Population coding of visual space: modeling

Cortical field maps across human sensory cortex

Cortical processing pathways for sensory information in the mammalian brain tend to be organized into topographical representations that encode various fundamental sensory dimensions. Numerous laboratories have now shown how these representations are organized into numerous cortical field maps (CMFs) across visual and auditory cortex, with each CFM supporting a specialized computation or set of computations that underlie the associated perceptual behaviors. An individual CFM is defined by two orthogonal topographical gradients that reflect two essential aspects of feature space for that sense. Multiple adjacent CFMs are then organized across visual and auditory cortex into macrostructural patterns termed cloverleaf clusters. CFMs within cloverleaf clusters are thought to share properties such as receptive field distribution, cortical magnification, and processing specialization. Recent measurements point to the likely existence of CFMs in the other senses, as well, with topographical representations of at least one sensory dimension demonstrated in somatosensory, gustatory, and possibly olfactory cortical pathways. Here we discuss the evidence for CFM and cloverleaf cluster organization across human sensory cortex as well as approaches used to identify such organizational patterns. Knowledge of how these topographical representations are organized across cortex provides us with insight into how our conscious perceptions are created from our basic sensory inputs. In addition, studying how these representations change during development, trauma, and disease serves as an important tool for developing improvements in clinical therapies and rehabilitation for sensory deficits.

Effects of normalized summation in the visual illusion of extent

In the present study, the features of summation of effects caused by contextual distracting dots in the length-matching task (a variant of the filled-space illusion) were investigated. In the first two series of psychophysical experiments, the illusion magnitude was measured as a function of the displacement of distractors (either single or double sets of dots) orthogonally to the main axis of the stimulus. It was demonstrated that with increasing displacement, the illusion smoothly decreases for a single set of distractors, while for two sets, the illusion first increases to a certain maximum value, and then gradually decreases. In the third and fourth series of experiments, magnitude of the illusion was measured as a function of the luminance of one set of distracting dots, while the luminance of the other set was fixed. It has been shown that increasing the luminance until the same value is reached for both sets leads to a monotonous growth in the illusion magnitude; after that, the illusion asymptotically decreases to an almost constant level. The theoretical interpretation of the established functional dependencies was performed using a quantitative model based on the assumption that the illusion may arise due to the weighted summation of the distractor-induced normalized neural activity, which leads to the perceptual mislocalization of terminators of stimulus spatial intervals.

Behavioral Prioritization Enhances Working Memory Precision and Neural Population Gain

Humans allocate visual working memory (WM) resource according to behavioral relevance, resulting in more precise memories for more important items. Theoretically, items may be maintained by feature-tuned neural populations, where the relative gain of the populations encoding each item determines precision. To test this hypothesis, we compared the amplitudes of delay period activity in the different parts of retinotopic maps representing each of several WM items, predicting the amplitudes would track behavioral priority. Using fMRI, we scanned participants while they remembered the location of multiple items over a WM delay and then reported the location of one probed item using a memory-guided saccade. Importantly, items were not equally probable to be probed (0.6, 0.3, 0.1, 0.0), which was indicated with a precue. We analyzed fMRI activity in 10 visual field maps in occipital, parietal, and frontal cortex known to be important for visual WM. In early visual cortex, but not association cortex, the amplitude of BOLD activation within voxels corresponding to the retinotopic location of visual WM items increased with the priority of the item. Interestingly, these results were contrasted with a common finding that higher-level brain regions had greater delay period activity, demonstrating a dissociation between the absolute amount of activity in a brain area and the activity of different spatially selective populations within it. These results suggest that the distribution of WM resources according to priority sculpts the relative gains of neural populations that encode items, offering a neural mechanism for how prioritization impacts memory precision.

Pseudosparse neural coding in the visual system of primates

When measuring sparseness in neural populations as an indicator of efficient coding, an implicit assumption is that each stimulus activates a different random set of neurons. In other words, population responses to different stimuli are, on average, uncorrelated. Here we examine neurophysiological data from four lobes of macaque monkey cortex, including V1, V2, MT, anterior inferotemporal cortex, lateral intraparietal cortex, the frontal eye fields, and perirhinal cortex, to determine how correlated population responses are. We call the mean correlation the pseudosparseness index, because high pseudosparseness can mimic statistical properties of sparseness without being authentically sparse. In every data set we find high levels of pseudosparseness ranging from 0.59-0.98, substantially greater than the value of 0.00 for authentic sparseness. This was true for synthetic and natural stimuli, as well as for single-electrode and multielectrode data. A model indicates that a key variable producing high pseudosparseness is the standard deviation of spontaneous activity across the population. Consistently high values of pseudosparseness in the data demand reconsideration of the sparse coding literature as well as consideration of the degree to which authentic sparseness provides a useful framework for understanding neural coding in the cortex.

