due to the limitations of naturalness and accuracy, research has been conducted to overcome this in various fields such as phonetics, linguistics, and statistics, and now the development of big data, AI technology, and parallel processing technology has increased to reflect human tone and timbre. If the voice information, which is one of the biometric information that can be used as personally identifiable information, can be made so natural that it is difficult for a machine to distinguish it from a human, personal information may be invaded by a security threat that may exist at any time. Therefore, in this paper, we explain representative speech synthesis models and introduce voice generation models, which are applications, to find out how to convert and synthesize Korean into other languages, and to look at what may become a security issue in the future.

It is argued that certain kinds of problem in modern society impose a requirement on ABM to represent all human agents on the planet. Eleven reason are given, including the need for realism in social models, the importance of boundary conditions and scale, the need for global social justice, the existence of global scale dynamics created by and impacting directly on individuals, and the need to interface with other global models in order to address pressing problems such as climate change and ecosystem destruction. An indication of the difficulties involved in creating such models is given, drawing on experience of creating models with RepastHPC. The paper concludes by suggesting that rather than creating ever more model platforms and frameworks, what we need is a series of shared and collectively developed models, in a similar way to existing traffic models or, models of the physical parts of the earth system.

In the wake of the developments of deep learning, more and more research focus on the intersection fields of natural language processing and machine vision, vision and language navigation (VLN) is one of them. The VLN task needs an embodied agent to carry out the natural language instruction and navigate inside a real 3D environment with the help of visual information, planning a trajectory from start point to goal location. In this paper, inspired by the previous works, we introduce a multi-head attention module with a parallel attention computing method which apply multi-head attention mechanism on visual and textual input to enhance the performance of the model. Specifically, first we design a multihead attention module with trainable parameters, which can extract associated attention from the textual and visual information, extracted attention variables are used to promote the agent aware of which part of the sentence or image is more important. Second, in order to help the agent perceives more useful information, we perform a parallel computing method on extracted attention and input, i.e. visual and textual input feature, then use a layer normalization to combine them. The results of experiments indicate our proposed module enables the model to obtain better performance and surpasses the baseline model. The success rate of our model is 51% and oracle success rate is 62% with low navigation error.

Iconic gesture-speech integration is a relatively recent field of investigation with numerous researchers studying its various aspects. The results obtained are just as diverse. The definition of iconic gestures is often overlooked in the interpretations of results. Furthermore, while most behavioral studies have demonstrated an advantage of bimodal presentation, brain activity studies show a diversity of results regarding the brain regions involved in the processing of this integration. Clinical studies also yield mixed results, some suggesting parallel processing channels, others a unique and integrated channel. This review aims to draw attention to the methodological variations in research on iconic gesture-speech integration and how they impact conclusions regarding the underlying phenomena. It will also attempt to draw together the findings from other relevant research and suggest potential areas for further investigation in order to better understand processes at play during speech integration process.

The dendrites of a retinal interneuron detect visual threats and initiate defensive responses through parallel pathways., Approaching predators cast expanding shadows (i.e., looming) that elicit innate defensive responses in most animals. Where looming is first detected and how critical parameters of predatory approaches are extracted are unclear. In mice, we identify a retinal interneuron (the VG3 amacrine cell) that responds robustly to looming, but not to related forms of motion. Looming-sensitive calcium transients are restricted to a specific layer of the VG3 dendrite arbor, which provides glutamatergic input to two ganglion cells (W3 and OFFα). These projection neurons combine shared excitation with dissimilar inhibition to signal approach onset and speed, respectively. Removal of VG3 amacrine cells reduces the excitation of W3 and OFFα ganglion cells and diminishes defensive responses of mice to looming without affecting other visual behaviors. Thus, the dendrites of a retinal interneuron detect visual threats, divergent circuits downstream extract critical threat parameters, and these retinal computations initiate an innate survival behavior.

Intense interests are a core symptom of autism spectrum disorders (ASD) and can be all-encompassing for affected individuals. This observation raises the hypothesis that intense interests in ASD are related to pervasive changes in visual processing for objects within that category, including visual search. We assayed visual processing with two novel tasks, targeting category search and exemplar search. For each task, three kinds of stimuli were used: faces, houses, and images personalized to each participant’s interest. 25 children and adults with ASD were compared to 25 neurotypical (NT) children and adults. We found no differences in either visual search task between ASD and NT controls for interests. Thus, pervasive alterations in perception are not likely to account for ASD behavioral symptoms.

