Physiological differences between neurons in layer 2 and layer 3 of primary visual cortex (V1) of alert macaque monkeys

Computational components of visual predictive coding circuitry

If a full visual percept can be said to be a 'hypothesis', so too can a neural 'prediction' - although the latter addresses one particular component of image content (such as 3-dimensional organisation, the interplay between lighting and surface colour, the future trajectory of moving objects, and so on). And, because processing is hierarchical, predictions generated at one level are conveyed in a backward direction to a lower level, seeking to predict, in fact, the neural activity at that prior stage of processing, and learning from errors signalled in the opposite direction. This is the essence of 'predictive coding', at once an algorithm for information processing and a theoretical basis for the nature of operations performed by the cerebral cortex. Neural models for the implementation of predictive coding invoke specific functional classes of neuron for generating, transmitting and receiving predictions, and for producing reciprocal error signals. Also a third general class, 'precision' neurons, tasked with regulating the magnitude of error signals contingent upon the confidence placed upon the prediction, i.e., the reliability and behavioural utility of the sensory data that it predicts. So, what is the ultimate source of a 'prediction'? The answer is multifactorial: knowledge of the current environmental context and the immediate past, allied to memory and lifetime experience of the way of the world, doubtless fine-tuned by evolutionary history too. There are, in consequence, numerous potential avenues for experimenters seeking to manipulate subjects' expectation, and examine the neural signals elicited by surprising, and less surprising visual stimuli. This review focuses upon the predictive physiology of mouse and monkey visual cortex, summarising and commenting on evidence to date, and placing it in the context of the broader field. It is concluded that predictive coding has a firm grounding in basic neuroscience and that, unsurprisingly, there remains much to learn.

Excitatory and inhibitory effects of HCN channel modulation on excitability of layer V pyramidal cells

Dendrites of cortical pyramidal cells are densely populated by hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, a.k.a. Ih channels. Ih channels are targeted by multiple neuromodulatory pathways, and thus are one of the key ion-channel populations regulating the pyramidal cell activity. Previous observations and theories attribute opposing effects of the Ih channels on neuronal excitability due to their mildly hyperpolarized reversal potential. These effects are difficult to measure experimentally due to the fine spatiotemporal landscape of the Ih activity in the dendrites, but computational models provide an efficient tool for studying this question in a reduced but generalizable setting. In this work, we build upon existing biophysically detailed models of thick-tufted layer V pyramidal cells and model the effects of over- and under-expression of Ih channels as well as their neuromodulation. We show that Ih channels facilitate the action potentials of layer V pyramidal cells in response to proximal dendritic stimulus while they hinder the action potentials in response to distal dendritic stimulus at the apical dendrite. We also show that the inhibitory action of the Ih channels in layer V pyramidal cells is due to the interactions between Ih channels and a hot zone of low voltage-activated Ca2+ channels at the apical dendrite. Our simulations suggest that a combination of Ih-enhancing neuromodulation at the proximal part of the apical dendrite and Ih-inhibiting modulation at the distal part of the apical dendrite can increase the layer V pyramidal excitability more than either of the two alone. Our analyses uncover the effects of Ih-channel neuromodulation of layer V pyramidal cells at a single-cell level and shed light on how these neurons integrate information and enable higher-order functions of the brain.

Functional and structural features of L2/3 pyramidal cells continuously covary with pial depth in mouse visual cortex

Pyramidal cells of neocortical layer 2/3 (L2/3 PyrCs) integrate signals from numerous brain areas and project throughout the neocortex. These PyrCs show pial depth-dependent functional and structural specializations, indicating participation in different functional microcircuits. However, whether these depth-dependent differences result from separable PyrC subtypes or whether their features display a continuum correlated with pial depth is unknown. Here, we assessed the stimulus selectivity, electrophysiological properties, dendritic morphology, and excitatory and inhibitory connectivity across the depth of L2/3 in the binocular visual cortex of mice. We find that the apical, but not the basal dendritic tree structure, varies with pial depth, which is accompanied by variation in subthreshold electrophysiological properties. Lower L2/3 PyrCs receive increased input from L4, while upper L2/3 PyrCs receive a larger proportion of intralaminar input. In vivo calcium imaging revealed a systematic change in visual responsiveness, with deeper PyrCs showing more robust responses than superficial PyrCs. Furthermore, deeper PyrCs are more driven by contralateral than ipsilateral eye stimulation. Importantly, the property value transitions are gradual, and L2/3 PyrCs do not display discrete subtypes based on these parameters. Therefore, L2/3 PyrCs' multiple functional and structural properties systematically correlate with their depth, forming a continuum rather than discrete subtypes.

