Cortical region interactions and the functional role of apical dendrites

Cellular psychology: relating cognition to context-sensitive pyramidal cells

'Cellular psychology' is a new field of inquiry that studies dendritic mechanisms for adapting mental events to the current context, thus increasing their coherence, flexibility, effectiveness, and comprehensibility. Apical dendrites of neocortical pyramidal cells have a crucial role in cognition - those dendrites receive input from diverse sources, including feedback, and can amplify the cell's feedforward transmission if relevant in that context. Specialized subsets of inhibitory interneurons regulate this cooperative context-sensitive processing by increasing or decreasing amplification. Apical input has different effects on cellular output depending on whether we are awake, deeply asleep, or dreaming. Furthermore, wakeful thought and imagery may depend on apical input. High-resolution neuroimaging in humans supports and complements evidence on these cellular mechanisms from other mammals.

Neurobiological Causal Models of Language Processing

The language faculty is physically realized in the neurobiological infrastructure of the human brain. Despite significant efforts, an integrated understanding of this system remains a formidable challenge. What is missing from most theoretical accounts is a specification of the neural mechanisms that implement language function. Computational models that have been put forward generally lack an explicit neurobiological foundation. We propose a neurobiologically informed causal modeling approach which offers a framework for how to bridge this gap. A neurobiological causal model is a mechanistic description of language processing that is grounded in, and constrained by, the characteristics of the neurobiological substrate. It intends to model the generators of language behavior at the level of implementational causality. We describe key features and neurobiological component parts from which causal models can be built and provide guidelines on how to implement them in model simulations. Then we outline how this approach can shed new light on the core computational machinery for language, the long-term storage of words in the mental lexicon and combinatorial processing in sentence comprehension. In contrast to cognitive theories of behavior, causal models are formulated in the "machine language" of neurobiology which is universal to human cognition. We argue that neurobiological causal modeling should be pursued in addition to existing approaches. Eventually, this approach will allow us to develop an explicit computational neurobiology of language.

Methodological advice for the young at heart investigator: Triangulation to build better foundations

In medicine and science, one is typically taught the main theories in a discipline or field along with standard models before receiving more instructions on how to apply certain methods. The aim of this work is not to address one method, but rather methodology, the study and evaluation of methods, by taking a philosophy of science detour. In this, a critique of biomedicine will be used as a starting point to address some positions regarding reductionism, specifying notions such as systems and mechanisms, as well as regarding the mind-body problem discussing psychosomatic medicine and psychoneuroimmunology. Some recommendations to make science more pluralistic, robust and translationally-relevant will then be made as a way to foster constructive debates on reductionism and the mind-body problem and, in turn, favor more interdisciplinary research.

Canonical circuit computations for computer vision

Advanced computer vision mechanisms have been inspired by neuroscientific findings. However, with the focus on improving benchmark achievements, technical solutions have been shaped by application and engineering constraints. This includes the training of neural networks which led to the development of feature detectors optimally suited to the application domain. However, the limitations of such approaches motivate the need to identify computational principles, or motifs, in biological vision that can enable further foundational advances in machine vision. We propose to utilize structural and functional principles of neural systems that have been largely overlooked. They potentially provide new inspirations for computer vision mechanisms and models. Recurrent feedforward, lateral, and feedback interactions characterize general principles underlying processing in mammals. We derive a formal specification of core computational motifs that utilize these principles. These are combined to define model mechanisms for visual shape and motion processing. We demonstrate how such a framework can be adopted to run on neuromorphic brain-inspired hardware platforms and can be extended to automatically adapt to environment statistics. We argue that the identified principles and their formalization inspires sophisticated computational mechanisms with improved explanatory scope. These and other elaborated, biologically inspired models can be employed to design computer vision solutions for different tasks and they can be used to advance neural network architectures of learning.