Extrinsic and intrinsic representations

We extend the discussion in the target article about distinctions between extrinsic coding (external references to known things, as required by information theory) and the alternative we and the target article both favor, intrinsic coding (internal relationships within sensory and motor signals). Central to our thinking about intrinsic coding is population coding and the concept of high-dimensional neural response spaces.

Representation of shape, space, and attention in monkey cortex

Attentional deficits are core to numerous developmental, neurological, and psychiatric disorders. At the single-cell level, much knowledge has been garnered from studies of shape and spatial properties, as well as from numerous demonstrations of attentional modulation of those properties. Despite this wealth of knowledge of single-cell responses across many brain regions, little is known about how these cellular characteristics relate to population level representations and how such representations relate to behavior; in particular, how these cellular responses relate to the representation of shape, space, and attention, and how these representations differ across cortical areas and streams. Here we will emphasize the role of population coding as a missing link for connecting single-cell properties with behavior. Using a data-driven intrinsic approach to population decoding, we show that both 'what' and 'where' cortical visual streams encode shape, space, and attention, yet demonstrate striking differences in these representations. We suggest that both pathways fully process shape and space, but that differences in representation may arise due to their differing functions and input and output constraints. Moreover, differences in the effects of attention on shape and spatial population representations in the two visual streams suggest two distinct strategies: in a ventral area, attention or task demands modulate the population representations themselves (perhaps to expand or enhance one part at the expense of other parts) while in a dorsal area, at a population representation level, attention effects are weak and nearly non-existent, perhaps in order to maintain veridical representations needed for visuomotor control. We show that an intrinsic approach, as opposed to theory-driven and labeled approaches, is useful for understanding how representations develop and differ across brain regions. Most importantly, these approaches help link cellular properties more tightly with behavior, a much-needed step to better understand and interpret cellular findings and key to providing insights to improve interventions in human disorders.

Attention Effects on Neural Population Representations for Shape and Location Are Stronger in the Ventral than Dorsal Stream

We examined how attention causes neural population representations of shape and location to change in ventral stream (AIT) and dorsal stream (LIP). Monkeys performed two identical delayed-match-to-sample (DMTS) tasks, attending either to shape or location. In AIT, shapes were more discriminable when directing attention to shape rather than location, measured by an increase in mean distance between population response vectors. In LIP, attending to location rather than shape did not increase the discriminability of different stimulus locations. Even when factoring out the change in mean vector response distance, multidimensional scaling (MDS) still showed a significant task difference in AIT, but not LIP, indicating that beyond increasing discriminability, attention also causes a nonlinear warping of representation space in AIT. Despite single-cell attentional modulations in both areas, our data show that attentional modulations of population representations are weaker in LIP, likely due to a need to maintain veridical representations for visuomotor control.

Analytic processing of distance

How does a human observer extract from the distance between two frontal points the component corresponding to an axis of a rectangular reference frame? To find out we had participants classify pairs of small circles, varying on the horizontal and vertical axes of a computer screen, in terms of the horizontal distance between them. A response signal controlled response time. The error rate depended on the irrelevant vertical as well as the relevant horizontal distance between the test circles with the relevant distance effect being larger than the irrelevant distance effect. The results implied that the horizontal distance between the test circles was imperfectly extracted from the overall distance between them. The results supported an account, derived from the Exemplar Based Random Walk model (Nosofsky & Palmieri, 1997), under which distance classification is based on the overall distance between the test circles, with relevant distance being extracted from overall distance to the extent that the relevant and irrelevant axes are differentially weighted so as to reduce the contribution of irrelevant distance to overall distance. The results did not support an account, derived from the General Recognition Theory (Ashby & Maddox, 1994), under which distance classification is based on the relevant distance between the test circles, with the irrelevant distance effect arising because a test circle's perceived location on the relevant axis depends on its location on the irrelevant axis, and with relevant distance being extracted from overall distance to the extent that this dependency is absent.