Primate neuronal connections are sparse in cortex as compared to mouse

Detailing how primate and mouse neurons differ is critical for creating generalized models of how neurons process information. We reconstruct 15,748 synapses in adult Rhesus macaques and mice and ask how connectivity differs on identified cell types in layer 2/3 of primary visual cortex. Primate excitatory and inhibitory neurons receive 2-5 times fewer excitatory and inhibitory synapses than similar mouse neurons. Primate excitatory neurons have lower excitatory-to-inhibitory (E/I) ratios than mouse but similar E/I ratios in inhibitory neurons. In both species, properties of inhibitory axons such as synapse size and frequency are unchanged, and inhibitory innervation of excitatory neurons is local and specific. Using artificial recurrent neural networks (RNNs) optimized for different cognitive tasks, we find that penalizing networks for creating and maintaining synapses, as opposed to neuronal firing, reduces the number of connections per node as the number of nodes increases, similar to primate neurons compared with mice.

Functional organization of spatial frequency tuning in macaque V1 revealed with two-photon calcium imaging

V1 neurons are functionally organized in orientation columns in primates. Whether spatial frequency (SF) columns also exist is less clear because mixed results have been reported. A definitive solution would be SF functional maps at single-neuron resolution. Here we used two-photon calcium imaging to construct first cellular SF maps in V1 superficial layers of five awake fixating macaques, and studied SF functional organization properties and neuronal tuning characteristics. The SF maps (850 × 850 μm²) showed weak horizontal SF clustering (median clustering index = 1.43 vs. unity baseline), about one sixth as strong as orientation clustering in the same sets of neurons, which argues against a meaningful orthogonal relationship between orientation and SF functional maps. These maps also displayed nearly absent vertical SF clustering between two cortical depths (150 & 300 μm), indicating a lack of SF columnar structures within the superficial layers. The underlying causes might be that most neurons were tuned to a narrow two-octave range of medium frequencies, and many neurons with different SF preferences were often spatially mixed, which disallowed finer grouping of SF tuning. In addition, individual SF tuning functions were often asymmetric, having wider lower frequency branches, which may help encode low SF information for later decoding.

Anatomy and Physiology of Macaque Visual Cortical Areas V1, V2, and V5/MT: Bases for Biologically Realistic Models

The cerebral cortex of primates encompasses multiple anatomically and physiologically distinct areas processing visual information. Areas V1, V2, and V5/MT are conserved across mammals and are central for visual behavior. To facilitate the generation of biologically accurate computational models of primate early visual processing, here we provide an overview of over 350 published studies of these three areas in the genus Macaca, whose visual system provides the closest model for human vision. The literature reports 14 anatomical connection types from the lateral geniculate nucleus of the thalamus to V1 having distinct layers of origin or termination, and 194 connection types between V1, V2, and V5, forming multiple parallel and interacting visual processing streams. Moreover, within V1, there are reports of 286 and 120 types of intrinsic excitatory and inhibitory connections, respectively. Physiologically, tuning of neuronal responses to 11 types of visual stimulus parameters has been consistently reported. Overall, the optimal spatial frequency (SF) of constituent neurons decreases with cortical hierarchy. Moreover, V5 neurons are distinct from neurons in other areas for their higher direction selectivity, higher contrast sensitivity, higher temporal frequency tuning, and wider SF bandwidth. We also discuss currently unavailable data that could be useful for biologically accurate models.

