117,575 research outputs found
The primary visual cortex creates a bottom-up saliency map
No full textIt has been proposed that the primary visual cortex (V1) creates a saliency map using autonomous intra-cortical mechanisms. This saliency of a visual location describes the location's ability to attract attention without top-down factors. It increases monotonously with the firing rate of the most active V1 cell responding to that location. Given the prevalent feature selectivities of V1 cells (many tuned to more than one feature dimension), no separate feature maps, or any subsequent combinations of them, are needed to create a saliency map. This proposal has been demonstrated in a biologically based V1 model. By relating the saliencies of the visual search targets or object (texture) boundaries to the eases of the visual search or segmentation tasks, the model accounted for behavioral data such as how task difficulties can be influenced by image features and their spatial configurations. This proposal links physiology with psychophysics, thereby making testable predictions some of which are subsequently confirmed experimentally
Peripheral vision in the central-peripheral dichotomy
No full textCompared to central vision, peripheral vision has not only a lower spatial sampling resolution in the retina, but also, according to the recently proposed central-peripheral dichotomy (CPD, Zhaoping 2017, 2019), has a primary role for looking rather than seeing in vision. Furthermore, for seeing (i.e., recognizing and discriminating visual objects), CPD asserts that peripheral vision has a weaker or absent feedback component in the feedforward and feedback processes along the visual pathway from the primary visual cortex (V1) to higher visual areas. Due to an attentional bottleneck assumed to start from V1's output to downstream areas (Zhaoping 2019), visual recognition in higher visual areas relies on impoverished sensory information fed forward from V1. To aid recognition in challenging or ambiguous situations, in which the perceptual outcome from viewing a scene could be one of multiple non-trivial possibilities, central vision uses feedback from higher to lower visual areas such as V1 to query for additional information. This query uses brain's internal model of the visual world to disambiguate between the possibilities for an eventual perceptual outcome. Peripheral vision, with a weaker or absent feedback query according to CPD, is therefore vulnerable to visual illusions due to misleading V1 inputs. I will show two visual illusions predicted by CPD using our knowledge about V1's neural response properties. One is called the reversed depth illusion (Zhaoping &Ackermann 2018) in perceiving the 3-dimensional depth of a surface from a viewer. The other is called the flip tilt illusion (Zhaoping 2020) in perceiving the orientation of an item in an image. Usually, both illusions are only visible peripherally. A relative of the flip tilt illusion is a surprising prediction of a parallel advantage: in a special visual search task, it is faster to find a target that is parallel rather than perpendicular to uniformly oriented nontargets (Zhaoping 2022). As in typical visual search tasks, time needed for completing the task is largely determined by looking, which is the process of deciding where in the peripheral visual field to make a saccade to, until the target is located at the saccadic destination. Hence, this predicted parallel advantage highlights the role of peripheral vision in looking. Indeed, this parallel advantage is stronger for targets at the more peripheral visual fields
Mathematical analysis and simulations of the neural circuit for locomotion in lampreys ;
No full textWe analyze the dynamics of the neural circuit of the lamprey central pattern generator. This analysis
provides insight into how neural interactions form oscillators and enable spontaneous oscillations in a
network of damped oscillators, which were not apparent in previous simulations or abstract phase
oscillator models. We also show how the different behavior regimes (characterized by phase and
amplitude relationships between oscillators) of forward or backward swimming, and turning, can be
controlled using the neural connection strengths and external inputs
A full reference quality metric for geometrically distorted images
