A computational model for the primate neocortex based on its functional architecture

Experimental evidence has shown that the primate neocortex consists in the main of a set of cortical regions which form a perception hierarchy, an action hierarchy and connections between them. By using a computer science analysis, we develop a computational architecture for the brain in which each cortical region is represented by a computational module with processing and storage abilities. Modules are interconnected according to the connectivity of the corresponding cortical regions. We develop computational principles for designing such a hierarchical and parallel computing system. We demonstrate this approach by proposing a causal functioning model of the brain. We report on results obtained with an implementation of this model. We conclude with a brief discussion of some consequences and predictions of our work.     

Benefits of hierarchical generalization and storage of the representations of "object-place" associations in the hippocampal subfields (a hypothesis)

We analyzed the results of experimental research of features of processing sensory information in the hippocampus and neocortex available in literature and results of modelling the perception of information in the neocortex. It is noted that "place" fields of neurons become wider, and overlapping of receptive fields increases during upward moving in trisynaptic hippocampal pathway. These effects specify the generalization of the information processed. The results of our analysis allow us to put forward a hypothesis that a hierarchical complication of"object - place" associations occurs during upward propagation of signals through all hippocampal subfields. Complexity of neural representations of "object - place" associations that are formed and permanently stored in the hippocampal areas increases in process of propagation of signals from the entorhinal cortex to the hierarchically higher dentate gyrus, area CA3 and area CA1. Therefore, with the aim to extract information about "object - place" associations with certain details it is necessary to access that hippocampal area in which associations were processed and stored with the required degree of elaboration. By analogy with the neocortex, it is proposed that such processing of information in the hippocampus makes it possible to avoid the combinatorial explosion and provides storing (memory) the associations accumulated during the life. The proposed mechanism can serve as an addition to the known multiple trace theory, which states that the hippocampus is an integrating part of memory trace and is always involved in recall of long-delayed episodes.     

Conditional control transfer mechanisms in the neocortex: 1. The basic model

A neural network model of the neocortex is developed which is derived from behavioural, neuroanatomical and neurophysiological findings. It is an attempt to provide a structural and functional account of the groundplan of the mammalian cerebral neocortex.Starting from the observation that the comparatively uniformly structured cortex is believed to be responsible for a wide range of behavioural tasks it is argued that all these tasks involve the integrated processing of two classes of information. The first of these, termed context information, concerns the current state of the world while the second, termed control information, concerns the animals goals, intentions and objectives. It is noted that this information dichotomy is a characteristic of all powerful information processing systems. In a digital computer the integration of the two types of information is accomplished by conditional jump instructions while in a linear control system it is accomplished by servomechanisms. Both of these may be regarded as examples of what are termed conditional control transfer devices. It is then postulated that the fundamental unit of neocortex is a conditional control transfer device.A simple neural element is then developed which has the properties of a conditional control transfer device. A number of examples are given to show how “programs” of behaviour may be built up from such units and computer simulations of these are described.In the discussion the computational power of the model is compared with that of other models of cortex and with current psychological findings regarding the role of central processes in perception.     

The temporal sequence of the mammalian neocortical neurogenetic program drives mediolateral pattern in the chick pallium

The six-layered neocortex permits complex information processing in all mammalian species. Because its homologous region (the pallium) in nonmammalian amniotes has a different architecture, the ability of neocortical progenitors to generate an orderly sequence of distinct cell types was thought to have arisen in the mammalian lineage. This study, however, shows that layer-specific neuron subtypes do exist in the chick pallium. Deep- and upper-layer neurons are not layered but are segregated in distinct mediolateral domains in vivo. Surprisingly, cultured chick neural progenitors produce multiple layer-specific neuronal subtypes in the same chronological sequence as seen in mammals. These results suggest that the temporal sequence of the neocortical neurogenetic program was already inherent in the last common ancestor of mammals and birds and that mammals use this conserved program to generate a uniformly layered neocortex, whereas birds impose spatial constraints on the sequence to pattern the pallium.     