Two distinct profiles of fMRI and neurophysiological activity elicited by acetylcholine in visual cortex

fMRI changes are typically assumed to be due to changes in neural activity, although whether this remains valid under the influence of neuromodulators is relatively unknown. Here, we found evidence that intracortical acetylcholine elicits distinct profiles of fMRI and electrophysiological activity in visual cortex. Two patterns of cholinergic activity were observed, depending on the distance to the injection site, although neurovascular coupling was preserved. Our results illustrate the effects of neuromodulators on fMRI and electrophysiological responses and show that these depend on neuromodulator concentration and kinetics.

Cholinergic neuromodulation is involved in all aspects of sensory processing and is crucial for processes such as attention, learning and memory, etc. However, despite the known roles of acetylcholine (ACh), we still do not how to disentangle ACh contributions from sensory or task-evoked changes in functional magnetic resonance imaging (fMRI). Here, we investigated the effects of local injection of ACh on fMRI and neural signals in the primary visual cortex (V1) of anesthetized macaques by combining pharmaco-based MRI (phMRI) with electrophysiological recordings, using single electrodes and electrode arrays. We found that local injection of ACh elicited two distinct profiles of fMRI and neurophysiological activity, depending on the distance from the injector. Near the injection site, we observed an increase in the baseline blood oxygen-level-dependent (BOLD) and cerebral blood flow (CBF) responses, while their visual modulation decreased. In contrast, further from the injection site, we observed an increase in the visually induced BOLD and CBF modulation without changes in baseline. Neurophysiological recordings suggest that the spatial correspondence between fMRI responses and neural activity does not change in the gamma, high-gamma, and multiunit activity (MUA) bands. The results near the injection site suggest increased inhibitory drive and decreased metabolism, contrasting to the far region. These changes are thought to reflect the kinetics of ACh and its metabolism to choline.

Apical Function in Neocortical Pyramidal Cells: A Common Pathway by Which General Anesthetics Can Affect Mental State

It has been argued that general anesthetics suppress the level of consciousness, or the contents of consciousness, or both. The distinction between level and content is important because, in addition to clarifying the mechanisms of anesthesia, it may help clarify the neural bases of consciousness. We assess these arguments in the light of evidence that both the level and the content of consciousness depend upon the contribution of apical input to the information processing capabilities of neocortical pyramidal cells which selectively amplify relevant signals. We summarize research suggesting that what neocortical pyramidal cells transmit information about can be distinguished from levels of arousal controlled by sub-cortical nuclei and from levels of prioritization specified by interactions within the thalamocortical system. Put simply, on the basis of the observations reviewed, we hypothesize that when conscious we have particular, directly experienced, percepts, thoughts, feelings and intentions, and that general anesthetics affect consciousness by interfering with the subcellular processes by which particular activities are selectively amplified when relevant to the current context.

Slot-like capacity and resource-like coding in a neural model of multiple-item working memory

For the past decade, research on the storage limitations of working memory has been dominated by two fundamentally different hypotheses. On the one hand, the contents of working memory may be stored in a limited number of "slots," each with a fixed resolution. On the other hand, any number of items may be stored but with decreasing resolution. These two hypotheses have been invaluable in characterizing the computational structure of working memory, but neither provides a complete account of the available experimental data or speaks to the neural basis of the limitations it characterizes. To address these shortcomings, we simulated a multiple-item working memory task with a cortical network model, the cellular resolution of which allowed us to quantify the coding fidelity of memoranda as a function of memory load, as measured by the discriminability, regularity, and reliability of simulated neural spiking. Our simulations account for a wealth of neural and behavioral data from human and nonhuman primate studies, and they demonstrate that feedback inhibition lowers both capacity and coding fidelity. Because the strength of inhibition scales with the number of items stored by the network, increasing this number progressively lowers fidelity until capacity is reached. Crucially, the model makes specific, testable predictions for neural activity on multiple-item working memory tasks. NEW & NOTEWORTHY Working memory is the ability to keep information in mind and is fundamental to cognition. It is actively debated whether the storage limitations of working memory reflect a small number of storage units (slots) or a decrease in coding resolution as a limited resource is allocated to more items. In a cortical model, we found that slot-like capacity and resource-like neural coding resulted from the same mechanism, offering an integrated explanation for storage limitations.