Hierarchical differences in population coding within auditory cortex

Most models of auditory cortical (AC) population coding have focused on primary auditory cortex (A1). Thus our understanding of how neural coding for sounds progresses along the cortical hierarchy remains obscure. To illuminate this, we recorded from two AC fields: A1 and middle lateral belt (ML) of rhesus macaques. We presented amplitude-modulated (AM) noise during both passive listening and while the animals performed an AM detection task ("active" condition). In both fields, neurons exhibit monotonic AM-depth tuning, with A1 neurons mostly exhibiting increasing rate-depth functions and ML neurons approximately evenly distributed between increasing and decreasing functions. We measured noise correlation (r_noise) between simultaneously recorded neurons and found that whereas engagement decreased average r_noise in A1, engagement increased average r_noise in ML. This finding surprised us, because attentive states are commonly reported to decrease average r_noise We analyzed the effect of r_noise on AM coding in both A1 and ML and found that whereas engagement-related shifts in r_noise in A1 enhance AM coding, r_noise shifts in ML have little effect. These results imply that the effect of r_noise differs between sensory areas, based on the distribution of tuning properties among the neurons within each population. A possible explanation of this is that higher areas need to encode nonsensory variables (e.g., attention, choice, and motor preparation), which impart common noise, thus increasing r_noise Therefore, the hierarchical emergence of r_noise-robust population coding (e.g., as we observed in ML) enhances the ability of sensory cortex to integrate cognitive and sensory information without a loss of sensory fidelity.NEW & NOTEWORTHY Prevailing models of population coding of sensory information are based on a limited subset of neural structures. An important and under-explored question in neuroscience is how distinct areas of sensory cortex differ in their population coding strategies. In this study, we compared population coding between primary and secondary auditory cortex. Our findings demonstrate striking differences between the two areas and highlight the importance of considering the diversity of neural structures as we develop models of population coding.

Visual Field Map Clusters in High-Order Visual Processing: Organization of V3A/V3B and a New Cloverleaf Cluster in the Posterior Superior Temporal Sulcus

The cortical hierarchy of the human visual system has been shown to be organized around retinal spatial coordinates throughout much of low- and mid-level visual processing. These regions contain visual field maps (VFMs) that each follows the organization of the retina, with neighboring aspects of the visual field processed in neighboring cortical locations. On a larger, macrostructural scale, groups of such sensory cortical field maps (CFMs) in both the visual and auditory systems are organized into roughly circular cloverleaf clusters. CFMs within clusters tend to share properties such as receptive field distribution, cortical magnification, and processing specialization. Here we use fMRI and population receptive field (pRF) modeling to investigate the extent of VFM and cluster organization with an examination of higher-level visual processing in temporal cortex and compare these measurements to mid-level visual processing in dorsal occipital cortex. In human temporal cortex, the posterior superior temporal sulcus (pSTS) has been implicated in various neuroimaging studies as subserving higher-order vision, including face processing, biological motion perception, and multimodal audiovisual integration. In human dorsal occipital cortex, the transverse occipital sulcus (TOS) contains the V3A/B cluster, which comprises two VFMs subserving mid-level motion perception and visuospatial attention. For the first time, we present the organization of VFMs in pSTS in a cloverleaf cluster. This pSTS cluster contains four VFMs bilaterally: pSTS-1:4. We characterize these pSTS VFMs as relatively small at ∼125 mm² with relatively large pRF sizes of ∼2-8° of visual angle across the central 10° of the visual field. V3A and V3B are ∼230 mm² in surface area, with pRF sizes here similarly ∼1-8° of visual angle across the same region. In addition, cortical magnification measurements show that a larger extent of the pSTS VFM surface areas are devoted to the peripheral visual field than those in the V3A/B cluster. Reliability measurements of VFMs in pSTS and V3A/B reveal that these cloverleaf clusters are remarkably consistent and functionally differentiable. Our findings add to the growing number of measurements of widespread sensory CFMs organized into cloverleaf clusters, indicating that CFMs and cloverleaf clusters may both be fundamental organizing principles in cortical sensory processing.