The Optimal Human Ventral Stream from Estimates of the Complexity of Visual Objects

The part of the primate visual cortex responsible for the recognition of objects is parcelled into about a dozen areas organized somewhat hierarchically (the region is called the ventral stream). Why are there approximately this many hierarchical levels? Here I put forth a generic information-processing hierarchical model, and show how the total number of neurons required depends on the number of hierarchical levels and on the complexity of visual objects that must be recognized. Because the recognition of written words appears to occur in a similar part of inferotemporal cortex as other visual objects, the complexity of written words may be similar to that of other visual objects for humans; for this reason, I measure the complexity of written words, and use it as an approximate estimate of the complexity more generally of visual objects. I then show that the information-processing hierarchy that accommodates visual objects of that complexity possesses the minimum number of neurons when the number of hierarchical levels is approximately 15.     

Growth/differentiation factor 7 is preferentially expressed in the primary motor area of the monkey neocortex

We applied a differential display PCR technique to isolate molecules that are area-specific in expression in the primate neocortex, and found that growth/differentiation factor 7 (GDF7), a member of the bone morphogenetic protein (BMP)/transforming growth factor (TGF) beta super-family, is preferentially expressed in the primary motor area of African green monkeys (Cercopithecus aethiops). We proved that GDF7 is 10 times more abundant in the motor cortex than in the visual cortex by northern blotting and quantitative RT-PCR. When we examined the neocortex of closely related rhesus monkeys (Macaca mulatta), GDF7 was also most abundant in the motor cortex, although the regional difference was reduced to 3-fold. This differential expression pattern was observed in both newborn and infant rhesus monkeys. We found that several type I/II receptors of BMP, candidates of the receptors for GDF7, are uniformly expressed in the mature neocortex. The unique expression pattern of GDF7 suggests that it may play an active role in the motor area of the primate neocortex.     

The mystery of structure and function of sensory processing areas of the neocortex: a resolution

Many different neural models have been proposed to account for major characteristics of the memory phenomenon family in primates. However, in spite of the large body of neurophysiological, anatomical and behavioral data, there is no direct evidence for supporting one model while falsifying the others. And yet, we can discriminate models based on their complexity and/or their predictive power. In this paper we present a computational framework with our basic assumption that neural information processing is performed by generative networks. A complex architecture is 'derived' by using information-theoretic principles. We find that our approach seems to uncover possible relations among the functional memory units (declarative and implicit memory) and the process of information encoding in primates. The architecture can also be related to the entorhinal-hippocampal loop. An effort is made to form a prototype of this computational architecture and to map it onto the functional units of the neocortex. This mapping leads us to claim that one may gain a better understanding by considering that anatomical and functional layers of the cortex differ. Philosophical consequences regarding the homunculus fallacy are also considered.     

Cortical Folding of the Primate Brain: An Interdisciplinary Examination of the Genetic Architecture,Modularity, and Evolvability of a Significant Neurological Trait in Pedigreed Baboons (Genus Papio)

Folding of the primate brain cortex allows for improved neural processing power by increasing cortical surface area for the allocation of neurons. The arrangement of folds (sulci) and ridges (gyri) across the cerebral cortex is thought to reflect the underlying neural network. Gyrification, an adaptive trait with a unique evolutionary history, is affected by genetic factors different from those affecting brain volume. Using a large pedigreed population of ∼1000 Papio baboons, we address critical questions about the genetic architecture of primate brain folding, the interplay between genetics, brain anatomy, development, patterns of cortical–cortical connectivity, and gyrification’s potential for future evolution. Through Mantel testing and cluster analyses, we find that the baboon cortex is quite evolvable, with high integration between the genotype and phenotype. We further find significantly similar partitioning of variation between cortical development, anatomy, and connectivity, supporting the predictions of tension-based models for sulcal development. We identify a significant, moderate degree of genetic control over variation in sulcal length, with gyrus-shape features being more susceptible to environmental effects. Finally, through QTL mapping, we identify novel chromosomal regions affecting variation in brain folding. The most significant QTL contain compelling candidate genes, including gene clusters associated with Williams and Down syndromes. The QTL distribution suggests a complex genetic architecture for gyrification with both polygeny and pleiotropy. Our results provide a solid preliminary characterization of the genetic basis of primate brain folding, a unique and biomedically relevant phenotype with significant implications in primate brain evolution.     