Very small faces are easily discriminated under long and short exposure times

Acuity measures related to overall face size that can be perceived have not been studied quantitatively. Consequently, experimenters use a wide range of sizes (usually large) without always providing a rationale for their choices. I studied thresholds for face discrimination by presenting both long (500 ms)- and short (17, 33, 50 ms)-duration stimuli. Face width threshold for the long presentation was ~0.2°, and thresholds for the flashed stimuli ranged from ~0.3° for the 17-ms flash to ~0.23° for the 33- and 50-ms flashes. Such thresholds indicate that face stimuli used in physiological or psychophysical experiments are often too large to tap human fine spatial capabilities, and thus interpretations of such experiments should take into account face discrimination acuity. The 0.2° threshold found in this study is incompatible with the prevalent view that faces are represented by a population of specialized “face cells” because those cells do not respond to <1° stimuli and are optimally tuned to >4° faces. Also, the ability to discriminate small, high-spatial frequency flashed face stimuli is inconsistent with models suggesting that fixational drift transforms retinal spatial patterns into a temporal code. It seems therefore that the small image motions occurring during fixation do not disrupt our perception, because all relevant processing is over with before those motions can have significant effects.

NEW & NOTEWORTHY Although face perception is central to human behavior, the minimally perceived face size is not known. This study shows that humans can discriminate very small (~0.2°) faces. Furthermore, even when flashed for tens of milliseconds, ~0.25° faces can be discriminated. Such fine acuity should impact modeling of physiological mechanisms of face perception. The ability to discriminate flashed faces where there is almost no eye movement indicates that eye drift is not essential for visibility.

Comparison of the Upper Marginal Neurons of Cortical Layer 2 with Layer 2/3 Pyramidal Neurons in Mouse Temporal Cortex

Layer 2/3 (L2/3) excitatory neurons in the neocortex make major contributions to corticocortical connections and therefore function to integrate information across cortical areas and hemispheres. Recent evidence suggests that excitatory neurons in L2/3 can have different properties. Sparse evidence from previous studies suggests that L2 neurons located at the border between L1 and L2 (referred to as L2 marginal neurons, L2MNs), have a morphology distinct from a typical pyramidal neuron. However, whether the membrane properties and input/output properties of L2MNs are different from those of typical pyramidal neurons in L2/3 is unknown. Here we addressed these questions in a slice preparation of mouse temporal cortex. We found that L2MNs were homogeneous in intrinsic membrane properties but appeared diverse in morphology. In agreement with previous studies, L2MNs either had oblique apical dendrites or had no obvious apical dendrites. The tufts of both apical and basal dendrites of these neurons invaded L1 extensively. All L2MNs showed a regular firing pattern with moderate adaptation. Compared with typical L2/3 pyramidal neurons that showed regular spiking (RS) activity (neurons), L2MNs showed a higher firing rate, larger sag ratio, and higher input resistance. No difference in the amplitude of excitatory and inhibitory postsynaptic potentials (EPSPs and IPSPs, respectively), evoked by stimulation of L1, was found between the two types of neurons, but the IPSPs in L2MNs had a slower time course than those in L2/3 RS cells. In paired recordings, unitary EPSPs showed no significant differences between synapses formed by L2MNs and those formed by L2/3 RS neurons. However, short-term synaptic depression (STSD) examined with a L2MN as the presynaptic neuron was greater when another L2MN was the postsynaptic neuron than when a L2/3 RS neuron was the postsynaptic neuron. The distinct morphological features of L2MNs found here have developmental implications, and the differences in electrophysiological properties between L2MNs and other L2/3 pyramidal neurons suggest that they play different functional roles in cortical networks.

Cell Type-Specific Structural Organization of the Six Layers in Rat Barrel Cortex

The cytoarchitectonic subdivision of the neocortex into six layers is often used to describe the organization of the cortical circuitry, sensory-evoked signal flow or cortical functions. However, each layer comprises neuronal cell types that have different genetic, functional and/or structural properties. Here, we reanalyze structural data from some of our recent work in the posterior-medial barrel-subfield of the vibrissal part of rat primary somatosensory cortex (vS1). We quantify the degree to which somata, dendrites and axons of the 10 major excitatory cell types of the cortex are distributed with respect to the cytoarchitectonic organization of vS1. We show that within each layer, somata of multiple cell types intermingle, but that each cell type displays dendrite and axon distributions that are aligned to specific cytoarchitectonic landmarks. The resultant quantification of the structural composition of each layer in terms of the cell type-specific number of somata, dendritic and axonal path lengths will aid future studies to bridge between layer- and cell type-specific analyses.