Comparing the ontogeny, neurobiology, and function of social play in hamsters and rats

Syrian hamsters show complex social play behavior and provide a valuable animal model for delineating the neurobiological mechanisms and functions of social play. In this review, we compare social play behavior of hamsters and rats and underlying neurobiological mechanisms. Juvenile rats play by competing for opportunities to pin one another and attack their partner's neck. A broad set of cortical, limbic, and striatal regions regulate the display of social play in rats. In hamsters, social play is characterized by attacks to the head in early puberty, which gradually transitions to the flanks in late puberty. The transition from juvenile social play to adult hamster aggression corresponds with engagement of neural ensembles controlling aggression. Play deprivation in rats and hamsters alters dendritic morphology in mPFC neurons and impairs flexible, context-dependent behavior in adulthood, which suggests these animals may have converged on a similar function for social play. Overall, dissecting the neurobiology of social play in hamsters and rats can provide a valuable comparative approach for evaluating the function of social play.

Atypical play experiences in the juvenile period has an impact on the development of the medial prefrontal cortex in both male and female rats

In rats reared without play, or with limited access to play during the juvenile period, the dendrites of pyramidal neurons of the medial prefrontal cortex (mPFC) exhibit more branching than rats reared with more typical levels of play. This suggests that play is critical for pruning the dendritic arbor of these neurons. However, the rearing paradigms typically used to limit play involve physical separation from a peer or sharing a cage with an adult, causing stress that may disrupt pruning. To limit this potentially confounding source of stress, we used an alternative approach in this study: pairing playful Long Evans rats (LE) with low playing Fischer 344 (F344) rats throughout the juvenile period. We then examined the morphology of medial prefrontal cortex (mPFC) neurons, predicting that pruning should be reduced. LE rats reared with another LE rat had significantly greater pruning of mPFC pyramidal neurons compared to LE rats reared with a F344 partner. Furthermore, in previous studies, only one sex or the other was used, whereas in the present rearing paradigm, both sexes were tested, showing that play influences neuronal pruning in both. The neurons of the play deficient LE rats not only occupied more space, as determined by convex hull analyses, but the dendrites were also longer than in rats with more typical play experiences. Unlike studies using more stressful rearing paradigms, the present effects were limited to the apical dendritic projections, suggesting that the previously reported effects on the basilar dendrites may have resulted from developmental disruptions caused by stress. If correct, the present findings indicate that play experienced over the juvenile period affects how mPFC neurons develop and function.

Error-based or target-based? A unified framework for learning in recurrent spiking networks

The field of recurrent neural networks is over-populated by a variety of proposed learning rules and protocols. The scope of this work is to define a generalized framework, to move a step forward towards the unification of this fragmented scenario. In the field of supervised learning, two opposite approaches stand out, error-based and target-based. This duality gave rise to a scientific debate on which learning framework is the most likely to be implemented in biological networks of neurons. Moreover, the existence of spikes raises the question of whether the coding of information is rate-based or spike-based. To face these questions, we proposed a learning model with two main parameters, the rank of the feedback learning matrix R and the tolerance to spike timing τ_⋆. We demonstrate that a low (high) rank R accounts for an error-based (target-based) learning rule, while high (low) tolerance to spike timing promotes rate-based (spike-based) coding. We show that in a store and recall task, high-ranks allow for lower MSE values, while low-ranks enable a faster convergence. Our framework naturally lends itself to Behavioral Cloning and allows for efficiently solving relevant closed-loop tasks, investigating what parameters (R,τ⋆) are optimal to solve a specific task. We found that a high R is essential for tasks that require retaining memory for a long time (Button and Food). On the other hand, this is not relevant for a motor task (the 2D Bipedal Walker). In this case, we find that precise spike-based coding enables optimal performances. Finally, we show that our theoretical formulation allows for defining protocols to estimate the rank of the feedback error in biological networks. We release a PyTorch implementation of our model supporting GPU parallelization.

Learning in biological or artificial networks means changing the laws governing the network dynamics in order to better behave in a specific situation. However, there exists no consensus on what rules regulate learning in biological systems. To face these questions, we propose a novel theoretical formulation for learning with two main parameters, the number of learning constraints (R) and the tolerance to spike timing (τ_⋆). We demonstrate that a low (high) rank R accounts for an error-based (target-based) learning rule, while high (low) tolerance to spike timing τ_⋆ promotes rate-based (spike-based) coding.