Gene expression profiling of primate neocortex: molecular neuroanatomy of cortical areas

One hundred years have passed since Brodmann's mapping of the mammalian neocortex. Solely on the basis of morphological observations, he envisaged the conservation and differentiation of cortical areal structures across various species. We now know that neurochemical, connectional and functional heterogeneity of the neocortex accompanies the morphological divergence observed in such cytoarchitectonic studies. Nevertheless, we are yet far from fully understanding the biological significance of this cortical heterogeneity. In this article, we review our past works on the gene expression profiling of the postnatal primate cortical areas, by quantitative real-time polymerase chain reaction (PCR), DNA array, differential display PCR and in situ hybridization analyses. These studies revealed both the overall homogeneity of gene expression across different cortical areas and the presence of a small number of genes that show markedly area-specific expression patterns. In situ hybridization showed that, among such genes, occ1 and retinol-binding protein (RBP) mRNAs are selectively expressed in the neuronal populations that seem to be involved in distinct neural processing such as sensory reception (for occ1 ) and associative function (for RBP). Such a molecular neuroanatomical approach has the promise to provide an important link between structure and function of the cerebral cortex.     

The neurobiology of syntax: beyond string sets

The human capacity to acquire language is an outstanding scientific challenge to understand. Somehow our language capacities arise from the way the human brain processes, develops and learns in interaction with its environment. To set the stage, we begin with a summary of what is known about the neural organization of language and what our artificial grammar learning (AGL) studies have revealed. We then review the Chomsky hierarchy in the context of the theory of computation and formal learning theory. Finally, we outline a neurobiological model of language acquisition and processing based on an adaptive, recurrent, spiking network architecture. This architecture implements an asynchronous, event-driven, parallel system for recursive processing. We conclude that the brain represents grammars (or more precisely, the parser/generator) in its connectivity, and its ability for syntax is based on neurobiological infrastructure for structured sequence processing. The acquisition of this ability is accounted for in an adaptive dynamical systems framework. Artificial language learning (ALL) paradigms might be used to study the acquisition process within such a framework, as well as the processing properties of the underlying neurobiological infrastructure. However, it is necessary to combine and constrain the interpretation of ALL results by theoretical models and empirical studies on natural language processing. Given that the faculty of language is captured by classical computational models to a significant extent, and that these can be embedded in dynamic network architectures, there is hope that significant progress can be made in understanding the neurobiology of the language faculty.     

Constructing the mammalian neocortex: the role of intrinsic factors

The mammalian neocortex is subdivided into regions that are specialised for the processing of particular forms of information. These regions are distinct in terms of their cytoarchitecture, electrophysiology, and connectivity. How this regional diversity is generated through development is currently a topic of considerable interest and has centered upon two main issues. First, to what extent are these regions prespecified by intrinsic genetic mechanisms? Second, what is the influence of extrinsic activity in transmitting signals that ultimately shape functional regions? Historically, experimental evidence has tended to emphasise the role of extrinsic influences, but the identification and analysis of several genes that are expressed asymmetrically in the developing neocortex have tempered this viewpoint. We review current literature from the standpoint that intrinsic influences act early in neocortical development to generate molecular patterning whose main role is the guidance of long-range projections from the dorsal thalamus. Extrinsic influences appear to generate receptive fields for peripheral input, the summation of which determines the areal extent of particular neocortical region.     

Decision-making in the leech nervous system

Previous models of behavioral choice have described two types of hierarchy, a decision hierarchy, in which different classes of decisions are made at each level (Tinbergen, 1951), and a behavioral hierarchy, in which one behavior will take precedence over others (Davis, 1985). Most experimental work on the neuronal basis of decision-making has focussed on the latter of these: a behavioral hierarchy is described for an animal, and the neuronal basis for this hierarchy, hypothesized to depend on inhibitory interactions, is investigated. Although the concept of "dedicated command neurons" has been useful for guiding these studies, it appears that such neurons are rare. We present evidence that in the leech, most neurons, including high-level decision neurons, are active in more than one behavior. We include data from one newly-identified neuron that elicits both swimming and crawling motor patterns. We suggest that decisions are made by a "combinatorial code": what behavior is produced depends on the specific combination of decision neurons that are active at a particular time. Finally, we discuss how decision neurons may be arranged into a decision hierarchy, with neurons at each sequential level responsible for choosing between a narrower range of behaviors. We suggest additional sensory information is incorporated at each level to inform the decision.     

Modularity in vertebrate brain development and evolution.