No full textIn multimedia applications, there has been an increasing interest in the use of quality measures based on human perception; however, research has not dealt with distortions due to geometric transformations. In this paper, we propose a method to objectively assess the perceptual quality of geometrically distorted images, based on image features processed by human vision. The proposed approach is a full-reference image quality metric focusing on the problem of local geometric distortions and is based on the use of Gabor filters that have received considerable attention because the characteristics of certain cells in the visual cortex of some mammals can be approximated by these filters. The novelty of the proposed technique is that it considers both the displacement field describing the distortion and the structure of the image. The experimental results show the good performances of the proposed metric
Parallel popout: further confirmation of the V1 Saliency Hypothesis (V1SH)
No full textFinding a target among uniformly oriented non-targets is typically faster when this target is perpendicular, rather than parallel, to the non-targets. Here, by exploiting the properties of saliency computations in primary visual cortex (V1), I demonstrate a special case when exactly the opposite is true. Each item, target or non-target, comprises two disks of the same size; the centre of one disk is displaced 1.2 disk diameters from that of the other along a line defining the item’s orientation. A target has two black disks or two white disks; each non-target has one white disk and one black disk. The target is oriented 45 degree clockwise or counter-clockwise from horizontal; the non-targets are uniformly oriented either perpendicular or parallel to the target in a grey background. Unlike the target, each non-target activates a neuron in V1 more strongly when its orientation is perpendicular rather than parallel to the neuron's preferred orientation, since the white and the black disks best activate, respectively, the on- and off- subfields of the neural receptive field (Zhaoping, L., 2020, i- Perception, 11(4):1--5.). V1 neurons are suppressed more strongly by neighbouring neurons tuned to similar rather than dissimilar orientations. Thus, a target parallel (rather than perpendicular) to the non-targets evokes a higher V1 response and, according to V1SH, is more salient. Our behavioural confirmation of faster search in this condition supports V1SH's proposal that V1 is the neural basis for saliency of exogenous attentional selection (Zhaoping, L., 2002, Trends in Cognitive Sciences 6(1):9-16)
Hebbian imprinting and retrieval in oscillatory neural networks
No full textWe introduce a model of generalized Hebbian learning and retrieval in oscillatory neural networks modeling cortical areas such as hippocampus and olfactory cortex. Recent experiments have shown that synaptic plasticity depends on spike timing, especially on synapses from excitatory pyramidal cells, in hippocampus, and in sensory and cerebellar cortex. Here we study how such plasticity can be used to form memories and input representations when the neural dynamics are oscillatory, as is common in the brain (particularly in the hippocampus and olfactory cortex). Learning is assumed to occur in a phase of neural plasticity, in which the network is clamped to external teaching signals. By suitable manipulation of the nonlinearity of the neurons or the oscillation frequencies during learning, the model can be made, in a retrieval phase, either to categorize new inputs or to map them, in a continuous fashion, onto the space spanned by the imprinted patterns. We identify the first of these possibilities with the function of olfactory cortex and the second with the observed response characteristics of place cells in hippocampus. We investigate both kinds of networks analytically and by computer simulations, and we link the models with experimental findings, exploring, in particular, how the spike timing dependence of the synaptic plasticity constrains the computational function of the network and vice versa
Relative contributions of 2D and 3D cues in a texture segmentation task, implications for the roles of striate and extrastriate cortex in attentional selection
No full textExperimental evidence has given strong support to the theory that the primary visual cortex (V1) realizes a bottom–up saliency map (A. R. Koene & L. Zhaoping, 2007; Z. Li, 2002; L. Zhaoping, 2008a; L. Zhaoping & K. A. May, 2007). Unlike the conventional models of texture segmentation, this theory predicted that segmenting two textures in an image Irel comprising obliquely oriented bars would become much more difficult when a task-irrelevant texture Iir of spatially alternating horizontal and vertical bars is superposed on the original texture Irel. The irrelevant texture Iir interferes with Irel's ability to direct attention. This predicted interference was confirmed (L. Zhaoping & K. A. May, 2007) in the form of a prolonged task reaction time (RT). In this study, we investigate whether and how 3D depth perception, believed to be processed mostly beyond V1 and starting in V2 (J. S. Bakin, K. Nakayama, & C. D. Gilbert, 2000; B. G. Cumming & A. J. Parker, 