Our approach naturally lends itself to Imitation Learning (and Behavioral Cloning in particular) and we apply it to solve relevant closed-loop tasks such as the button-and-food task, and the 2D Bipedal Walker. The button-and-food is a navigation task that requires retaining a long-term memory, and benefits from a high R. On the other hand, the 2D Bipedal Walker is a motor task and benefits from a low τ_⋆.

Finally, we show that our theoretical formulation suggests protocols to deduce the structure of learning feedback in biological networks.

Are the Anterior and Mid-Cingulate Cortices Distinct in Rodents?

The prefrontal cortex (PFC) is involved in cognitive control, emotional regulation, and motivation. In this Perspective article, we discuss the nomenclature of the subdivisions of the medial prefrontal cortex (mPFC), since the anatomical definitions of the PFC subregions have been confusing. Although the mid-cingulate cortex (MCC) and anterior cingulate cortex (ACC) have distinct features in humans and non-human primates, it is unclear whether these regions serve different functions in rodents. Accurate mapping of the cingulate cortex in rodents is important to allow comparisons between species. A proposed change in the nomenclature of the rodent cingulate cortex to anterior cingulate cortex (aCg) and mid-cingulate cortex (mCg) is presented based on our data. We show evidence for distinct cortico-cortical projections from the aCg and mCg to the PrL. The aCg→PrL neurons were abundant in layer VI, while the mCg→PrL neurons were mainly distributed in layer V. In addition, a sex difference was detected in the aCg, with males having a higher proportion of layer V neurons projecting to the PrL than females. Based on this laminar distribution and considering that layer V and VI send efferent projections to different brain areas such as the brain stem, amygdala, and thalamus, we propose that aCg and mCg need to be considered separate entities for future rodent studies. This new definition will put into perspective the role of rodent cingulate cortex in diverse aspects of cognition and facilitate interspecies comparisons in cingulate cortex research.

Long-Range Monosynaptic Inputs Targeting Apical and Basal Dendrites of Primary Motor Cortex Deep Output Neurons

The primary motor cortex (M1) integrates various long-range signals from other brain regions for the learning and execution of goal-directed movements. How the different inputs target the distinct apical and basal dendrites of M1 pyramidal neurons is crucial in understanding the functions of M1, but the detailed connectivity pattern is still largely unknown. Here, by combining cre-dependent rabies virus tracing, layer-specific chemical retrograde tracing, optogenetic stimulation, and electrophysiological recording, we mapped all long-range monosynaptic inputs to M1 deep output neurons in layer 5 (L5) in mice. We revealed that most upstream areas innervate both dendritic compartments concurrently. These include the sensory cortices, higher motor cortices, sensory and motor thalamus, association cortices, as well as many subcortical nuclei. Furthermore, the dichotomous inputs arise mostly from spatially segregated neuronal subpopulations within an upstream nucleus, and even in the case of an individual cortical layer. Therefore, these input areas could serve as both feedforward and feedback sources albeit via different subpopulations. Taken together, our findings revealed a previously unknown and highly intricate synaptic input pattern of M1L5 neurons, which implicates that the dendritic computations carried out by these neurons during motor execution or learning are far more complicated than we currently understand.

Pharmacological rescue of cognitive function in a mouse model of chemobrain

BACKGROUND

After chemotherapy, many cancer survivors suffer from long-lasting cognitive impairment, colloquially known as "chemobrain." However, the trajectories of cognitive changes and the underlying mechanisms remain unclear. We previously established paclitaxel-induced inositol trisphosphate receptor (InsP3R)-dependent calcium oscillations as a mechanism for peripheral neuropathy, which was prevented by lithium pretreatment. Here, we investigated if a similar mechanism also underlay paclitaxel-induced chemobrain.

METHOD

Mice were injected with 4 doses of 20 mg/kg paclitaxel every other day to induced cognitive impairment. Memory acquisition was assessed with the displaced object recognition test. The morphology of neurons in the prefrontal cortex and the hippocampus was analyzed using Golgi-Cox staining, followed by Sholl analyses. Changes in protein expression were measured by Western blot.