Embryonic modularity and functional modularity are two principles of brain organization. Embryonic modules are histogenetic fields that are specified by position-dependent expression of patterning genes. Within each embryonic module, secondary and higher-level pattern formation takes places during development, finally giving rise to brain nuclei and cortical layers. Defined subsets of these structures become connected by fiber tracts to form the information-processing neural circuits, which represent the functional modules of the brain. We review evidence that a group of cell adhesion molecules, the cadherins, provides an adhesive code for both types of modularity, based on a preferentially homotypic binding mechanism. Embryonic modularity is transformed into functional modularity, in part by translating early-generated positional information into an array of adhesive cues, which regulate the binding of functional neural structures distributed across the embryonic modules. Brain modularity may provide a basis for adaptability in evolution.     

The case for primate V3

The visual system in primates is represented by a remarkably large expanse of the cerebral cortex. While more precise investigative studies that can be performed in non-human primates contribute towards understanding the organization of the human brain, there are several issues of visual cortex organization in monkey species that remain unresolved. In all, more than 20 areas comprise the primate visual cortex, yet there is little agreement as to the exact number, size and visual field representation of all but three. A case in point is the third visual area, V3. It is found relatively early in the visual system hierarchy, yet over the last 40 years its organization and even its very existence have been a matter of debate among prominent neuroscientists. In this review, we discuss a large body of recent work that provides straightforward evidence for the existence of V3. In light of this, we then re-examine results from several seminal reports and provide parsimonious re-interpretations in favour of V3. We conclude with analysis of human and monkey functional magnetic resonance imaging literature to make the case that a complete V3 is an organizational feature of all primate species and may play a greater role in the dorsal stream of visual processing.     

Hierarchical organization of macaque and cat cortical sensory systems explored with a novel network processor

Neuroanatomists have described a large number of connections between the various structures of monkey and cat cortical sensory systems. Because of the complexity of the connection data, analysis is required to unravel what principles of organization they imply. To date, analysis of laminar origin and termination connection data to reveal hierarchical relationships between the cortical areas has been the most widely acknowledged approach. We programmed a network processor that searches for optimal hierarchical orderings of cortical areas given known hierarchical constraints and rules for their interpretation. For all cortical systems and all cost functions, the processor found a multitude of equally low-cost hierarchies. Laminar hierarchical constraints that are presently available in the anatomical literature were therefore insufficient to constrain a unique ordering for any of the sensory systems we analysed. Hierarchical orderings of the monkey visual system that have been widely reported, but which were derived by hand, were not among the optimal orderings. All the cortical systems we studied displayed a significant degree of hierarchical organization, and the anatomical constraints from the monkey visual and somato-motor systems were satisfied with very few constraint violations in the optimal hierarchies. The visual and somato-motor systems in that animal were therefore surprisingly strictly hierarchical. Most inconsistencies between the constraints and the hierarchical relationships in the optimal structures for the visual system were related to connections of area FST (fundus of superior temporal sulcus). We found that the hierarchical solutions could be further improved by assuming that FST consists of two areas, which differ in the nature of their projections. Indeed, we found that perfect hierarchical arrangements of the primate visual system, without any violation of anatomical constraints, could be obtained under two reasonable conditions, namely the subdivision of FST into two distinct areas, whose connectivity we predict, and the abolition of at least one of the less reliable rule constraints. Our analyses showed that the future collection of the same type of laminar constraints, or the inclusion of new hierarchical constraints from thalamocortical connections, will not resolve the problem of multiple optimal hierarchical representations for the primate visual system. Further data, however, may help to specify the relative ordering of some more areas. This indeterminacy of the visual hierarchy is in part due to the reported absence of some connections between cortical areas. These absences are consistent with limited cross-talk between differentiated processing streams in the system. Hence, hierarchical representation of the visual system is affected by, and must take into account, other organizational features, such as processing streams.     

Neuronal selectivity to complex vocalization features emerges in the superficial layers of primary auditory cortex