International audience; The sentence superiority effect observed with skilled adult readers has been taken to reflect parallel processing of word identities and the rapid construction of a preliminary syntactic structure. Here we examined if such processing is already present in primary school children in Grade 3 (average age 8.9 years). Children saw sequences of four horizontally aligned words presented simultaneously for 500 ms and followed by a post-mask and post-cue indicating the position for report of one of the four words. Word identification was more accurate in grammatically correct sequences compared with ungrammatical scrambled sequences of the same words, and this sentence superiority effect did not interact with position. This replicates the pattern found in prior research with adults and suggests that parallel word processing and the associated efficiency in syntactic processing arealready in place in Grade 3. We also found that accuracy in identifying words, independently of the surrounding context, correlated with reading age. This points to efficient word-in-sequence identification as one key ingredient of the process of becoming a skilled reader.

Partial awareness—an intermediate state between complete consciousness and unconsciousness—has been explained by independent cognitive access to different levels of representation in hierarchical visual processing. This account, however, cannot explain graded visual experiences in low levels. We aimed to explain partial awareness in low levels of visual processing by independent cognitive access to different spatial frequencies. To observe partial awareness stably, we used a novel method. Stimuli were presented briefly (12 ms) and repeatedly with a specific inter-stimulus interval, ranging from 0 to 235 ms. By using various stimuli containing high and low spatial frequencies (superimposed sinusoidal gratings, Navon letters, and scenes), we found that conscious percept was degraded with increasing inter-stimulus intervals. However, the degree of degradation was smaller for low spatial frequency than for high spatial frequency information. Our results reveal that cognitive access to different spatial frequencies can occur independently and this can explain partial awareness in low levels of visual processing.

International audience; When reading, orthographic information is extracted not only from the word the reader is looking at, but also from adjacent words in the parafovea. Here we examined, using the recently introduced OB1-reader computational model, how orthographic information can be processed in parallel across multiple words and how orthographic information can be integrated across time and space. Although OB1-reader is a model of text reading, here we used it to simulate single-word recognition experiments in which parallel processing has been shown to play a role by manipulating the surrounding context in flanker and priming paradigms. In flanker paradigms, observers recognize a central word flanked by other letter strings located left and right of the target and separated from the target by a space. The model successfully accounts for the finding that such flankers can aid word recognition when they contain bigrams of the target word, independent of where those flankers are in the visual field. In priming experiments, in which the target word is preceded by a masked prime, the model accounts for the finding that priming occurs independent of whether the prime and target word are in the same location or not. Crucial to these successes is the key role that spatial attention plays within OB1-reader, as it allows the model to receive visual input from multiple locations in parallel, while limiting the kinds of errors that can potentially occur under such spatial pooling of orthographic information.

SummaryLarval zebrafish (Danio rerio) are an ideal organism to study color vision, as their eye possesses four types of cone photoreceptors, covering most of the visible range and into the UV [1,2]. Additionally, their entire eye and nervous system are accessible to imaging, given they are naturally transparent [3–5]. Relying on this advantage, recent research has found that, through a set of color specific horizontal, bipolar and retinal ganglion cells (RGCs) [6–8], the eye then relays tetrachromatic information to several retino-recipient areas (RAs) [9,10]. The main RA is the optic tectum, receiving 97% of the RGC axons via the neuropil mass termed Arborization Field 10 (AF10) [11,12]. In this work, we aim to understand the processing of color signals at the interface between RGCs and their targets in the brain. We used 2-photon calcium imaging to separately measure the responses of RGCs and neurons in the dorsal brain to stimulation with four different colors in awake animals. We find that color information is widespread throughout the larval brain, with a large variety of color responses among RGCs, and an even greater diversity in their targets. Specific combinations of response types are localized to specific nuclei, but we observe no single color processing structure. In the main interface in this pathway, the connection between Arborization Field 10 and the tectum, we observe key elements of color processing such as enhanced signal decorrelation and improved decoding [13,14]. Finally, when presenting a richer set of stimuli, we identify parallel processing of color, motion and luminance information in the same cells/terminals, evidence of a rich color vision machinery in this small vertebrate brain.