RESULTS

Mice receiving paclitaxel showed impaired short-term spatial memory acquisition both acutely 5 days post injection and chronically 23 days post injection. Dendritic length and complexity were reduced in the hippocampus and the prefrontal cortex after paclitaxel injection. Concurrently, the expression of protein kinase C α (PKCα), an effector in the InsP3R pathway, was increased. Treatment with lithium before or shortly after paclitaxel injection rescued the behavioral, cellular, and molecular deficits observed. Similarly, memory and morphological deficits could be rescued by pretreatment with chelerythrine, a PKC inhibitor.

CONCLUSION

We establish the InsP3R calcium pathway and impaired neuronal morphology as mechanisms for paclitaxel-induced cognitive impairment. Our findings suggest lithium and PKC inhibitors as candidate agents for preventing chemotherapy-induced cognitive impairment.

Sex-Specific Role for SLIT1 in Regulating Stress Susceptibility

BACKGROUND

Major depressive disorder is a pervasive and debilitating syndrome characterized by mood disturbances, anhedonia, and alterations in cognition. While the prevalence of major depressive disorder is twice as high for women as men, little is known about the molecular mechanisms that drive sex differences in depression susceptibility.

METHODS

We discovered that SLIT1, a secreted protein essential for axonal navigation and molecular guidance during development, is downregulated in the adult ventromedial prefrontal cortex (vmPFC) of women with depression compared with healthy control subjects, but not in men with depression. This sex-specific downregulation of Slit1 was also observed in the vmPFC of mice exposed to chronic variable stress. To identify a causal, sex-specific role for SLIT1 in depression-related behavioral abnormalities, we performed knockdown (KD) of Slit1 expression in the vmPFC of male and female mice.

RESULTS

When combined with stress exposure, vmPFC Slit1 KD reflected the human condition by inducing a sex-specific increase in anxiety- and depression-related behaviors. Furthermore, we found that vmPFC Slit1 KD decreased the dendritic arborization of vmPFC pyramidal neurons and decreased the excitability of the neurons in female mice, effects not observed in males. RNA sequencing analysis of the vmPFC after Slit1 KD in female mice revealed an augmented transcriptional stress signature.

CONCLUSIONS

Together, our findings establish a crucial role for SLIT1 in regulating neurophysiological and transcriptional responses to stress within the female vmPFC and provide mechanistic insight into novel signaling pathways and molecular factors influencing sex differences in depression susceptibility.

Efficient Spike-Driven Learning With Dendritic Event-Based Processing

A critical challenge in neuromorphic computing is to present computationally efficient algorithms of learning. When implementing gradient-based learning, error information must be routed through the network, such that each neuron knows its contribution to output, and thus how to adjust its weight. This is known as the credit assignment problem. Exactly implementing a solution like backpropagation involves weight sharing, which requires additional bandwidth and computations in a neuromorphic system. Instead, models of learning from neuroscience can provide inspiration for how to communicate error information efficiently, without weight sharing. Here we present a novel dendritic event-based processing (DEP) algorithm, using a two-compartment leaky integrate-and-fire neuron with partially segregated dendrites that effectively solves the credit assignment problem. In order to optimize the proposed algorithm, a dynamic fixed-point representation method and piecewise linear approximation approach are presented, while the synaptic events are binarized during learning. The presented optimization makes the proposed DEP algorithm very suitable for implementation in digital or mixed-signal neuromorphic hardware. The experimental results show that spiking representations can rapidly learn, achieving high performance by using the proposed DEP algorithm. We find the learning capability is affected by the degree of dendritic segregation, and the form of synaptic feedback connections. This study provides a bridge between the biological learning and neuromorphic learning, and is meaningful for the real-time applications in the field of artificial intelligence.