2000; F. T. Qiu & R. von der Heydt, 2005; R. von der Heydt, H. Zhou, & H. S. Friedman, 2000), contribute additionally to direct attention. We measured the reduction of the interference or the RT when the position of the texture grid for Iir was offset horizontally from that for Irel, forming an offset, 2D, stimulus. This reduction was compared with that when this positional offset was only present in the input image to one eye, or when it was in the opposite directions in the images for the two eyes, creating a 3D stimulus with a depth separation between Iir and Irel. The contribution by 3D processes to attentional guidance would be manifested by any extra RT reduction associated with the 3D stimulus over the offset 2D stimulus. This 3D contribution was not present unless the task was so difficult that RT (by button press) based on 2D cues alone was longer than about 1 second. Our findings suggest that, without other top–down factors, V1 plays a dominant role in attentional guidance during an initial window of processing, while cortical areas beyond V1 play an increasing role in later processing. Subject-dependent variations in the manifestations of the 3D effects also suggest that this later, 3D, contribution to attentional guidance can be easily influenced by top–down control
Evaluating the Central-Peripheral Dichotomy in human visual cortex using anatomical and retinotopic data in Human Connectome Project
No full textThe central-peripheral dichotomy (CPD, Zhaoping 2017, 2019) is the hypothesis that central vision is mainly devoted to seeing or recognition, peripheral vision is mainly concerned with looking or attentional selection, and top-down feedback to aid seeing is more strongly directed to the central visual field. CPD has led to behavioral predictions, such as illusions that are stronger or only present in the peripheral visual field, that have subsequently been confirmed (Zhaoping & Ackermann 2018, Zhaoping 2020). CPD’s anatomical prediction that central vision should receive more top-down feedback is consistent with functional connectivity data from human functional magnetic resonance imaging (fMRI) (Sims et al 2021). Here, I evaluate CPD further using data from the Human Connectome Project (Benson et al 2018). Firstly, I examine whether, on average, the proportion of fMRI voxels devoted to central visual field locations increases from V1 to V2, V3, V4, IT, etc, along the ventral visual pathway. The rationale is that higher areas along the (recognition-preferring) ventral stream should be more strongly devoted to seeing within the attentional spotlight rather than looking (by shifting gaze to the attended objects). Hence, CPD predicts increasingly higher proportions of voxels for central vision in higher processing levels along the ventral stream. Secondly, I evaluate whether the physical distance (in 3-dimensional space) between retinotopically corresponding locations in different visual cortical areas is on average shorter for cortical locations representing visual locations having smaller eccentricities. The rationale is that the stronger feedback that CPD predicts for the central field, requires more feedback fibers, which, to save space, should be shorter. Preliminary data analyses support both these predictions of CPD
Pre-attentive segmentation in the primary visual cortex
No full textThe activities of neurons in primary visual cortex have been shown to be significantly influenced by stimuli outside their classical receptive fields. We propose that these contextual influences serve pre-attentive visual segmentation by causing relatively higher neural responses to important or conspicuous image locations, making them more salient for perceptual pop-out. These locations include boundaries between regions, smooth contours, and pop-out targets against backgrounds. The mark of these locations is the breakdown of spatial homogeneity in the input, for instance, at the border between two texture regions of equal mean luminance. This breakdown causes changes in contextual influences, often resulting in higher responses at the border than at surrounding locations. This proposal is implemented in a biologically based model of V1 in which contextual influences are mediated by intra-cortical horizontal connections. The behavior of the model is demonstrated using examples of texture segmentation, figure-ground segregation, target-distractor asymmetry, and contour enhancement, and is compared with psychophysical and physiological data. The model predicts (1) how neural responses should be tuned to the orientation of nearby texture borders, (2) a set of qualitative constraints on the structure of the intracortical connections, and (3) stimulus-dependent biases in estimating the locations of the region borders by pre-attentive vision
- …