Two Visual Pathways in Primates Based on Sampling of Space: Exploitation and Exploration of Visual Information

Evidence is strong that the visual pathway is segregated into two distinct streams—ventral and dorsal. Two proposals theorize that the pathways are segregated in function: The ventral stream processes information about object identity, whereas the dorsal stream, according to one model, processes information about either object location, and according to another, is responsible in executing movements under visual control. The models are influential; however recent experimental evidence challenges them, e.g., the ventral stream is not solely responsible for object recognition; conversely, its function is not strictly limited to object vision; the dorsal stream is not responsible by itself for spatial vision or visuomotor control; conversely, its function extends beyond vision or visuomotor control. In their place, we suggest a robust dichotomy consisting of a ventral stream selectively sampling high-resolution/focal spaces, and a dorsal stream sampling nearly all of space with reduced foveal bias. The proposal hews closely to the theme of embodied cognition: Function arises as a consequence of an extant sensory underpinning. A continuous, not sharp, segregation based on function emerges, and carries with it an undercurrent of an exploitation-exploration dichotomy. Under this interpretation, cells of the ventral stream, which individually have more punctate receptive fields that generally include the fovea or parafovea, provide detailed information about object shapes and features and lead to the systematic exploitation of said information; cells of the dorsal stream, which individually have large receptive fields, contribute to visuospatial perception, provide information about the presence/absence of salient objects and their locations for novel exploration and subsequent exploitation by the ventral stream or, under certain conditions, the dorsal stream. We leverage the dichotomy to unify neuropsychological cases under a common umbrella, account for the increased prevalence of multisensory integration in the dorsal stream under a Bayesian framework, predict conditions under which object recognition utilizes the ventral or dorsal stream, and explain why cells of the dorsal stream drive sensorimotor control and motion processing and have poorer feature selectivity. Finally, the model speculates on a dynamic interaction between the two streams that underscores a unified, seamless perception. Existing theories are subsumed under our proposal.

Functional MRI Representational Similarity Analysis Reveals a Dissociation between Discriminative and Relative Location Information in the Human Visual System

Neural responses in visual cortex are governed by a topographic mapping from retinal locations to cortical responses. Moreover, at the voxel population level early visual cortex (EVC) activity enables accurate decoding of stimuli locations. However, in many cases information enabling one to discriminate between locations (i.e., discriminative information) may be less relevant than information regarding the relative location of two objects (i.e., relative information). For example, when planning to grab a cup, determining whether the cup is located at the same retinal location as the hand is hardly relevant, whereas the location of the cup relative to the hand is crucial for performing the action. We have previously used multivariate pattern analysis techniques to measure discriminative location information, and found the highest levels in EVC, in line with other studies. Here we show, using representational similarity analysis, that availability of discriminative information in fMRI activation patterns does not entail availability of relative information. Specifically, we find that relative location information can be reliably extracted from activity patterns in posterior intraparietal sulcus (pIPS), but not from EVC, where we find the spatial representation to be warped. We further show that this variability in relative information levels between regions can be explained by a computational model based on an array of receptive fields. Moreover, when the model's receptive fields are extended to include inhibitory surround regions, the model can account for the spatial warping in EVC. These results demonstrate how size and shape properties of receptive fields in human visual cortex contribute to the transformation of discriminative spatial representations into relative spatial representations along the visual stream.

Characteristics of Eye-Position Gain Field Populations Determine Geometry of Visual Space

We have previously demonstrated differences in eye-position spatial maps for anterior inferotemporal cortex (AIT) in the ventral stream and lateral intraparietal cortex (LIP) in the dorsal stream, based on population decoding of gaze angle modulations of neural visual responses (i.e., eye-position gain fields). Here we explore the basis of such spatial encoding differences through modeling of gain field characteristics. We created a population of model neurons, each having a different eye-position gain field. This population was used to reconstruct eye-position visual space using multidimensional scaling. As gain field shapes have never been well-established experimentally, we examined different functions, including planar, sigmoidal, elliptical, hyperbolic, and mixtures of those functions. All functions successfully recovered positions, indicating weak constraints on allowable gain field shapes. We then used a genetic algorithm to modify the characteristics of model gain field populations until the recovered spatial maps closely matched those derived from monkey neurophysiological data in AIT and LIP. The primary differences found between model AIT and LIP gain fields were that AIT gain fields were more foveally dominated. That is, gain fields in AIT operated on smaller spatial scales and smaller dispersions than in LIP. Thus, we show that the geometry of eye-position visual space depends on the population characteristics of gain fields, and that differences in gain field characteristics for different cortical areas may underlie differences in the representation of space.