The thalamus integrates the macrosystems of the brain to facilitate complex, adaptive brain network dynamics

The human brain is a complex, adaptive system comprised of billions of cells with trillions of connections. The interactions between the elements of the system oppose this seemingly limitless capacity by constraining the system's dynamic repertoire, enforcing distributed neural states that balance integration and differentiation. How this trade-off is mediated by the brain, and how the emergent, distributed neural patterns give rise to cognition and awareness, remains poorly understood. Here, I argue that the thalamus is well-placed to arbitrate the interactions between distributed neural assemblies in the cerebral cortex. Different classes of thalamocortical connections are hypothesized to promote either feed-forward or feedback processing modes in the cerebral cortex. This activity can be conceptualized as emerging dynamically from an evolving attractor landscape, with the relative engagement of distinct distributed circuits providing differing constraints over the manner in which brain state trajectories change over time. In addition, inputs to the distinct thalamic populations from the cerebellum and basal ganglia, respectively, are proposed to differentially shape the attractor landscape, and hence, the temporal evolution of cortical assemblies. The coordinated engagement of these neural macrosystems is then shown to share key characteristics with prominent models of cognition, attention and conscious awareness. In this way, the crucial role of the thalamus in mediating the distributed, multi-scale network organization of the central nervous system can be related to higher brain function.

The contribution of stem cell factor and granulocyte colony-stimulating factor in reducing neurodegeneration and promoting neurostructure network reorganization after traumatic brain injury

Traumatic brain injury (TBI) is a major cause of death and disability in young adults worldwide. TBI-induced long-term cognitive deficits represent a growing clinical problem. Stem cell factor (SCF) and granulocyte colony-stimulating factor (G-CSF) are involved in neuroprotection and neuronal plasticity. However, the knowledge concerning reparative efficacy of SCF + G-CSF treatment in post-acute TBI recovery remains incomplete. This study aims to determine the efficacy of SCF + G-CSF on post-acute TBI recovery in young adult mice. The controlled cortical impact model of TBI was used for inducing a severe damage in the motor cortex of the right hemisphere in 8-week-old male C57BL mice. SCF + G-CSF treatment was initiated 3 weeks after induction of TBI. Severe TBI led to persistent motor functional deficits (Rota-Rod test) and impaired spatial learning function (water maze test). SCF + G-CSF treatment significantly improved the severe TBI-impaired spatial learning function 6 weeks after treatment. TBI also caused significant increases of Fluoro-Jade C positive degenerating neurons in bilateral frontal cortex, striatum and hippocampus, and significant reductions in MAP2⁺ apical dendrites and overgrowth of SMI312⁺ axons in peri-TBI cavity frontal cortex and in the ipsilateral hippocampal CA1 at 24 weeks post-TBI. SCF + G-CSF treatment significantly reduced TBI-induced neurodegeneration in the contralateral frontal cortex and hippocampal CA1, increased MAP2⁺ apical dendrites in the peri-TBI cavity frontal cortex, and prevented TBI-induced axonal overgrowth in both the peri-TBI cavity frontal cortex and ipsilateral hippocampal CA1.These findings reveal a novel pathology of axonal overgrowth after severe TBI and demonstrate a therapeutic potential of SCF + G-CSF in ameliorating severe TBI-induced long-term neuronal pathology, neurostructural network malformation, and impairments in spatial learning.

Perceptron Learning and Classification in a Modeled Cortical Pyramidal Cell

The perceptron learning algorithm and its multiple-layer extension, the backpropagation algorithm, are the foundations of the present-day machine learning revolution. However, these algorithms utilize a highly simplified mathematical abstraction of a neuron; it is not clear to what extent real biophysical neurons with morphologically-extended non-linear dendritic trees and conductance-based synapses can realize perceptron-like learning. Here we implemented the perceptron learning algorithm in a realistic biophysical model of a layer 5 cortical pyramidal cell with a full complement of non-linear dendritic channels. We tested this biophysical perceptron (BP) on a classification task, where it needed to correctly binarily classify 100, 1,000, or 2,000 patterns, and a generalization task, where it was required to discriminate between two "noisy" patterns. We show that the BP performs these tasks with an accuracy comparable to that of the original perceptron, though the classification capacity of the apical tuft is somewhat limited. We concluded that cortical pyramidal neurons can act as powerful classification devices.