Early in auditory processing, neural responses faithfully reflect acoustic input. At higher stages of auditory processing, however, neurons become selective for particular call types, eventually leading to specialized regions of cortex that preferentially process calls at the highest auditory processing stages. We previously proposed that an intermediate step in how nonselective responses are transformed into call-selective responses is the detection of informative call features. But how neural selectivity for informative call features emerges from nonselective inputs, whether feature selectivity gradually emerges over the processing hierarchy, and how stimulus information is represented in nonselective and feature-selective populations remain open question. In this study, using unanesthetized guinea pigs (GPs), a highly vocal and social rodent, as an animal model, we characterized the neural representation of calls in 3 auditory processing stages—the thalamus (ventral medial geniculate body (vMGB)), and thalamorecipient (L4) and superficial layers (L2/3) of primary auditory cortex (A1). We found that neurons in vMGB and A1 L4 did not exhibit call-selective responses and responded throughout the call durations. However, A1 L2/3 neurons showed high call selectivity with about a third of neurons responding to only 1 or 2 call types. These A1 L2/3 neurons only responded to restricted portions of calls suggesting that they were highly selective for call features. Receptive fields of these A1 L2/3 neurons showed complex spectrotemporal structures that could underlie their high call feature selectivity. Information theoretic analysis revealed that in A1 L4, stimulus information was distributed over the population and was spread out over the call durations. In contrast, in A1 L2/3, individual neurons showed brief bursts of high stimulus-specific information and conveyed high levels of information per spike. These data demonstrate that a transformation in the neural representation of calls occurs between A1 L4 and A1 L2/3, leading to the emergence of a feature-based representation of calls in A1 L2/3. Our data thus suggest that observed cortical specializations for call processing emerge in A1 and set the stage for further mechanistic studies.

A study of the neuronal representations elicited in guinea pigs by conspecific calls at different auditory processing stages reveals insights into where call-selective neuronal responses emerge; the transformation from nonselective to call-selective responses occurs in the superficial layers of the primary auditory cortex.     

A case for spiking neural network simulation based on configurable multiple-FPGA systems

Recent neuropsychological research has begun to reveal that neurons encode information in the timing of spikes. Spiking neural network simulations are a flexible and powerful method for investigating the behaviour of neuronal systems. Simulation of the spiking neural networks in software is unable to rapidly generate output spikes in large-scale of neural network. An alternative approach, hardware implementation of such system, provides the possibility to generate independent spikes precisely and simultaneously output spike waves in real time, under the premise that spiking neural network can take full advantage of hardware inherent parallelism. We introduce a configurable FPGA-oriented hardware platform for spiking neural network simulation in this work. We aim to use this platform to combine the speed of dedicated hardware with the programmability of software so that it might allow neuroscientists to put together sophisticated computation experiments of their own model. A feed-forward hierarchy network is developed as a case study to describe the operation of biological neural systems (such as orientation selectivity of visual cortex) and computational models of such systems. This model demonstrates how a feed-forward neural network constructs the circuitry required for orientation selectivity and provides platform for reaching a deeper understanding of the primate visual system. In the future, larger scale models based on this framework can be used to replicate the actual architecture in visual cortex, leading to more detailed predictions and insights into visual perception phenomenon.     

Probabilistic pathway representation of cognitive information

We present for mental processes the program of mathematical mapping which has been successfully realized for physical processes. We emphasize that our project is not about mathematical simulation of the brain's functioning as a complex physical system, i.e., mapping of physical and chemical processes in the brain on mathematical spaces. The project is about mapping of purely mental processes on mathematical spaces. We present various arguments--philosophic, mathematical, information, and neurophysiological--in favor of the p-adic model of mental space. p-adic spaces have structures of hierarchic trees and in our model such a tree hierarchy is considered as an image of neuronal hierarchy. Hierarchic neural pathways are considered as fundamental units of information processing. As neural pathways can go through the whole body, the mental space is produced by the whole neural system. Finally, we develop the probabilistic neural pathway model in that mental states are represented by probability distributions on mental space.     

The primate neocortex in comparative perspective using magnetic resonance imaging.

In this study we use neuroanatomic data from living anthropoid primate subjects to test the following three hypotheses: (1) that the human neocortex is significantly larger than expected for a primate of our brain size, (2) that the human prefrontal cortex is significantly more convoluted than expected for our brain size, and (3) that increases in cerebral white matter volume outpace increases in neocortical gray matter volume among anthropoid primates. Whole brain MRI scans were obtained from 44 living primate subjects from 11 different species. Image analysis software was used to calculate total brain volume, neocortical gray matter volume, cerebral white matter volume, and the cross sectional area of the spinal cord in each scan. Allometric regression analyses were used to compare the relative size of these brain structures across species, with an emphasis on determining whether human brain proportions correspond with predictions based on nonhuman primate allometric trajectories. All three hypotheses were supported by our analysis. The results of this study provide additional insights into human brain evolution beyond the important observation that brain volume approximately tripled in the hominid lineage by demonstrating that the neocortex was uniquely modified throughout hominid evolution. These modifications may constitute part of the neurobiological substrate that supports some of our species most distinctive cognitive abilities.