The anatomical and physiological properties of the visual cortex argue against cognitive penetration

We are consciously aware of visual objects together with the minute details that characterize each object. Those details are perceived instantaneously and in parallel. V1 is the only visual area with spatial resolution and topographical exactitude matching perceptual abilities. For cognition to penetrate perception, it needs to affect V1 image representation. That is unlikely because of the detailed parallel V1 organization and the nature of top-down connections, which can influence only large parts of the visual field.

Cognitive functions of intracellular mechanisms for contextual amplification

Evidence for the hypothesis that input to the apical tufts of neocortical pyramidal cells plays a central role in cognition by amplifying their responses to feedforward input is reviewed. Apical tufts are electrically remote from the soma, and their inputs come from diverse sources including direct feedback from higher cortical regions, indirect feedback via the thalamus, and long-range lateral connections both within and between cortical regions. This suggests that input to tuft dendrites may amplify the cell's response to basal inputs that they receive via layer 4 and which have synapses closer to the soma. ERP data supporting this inference is noted. Intracellular studies of apical amplification (AA) and of disamplification by inhibitory interneurons targeted only at tufts are reviewed. Cognitive processes that have been related to them by computational, electrophysiological, and psychopathological studies are then outlined. These processes include: figure-ground segregation and Gestalt grouping; contextual disambiguation in perception and sentence comprehension; priming; winner-take-all competition; attention and working memory; setting the level of consciousness; cognitive control; and learning. It is argued that theories in cognitive neuroscience should not assume that all neurons function as integrate-and-fire point processors, but should use the capabilities of cells with distinct sites of integration for driving and modulatory inputs. Potentially 'unifying' theories that depend upon these capabilities are reviewed. It is concluded that evolution of the primitives of AA and disamplification in neocortex may have extended cognitive capabilities beyond those built from the long-established primitives of excitation, inhibition, and disinhibition.

Spatial diversity of spontaneous activity in the cortex

The neocortex is a layered sheet across which a basic organization is thought to widely apply. The variety of spontaneous activity patterns is similar throughout the cortex, consistent with the notion of a basic cortical organization. However, the basic organization is only an outline which needs adjustments and additions to account for the structural and functional diversity across cortical layers and areas. Such diversity suggests that spontaneous activity is spatially diverse in any particular behavioral state. Accordingly, this review summarizes the laminar and areal diversity in cortical activity during fixation and slow oscillations, and the effects of attention, anesthesia and plasticity on the cortical distribution of spontaneous activity. Among questions that remain open, characterizing the spatial diversity in spontaneous membrane potential may help elucidate how differences in circuitry among cortical regions supports their varied functions. More work is also needed to understand whether cortical spontaneous activity not only reflects cortical circuitry, but also contributes to determining the outcome of plasticity, so that it is itself a factor shaping the functional diversity of the cortex.

A physiological perspective on fixational eye movements

For a behavioral neuroscientist, fixational eye movements are a double-edged sword. On one edge, they make control of visual stimuli difficult, but on the other edge they provide insight into the ways the visual system acquires information from the environment. We have studied macaque monkeys as models for human visual systems. Fixational eye movements of monkeys are similar to those of humans but they are more often vertically biased and spatially more dispersed. Eye movements scatter stimuli from their intended retinal locations, increase variability of neuronal responses, inflate estimates of receptive field size, and decrease measures of response amplitude. They also bias against successful stimulation of extremely selective cells. Compensating for eye movements reduced these errors and revealed a fine-grained motion pathway from V1 feeding the cortical ventral stream. Compensation is a useful tool for the experimenter, but rather than compensating for eye movements, the brain utilizes them as part of its input. The saccades and drifts that occur during fixation selectively activate different types of V1 neurons. Cells that prefer slower speeds respond during the drift periods with maintained discharges and tend to have smaller receptive fields that are selective for sign of contrast. They are well suited to code small details of the image and to enable our fine detailed vision. Cells that prefer higher speeds fire transient bursts of spikes when the receptive field leaves, crosses, or lands on a stimulus, but only the most transient ones (about one-third of our sample) failed to respond during drifts. Voluntary and fixational saccades had very similar effects, including the presence of a biphasic extraretinal modulation that interacted with stimulus-driven responses. Saccades evoke synchronous bursts that can enhance visibility but these bursts may also participate in the visual masking that contributes to saccadic suppression. Study of the small eye movements of fixation may illuminate some of the big problems in vision.