Synaptic Potentiation at Basal and Apical Dendrites of Hippocampal Pyramidal Neurons Involves Activation of a Distinct Set of Extracellular and Intracellular Molecular Cues

In the central nervous system, several forms of experience-dependent plasticity, learning and memory require the activity-dependent control of synaptic efficacy. Despite substantial progress in describing synaptic plasticity, mechanisms related to heterogeneity of synaptic functions at local circuits remain elusive. Here we studied the functional and molecular aspects of hippocampal circuit plasticity by analyzing excitatory synapses at basal and apical dendrites of mouse hippocampal pyramidal cells (CA1 region) in acute brain slices. In the past decade, activity of metalloproteinases (MMPs) has been implicated as a widespread and critical factor in plasticity mechanisms at various projections in the CNS. However, in the present study we discovered that in striking contrast to apical dendrites, synapses located within basal dendrites undergo MMP-independent synaptic potentiation. We demonstrate that synapse-specific molecular pathway allowing MMPs to rapidly upregulate function of NMDARs in stratum radiatum involved protease activated receptor 1 and intracellular kinases and GTPases activity. In contrast, MMP-independent scaling of synaptic strength in stratum oriens involved dopamine D1/D5 receptors and Src kinases. Results of this study reveal that 2 neighboring synaptic systems differ significantly in extracellular and intracellular cascades that control synaptic gain and provide long-searched transduction pathways relevant for MMP-dependent synaptic plasticity.

Amplifying the redistribution of somato-dendritic inhibition by the interplay of three interneuron types

GABAergic interneurons play an important role in shaping the activity of excitatory pyramidal cells (PCs). How the various inhibitory cell types contribute to neuronal information processing, however, is not resolved. Here, we propose a functional role for a widespread network motif consisting of parvalbumin- (PV), somatostatin- (SOM) and vasoactive intestinal peptide (VIP)-expressing interneurons. Following the idea that PV and SOM interneurons control the distribution of somatic and dendritic inhibition onto PCs, we suggest that mutual inhibition between VIP and SOM cells translates weak inputs to VIP interneurons into large changes of somato-dendritic inhibition of PCs. Using a computational model, we show that the neuronal and synaptic properties of the circuit support this hypothesis. Moreover, we demonstrate that the SOM-VIP motif allows transient inputs to persistently switch the circuit between two processing modes, in which top-down inputs onto apical dendrites of PCs are either integrated or cancelled.

Neurons in the brain can be classified as excitatory or inhibitory based on whether they activate or deactivate the cells to whom they send signals. Compared to their excitatory counterpart, inhibitory neurons present themselves as a wild diversity of cell classes. It is broadly believed that these classes serve different purposes, but as of now, those are poorly understood. In this article, we suggest how an intricate interplay of three inhibitory cell classes can control whether internal signals—such as predictions, memory signals or motor commands—are taken into account when sensory signals are interpreted. Using a mathematical model and computer simulations, we show that such internal signals can be shut down by regulating which inhibitory cell types are active, and that the interaction of different cell classes allows weak control signals to do so.

A Perspective on Cortical Layering and Layer-Spanning Neuronal Elements

This review article addresses the function of the layers of the cerebral cortex. We develop the perspective that cortical layering needs to be understood in terms of its functional anatomy, i.e., the terminations of synaptic inputs on distinct cellular compartments and their effect on cortical activity. The cortex is a hierarchical structure in which feed forward and feedback pathways have a layer-specific termination pattern. We take the view that the influence of synaptic inputs arriving at different cortical layers can only be understood in terms of their complex interaction with cellular biophysics and the subsequent computation that occurs at the cellular level. We use high-resolution fMRI, which can resolve activity across layers, as a case study for implementing this approach by describing how cognitive events arising from the laminar distribution of inputs can be interpreted by taking into account the properties of neurons that span different layers. This perspective is based on recent advances in measuring subcellular activity in distinct feed-forward and feedback axons and in dendrites as they span across layers.