Neural representation for object recognition in inferotemporal cortex

We suggest that population representation of objects in inferotemporal cortex lie on a continuum between a purely structural, parts-based description and a purely holistic description. The intrinsic dimensionality of object representation is estimated to be around 100, perhaps with lower dimensionalities for object representations more toward the holistic end of the spectrum. Cognitive knowledge in the form of semantic information and task information feed back to inferotemporal cortex from perirhinal and prefrontal cortex respectively, providing high-level multimodal-based expectations that assist in the interpretation of object stimuli. Integration of object information across eye movements may also contribute to object recognition through a process of active vision.

fMRI of the rod scotoma elucidates cortical rod pathways and implications for lesion measurements

Are silencing, ectopic shifts, and receptive field (RF) scaling in cortical scotoma projection zones (SPZs) the result of long-term reorganization (plasticity) or short-term adaptation? Electrophysiological studies of SPZs after retinal lesions in animal models remain controversial, because they are unable to conclusively answer this question because of limitations of the methodology. Here, we used functional MRI (fMRI) visual field mapping through population RF (pRF) modeling with moving bar stimuli under photopic and scotopic conditions to measure the effects of the rod scotoma in human early visual cortex. As a naturally occurring central scotoma, it has a large cortical representation, is free of traumatic lesion complications, is completely reversible, and has not reorganized under normal conditions (but can as seen in rod monochromats). We found that the pRFs overlapping the SPZ in V1, V2, V3, hV4, and VO-1 generally (i) reduced their blood oxygen level-dependent signal coherence and (ii) shifted their pRFs more eccentric but (iii) scaled their pRF sizes in variable ways. Thus, silencing, ectopic shifts, and pRF scaling in SPZs are not unique identifiers of cortical reorganization; rather, they can be the expected result of short-term adaptation. However, are there differences between rod and cone signals in V1, V2, V3, hV4, and VO-1? We did not find differences for all five maps in more peripheral eccentricities outside of rod scotoma influence in coherence, eccentricity representation, or pRF size. Thus, rod and cone signals seem to be processed similarly in cortex.

Recovering stimulus locations using populations of eye-position modulated neurons in dorsal and ventral visual streams of non-human primates

We recorded visual responses while monkeys fixated the same target at different gaze angles, both dorsally (lateral intraparietal cortex, LIP) and ventrally (anterior inferotemporal cortex, AIT). While eye-position modulations occurred in both areas, they were both more frequent and stronger in LIP neurons. We used an intrinsic population decoding technique, multidimensional scaling (MDS), to recover eye positions, equivalent to recovering fixated target locations. We report that eye-position based visual space in LIP was more accurate (i.e., metric). Nevertheless, the AIT spatial representation remained largely topologically correct, perhaps indicative of a categorical spatial representation (i.e., a qualitative description such as "left of" or "above" as opposed to a quantitative, metrically precise description). Additionally, we developed a simple neural model of eye position signals and illustrate that differences in single cell characteristics can influence the ability to recover target position in a population of cells. We demonstrate for the first time that the ventral stream contains sufficient information for constructing an eye-position based spatial representation. Furthermore we demonstrate, in dorsal and ventral streams as well as modeling, that target locations can be extracted directly from eye position signals in cortical visual responses without computing coordinate transforms of visual space.

Overlapping structures in sensory-motor mappings

This paper examines a biologically-inspired representation technique designed for the support of sensory-motor learning in developmental robotics. An interesting feature of the many topographic neural sheets in the brain is that closely packed receptive fields must overlap in order to fully cover a spatial region. This raises interesting scientific questions with engineering implications: e.g. is overlap detrimental? does it have any benefits? This paper examines the effects and properties of overlap between elements arranged in arrays or maps. In particular we investigate how overlap affects the representation and transmission of spatial location information on and between topographic maps. Through a series of experiments we determine the conditions under which overlap offers advantages and identify useful ranges of overlap for building mappings in cognitive robotic systems. Our motivation is to understand the phenomena of overlap in order to provide guidance for application in sensory-motor learning robots.