Relationship between the local structure of orientation map and the strength of orientation tuning of neurons in monkey V1: a 2-photon calcium imaging study

A majority of neurons in the monkey primary visual cortex (V1) are tuned to stimulus orientations. Preferred orientations and tuning strengths vary among V1 neurons. The preferred orientation of neurons gradually changes across the cortex with occasional failures of this organization. How V1 neurons are arranged by the strength of orientation tuning and whether neuronal arrangement for tuning strength relates to orientation preference maps remains controversial. In this study, we performed in vivo two-photon calcium imaging in macaque V1 to examine the local spatial organization of orientation tuning at the level of single cells. We recorded fluorescence signals from individual neurons loaded with a calcium-sensitive dye in layer 2 and the uppermost tier of layer 3. The strength of orientation tuning was shared by nearby neurons, and changed across the cortex. The neurons with similar tuning strength were distributed across at least the entire thickness of layer 2. The tuning strength was weaker in regions where neurons exhibited heterogeneous preferred orientations, as compared with regions where neurons shared similar orientation preferences. Nearby direction-selective neurons often shared their preferred directions, although only a few neurons were direction selective in the layers examined. Thus, the orientation tuning strength of V1 neurons is partially predictable from the local structure of orientation map. The weaker orientation tuning we found in regions with heterogeneous orientation preferences suggests that orientation-independent interactions among local populations of V1 neurons play a critical role in determining their orientation tuning.

Inhibitory interneurons in a cortical column form hot zones of inhibition in layers 2 and 5A

Although physiological data on microcircuits involving a few inhibitory neurons in the mammalian cerebral cortex are available, data on the quantitative relation between inhibition and excitation in cortical circuits involving thousands of neurons are largely missing. Because the distribution of neurons is very inhomogeneous in the cerebral cortex, it is critical to map all neurons in a given volume rather than to rely on sparse sampling methods. Here, we report the comprehensive mapping of interneurons (INs) in cortical columns of rat somatosensory cortex, immunolabeled for neuron-specific nuclear protein and glutamate decarboxylase. We found that a column contains ~2,200 INs (11.5% of ~19,000 neurons), almost a factor of 2 less than previously estimated. The density of GABAergic neurons was inhomogeneous between layers, with peaks in the upper third of L2/3 and in L5A. IN density therefore defines a distinct layer 2 in the sensory neocortex. In addition, immunohistochemical markers of IN subtypes were layer-specific. The "hot zones" of inhibition in L2 and L5A match the reported low stimulus-evoked spiking rates of excitatory neurons in these layers, suggesting that these inhibitory hot zones substantially suppress activity in the neocortex.

Four projection streams from primate V1 to the cytochrome oxidase stripes of V2

In the primate visual system, areas V1 and V2 distribute information they receive from the retina to all higher cortical areas, sorting this information into dorsal and ventral streams. Therefore, knowledge of the organization of projections between V1 and V2 is crucial to understand how the cortex processes visual information. In primates, parallel output pathways from V1 project to distinct V2 stripes. The traditional tripartite division of V1-to-V2 projections was recently replaced by a bipartite scheme, in which thin stripes receive V1 inputs from blob columns, and thick and pale stripes receive common input from interblob columns. Here, we demonstrate that thick and pale stripes, instead, receive spatially segregated V1 inputs and that the interblob is partitioned into two compartments: the middle of the interblob projecting to pale stripes and the blob/interblob border region projecting to thick stripes. Double-labeling experiments further demonstrate that V1 cells project to either thick or pale stripes, but rarely to both. We also find laminar specialization of V1 outputs, with layer 4B contributing projections mainly to thick stripes, and no projections to one set of pale stripes. These laminar differences suggest different contribution of magno, parvo, and konio inputs to each V1 output pathway. These results provide a new foundation for parallel processing models of the visual system by demonstrating four V1-to-V2 pathways: blob columns-to-thin stripes, blob/interblob border columns-to-thick stripes, interblob columns-to-pale(lateral) stripes, layer 2/3-4A interblobs-to-pale(medial) stripes.