In this paper, we present a learning-based method to the keyframe-based video stylization that allows an artist to propagate the style from a few selected keyframes to the rest of the sequence. Its key advantage is that the resulting stylization is semantically meaningful, i.e., specific parts of moving objects are stylized according to the artist's intention. In contrast to previous style transfer techniques, our approach does not require any lengthy pre-training process nor a large training dataset. We demonstrate how to train an appearance translation network from scratch using only a few stylized exemplars while implicitly preserving temporal consistency. This leads to a video stylization framework that supports real-time inference, parallel processing, and random access to an arbitrary output frame. It can also merge the content from multiple keyframes without the need to perform an explicit blending operation. We demonstrate its practical utility in various interactive scenarios, where the user paints over a selected keyframe and sees her style transferred to an existing recorded sequence or a live video stream.

In visual search tasks, observers look for targets among distractors. In the lab, this often takes the form of multiple searches for a simple shape that may or may not be present among other items scattered at random on a computer screen (e.g., Find a red T among other letters that are either black or red.). In the real world, observers may search for multiple classes of target in complex scenes that occur only once (e.g., As I emerge from the subway, can I find lunch, my friend, and a street sign in the scene before me?). This article reviews work on how search is guided intelligently. I ask how serial and parallel processes collaborate in visual search, describe the distinction between search templates in working memory and target templates in long-term memory, and consider how searches are terminated.

SummaryCortical parallel processing streams segregate many diverse features of a sensory scene. However, some features are distributed across streams, begging the question of whether and how such distributed representations contribute to perception. We determined the necessity of primary visual cortex (V1) and three key higher visual areas (LM, AL and PM) for perception of orientation and contrast, two features that are robustly encoded across all four areas. Suppressing V1, LM or AL decreased sensitivity for both orientation discrimination and contrast detection, consistent with a role for these areas in sensory perception. In comparison, suppressing PM selectively increased false alarm rates during contrast detection, without any effect on orientation discrimination. This effect was not retinotopically-specific, suggesting a distinct role for PM in the regulation of noise during decision-making. Thus, we find that distributed representations in the visual system can nonetheless support specialized perceptual roles for higher visual cortical areas.

Cortical parallel processing streams segregate many diverse features of a sensory scene. However, some features are distributed across streams, begging the question of whether and how such distributed representations contribute to perception. We determined the necessity of the primary visual cortex (V1) and three key higher visual areas (lateromedial [LM], anterolateral [AL], and posteromedial [PM]) for perception of orientation and contrast, two features that are robustly encoded across all four areas. Suppressing V1, LM, or AL decreased sensitivity for both orientation discrimination and contrast detection, consistent with a role for these areas in sensory perception. In comparison, suppressing PM selectively increased false alarm (FA) rates during contrast detection, without any effect on orientation discrimination. This effect was not retinotopically specific, suggesting that suppression of PM altered sensory integration or the decision-making process rather than processing of local visual features. Thus, we find that distributed representations in the visual system can nonetheless support specialized perceptual roles for higher visual cortical areas.

Parallel processing circuits are thought to dramatically expand the network capabilities of the nervous system. Magnocellular and parvocellular oxytocin neurons have been proposed to subserve two parallel streams of social information processing, which allow a single molecule to encode a diverse array of ethologically distinct behaviors, although to date direct evidence to support this hypothesis is lacking. Here we provide the first comprehensive characterization of magnocellular and parvocellular oxytocin neurons, validated across anatomical, projection target, electrophysiological, and transcriptional criteria. We next used novel multiple feature selection tools in Fmr1 KO mice to provide direct evidence that normal functioning of the parvocellular but not magnocellular oxytocin pathway is required for autism-relevant social reward behavior. Finally, we demonstrate that autism risk genes are uniquely enriched in parvocellular oxytocin neurons. Taken together these results provide the first evidence that oxytocin pathway specific pathogenic mechanisms account for social impairments across a broad range of autism etiologies.One Sentence SummaryPathoclisis of parvocellular oxytocin neurons plays an important role in the pathogenesis of social impairments in autism.