Monkeys in space: Primate neural data suggest volumetric representations

The target article does not consider neural data on primate spatial representations, which we suggest provide grounds for believing that navigational space may be three-dimensional rather than quasi–two-dimensional. Furthermore, we question the authors' interpretation of rat neurophysiological data as indicating that the vertical dimension may be encoded in a neural structure separate from the two horizontal dimensions.

When the body is time: spatial and temporal deixis in children with visual impairments and sighted children

While there is mounting evidence explaining how concrete concepts are processed, the evidence demonstrating how abstract concepts are processed is rather scant. Most research illustrating how concrete and abstract concepts are processed has been obtained from adult populations. Consequently, not much is known about how these concepts are processed by children, especially those with sensorimotor impairments. This paper reports a study in which groups of children who were either visual-motor impaired (VMG), blind (BG), or sighted (CG) were requested to perform deictic gestures for temporal and spatial concepts. The results showed that: (i) spatial pointing was performed faster than temporal pointing across all groups of children; (ii) such difference in pointing times occurred also within groups; and (iii) the slowest pointing times were those of the blind children followed by the VMG and the CG children, respectively. Additionally, while CG children correctly performed the pointing tasks, VMG and, particularly, BG children relied on a form of deixis known as autotopological (or personal) deixis. The results thus suggest that deprivation or lack of sensorimotor experience with the environment affects the processing of abstract concepts and that a compensatory mechanism may be to rely on the body as a reference frame.

Statistics of visual responses in primate inferotemporal cortex to object stimuli

We have characterized selectivity and sparseness in anterior inferotemporal cortex, using a large data set. Responses were collected from 674 monkey inferotemporal cells, each stimulated by 806 object photographs. This 806 × 674 matrix was examined in two ways: columnwise, looking at responses of a single neuron to all images (single-neuron selectivity), and rowwise, looking at the responses of all neurons caused by a single image (population sparseness). Selectivity and sparseness were measured as kurtosis of probability distributions. Population sparseness exceeded single-neuron selectivity, with specific values dependent on the size of the data sample. This difference was principally caused by inclusion, within the population, of neurons with a variety of dynamic ranges (standard deviations of responses over all images). Statistics of large responses were examined by quantifying how quickly the upper tail of the probability distribution decreased (tail heaviness). This analysis demonstrated that population responses had heavier tails than single-neuron responses, consistent with the difference between sparseness and selectivity measurements. Population responses with spontaneous activity subtracted had the heaviest tails, following a power law. The very light tails of single-neuron responses indicate that the critical feature for each neuron is simple enough to have a high probability of occurring within a limited stimulus set. Heavy tails of population responses indicate that there are a large number of different critical features to which different neurons are tuned. These results are inconsistent with some structural models of object recognition that posit that objects are decomposed into a small number of standard features.

Population coding of visual space: comparison of spatial representations in dorsal and ventral pathways

Although the representation of space is as fundamental to visual processing as the representation of shape, it has received relatively little attention from neurophysiological investigations. In this study we characterize representations of space within visual cortex, and examine how they differ in a first direct comparison between dorsal and ventral subdivisions of the visual pathways. Neural activities were recorded in anterior inferotemporal cortex (AIT) and lateral intraparietal cortex (LIP) of awake behaving monkeys, structures associated with the ventral and dorsal visual pathways respectively, as a stimulus was presented at different locations within the visual field. In spatially selective cells, we find greater modulation of cell responses in LIP with changes in stimulus position. Further, using a novel population-based statistical approach (namely, multidimensional scaling), we recover the spatial map implicit within activities of neural populations, allowing us to quantitatively compare the geometry of neural space with physical space. We show that a population of spatially selective LIP neurons, despite having large receptive fields, is able to almost perfectly reconstruct stimulus locations within a low-dimensional representation. In contrast, a population of AIT neurons, despite each cell being spatially selective, provide less accurate low-dimensional reconstructions of stimulus locations. They produce instead only a topologically (categorically) correct rendition of space, which nevertheless might be critical for object and scene recognition. Furthermore, we found that the spatial representation recovered from population activity shows greater translation invariance in LIP than in AIT. We suggest that LIP spatial representations may be dimensionally isomorphic with 3D physical space, while in AIT spatial representations may reflect a more categorical representation of space (e.g., "next to" or "above").