Cortext: a columnar model of bottom-up and top-down processing in the neocortex

Experimental data suggests that a first hypothesis about the content of a complex visual scene is available as early as 150 ms after stimulus presentation. Other evidence suggests that recognition in the visual cortex of mammals is a bidirectional, often top-down driven process. Here, we present a spiking neural network model that demonstrates how the cortex can use both strategies: Faced with a new stimulus, the cortex first tries to catch the gist of the scene. The gist is then fed back as global hypothesis to influence and redirect further bottom-up processing. We propose that these two modes of processing are carried out in different layers of the cortex. A cortical column may, thus, be primarily defined by the specific connectivity that links neurons in different layers into a functional circuit. Given an input, our model generates an initial hypothesis after only a few milliseconds. The first wave of action potentials traveling up the hierarchy activates representations of features and feature combinations. In most cases, the correct feature representation is activated strongest and precedes all other candidates with millisecond precision. Thus, our model codes the reliability of a response in the relative latency of spikes. In the subsequent refinement stage where high-level activity modulates lower stages, this activation dominance is propagated back, influencing its own afferent activity to establish a unique decision. Thus, top-down influence de-activates representations that have contributed to the initial hypothesis about the current stimulus, comparable to predictive coding. Features that do not match the top-down prediction trigger an error signal that can be the basis for learning new representations.

Paraneoplastic antigen-like 5 gene (PNMA5) is preferentially expressed in the association areas in a primate specific manner

To understand the relationship between the structure and function of primate neocortical areas at a molecular level, we have been screening for genes differentially expressed across macaque neocortical areas by restriction landmark cDNA scanning (RLCS). Here, we report enriched expression of the paraneoplastic antigen-like 5 gene (PNMA5) in association areas but not in primary sensory areas, with the lowest expression level in primary visual cortex. In situ hybridization in the primary sensory areas revealed PNMA5 mRNA expression restricted to layer II. Along the ventral visual pathway, the expression gradually increased in the excitatory neurons from the primary to higher visual areas. This differential expression pattern was very similar to that of retinol-binding protein (RBP) mRNA, another association-area-enriched gene that we reported previously. Additional expression analysis for comparison of other genes in the PNMA gene family, PNMA1, PNMA2, PNMA3, and MOAP1 (PNMA4), showed that they were widely expressed across areas and layers but without the differentiated pattern of PNMA5. In mouse brains, PNMA1 was only faintly expressed and PNMA5 was not detected. Sequence analysis showed divergence of PNMA5 sequences among mammals. These findings suggest that PNMA5 acquired a certain specialized role in the association areas of the neocortex during primate evolution.

Feature binding in the feedback layers of area V2

The visual features of an object are processed by multiple, functionally specialized areas of cerebral cortex. When several objects are seen simultaneously, what mechanism preserves the association of features that belong to a single item? We address this question-known as the "binding problem"-by examining combinatorial feature selectivity of neurons in area V2. In recording from anesthetized macaques, we estimate that dual selectivity for chromatic and spatiotemporal attributes is 50% more common (27% vs. 18% sampling frequency) in superficial and deep layer neurons receiving feedback connections from higher areas, compared with layer 4-3 neurons relaying ascending signals. The operation of feedback pathways is thought to mediate attentional modulation of activity that may achieve binding through acting to select one single object for higher representation and filtering out competing objects. We propose that dual-selective neurons perform a "bridging" function, mediating the transfer of feedback-induced bias between feature dimensions. The bias can be propagated through V2 output neurons (of unitary selectivity) to higher levels of specialized processing and so promote selection of the target object's representation among multiple feature maps. The bridging function would thus act to unify the outcome of parallel, object-selective processes taking place along segregated visual pathways.