Neuronal gain modulability is determined by dendritic morphology: A computational optogenetic study

The mechanisms by which the gain of the neuronal input-output function may be modulated have been the subject of much investigation. However, little is known of the role of dendrites in neuronal gain control. New optogenetic experimental paradigms based on spatial profiles or patterns of light stimulation offer the prospect of elucidating many aspects of single cell function, including the role of dendrites in gain control. We thus developed a model to investigate how competing excitatory and inhibitory input within the dendritic arbor alters neuronal gain, incorporating kinetic models of opsins into our modeling to ensure it is experimentally testable. To investigate how different topologies of the neuronal dendritic tree affect the neuron’s input-output characteristics we generate branching geometries which replicate morphological features of most common neurons, but keep the number of branches and overall area of dendrites approximately constant. We found a relationship between a neuron’s gain modulability and its dendritic morphology, with neurons with bipolar dendrites with a moderate degree of branching being most receptive to control of the gain of their input-output relationship. The theory was then tested and confirmed on two examples of realistic neurons: 1) layer V pyramidal cells—confirming their role in neural circuits as a regulator of the gain in the circuit in addition to acting as the primary excitatory neurons, and 2) stellate cells. In addition to providing testable predictions and a novel application of dual-opsins, our model suggests that innervation of all dendritic subdomains is required for full gain modulation, revealing the importance of dendritic targeting in the generation of neuronal gain control and the functions that it subserves. Finally, our study also demonstrates that neurophysiological investigations which use direct current injection into the soma and bypass the dendrites may miss some important neuronal functions, such as gain modulation.

New experimental techniques based on optogenetics allow neuronal activity to be manipulated with a high degree of spatial and temporal precision. This opens up new prospects for testing computational models of neuronal function, including questions such as the role of dendrites in neuronal gain control. However, compartmental models in computational neuroscience have not, until now, incorporated the kinetic models of opsins that are required in order to directly match the predictions of a computational model with observed optogenetic experimental results. Here, we introduce an approach for computational optogenetic modeling to test hypotheses, demonstrating it with application to the role of dendrites in neuronal gain control. We find that gain modulability is indicated by dendritic morphology, with pyramidal cell-like shapes optimally receptive to modulation. All dendritic subdomains are required for gain modulation—partial illumination is insufficient. Due to the simulation framework used, these results are directly testable through optogenetic experiments. Computational optogenetic models thus can be used to improve and refine experimental protocols for direct testing of theories of neural function.

Towards deep learning with segregated dendrites

Deep learning has led to significant advances in artificial intelligence, in part, by adopting strategies motivated by neurophysiology. However, it is unclear whether deep learning could occur in the real brain. Here, we show that a deep learning algorithm that utilizes multi-compartment neurons might help us to understand how the neocortex optimizes cost functions. Like neocortical pyramidal neurons, neurons in our model receive sensory information and higher-order feedback in electrotonically segregated compartments. Thanks to this segregation, neurons in different layers of the network can coordinate synaptic weight updates. As a result, the network learns to categorize images better than a single layer network. Furthermore, we show that our algorithm takes advantage of multilayer architectures to identify useful higher-order representations—the hallmark of deep learning. This work demonstrates that deep learning can be achieved using segregated dendritic compartments, which may help to explain the morphology of neocortical pyramidal neurons.

Artificial intelligence has made major progress in recent years thanks to a technique known as deep learning, which works by mimicking the human brain. When computers employ deep learning, they learn by using networks made up of many layers of simulated neurons. Deep learning has opened the door to computers with human – or even super-human – levels of skill in recognizing images, processing speech and controlling vehicles. But many neuroscientists are skeptical about whether the brain itself performs deep learning.

The patterns of activity that occur in computer networks during deep learning resemble those seen in human brains. But some features of deep learning seem incompatible with how the brain works. Moreover, neurons in artificial networks are much simpler than our own neurons. For instance, in the region of the brain responsible for thinking and planning, most neurons have complex tree-like shapes. Each cell has ‘roots’ deep inside the brain and ‘branches’ close to the surface. By contrast, simulated neurons have a uniform structure.

To find out whether networks made up of more realistic simulated neurons could be used to make deep learning more biologically realistic, Guerguiev et al. designed artificial neurons with two compartments, similar to the ‘roots’ and ‘branches’. The network learned to recognize hand-written digits more easily when it had many layers than when it had only a few. This shows that artificial neurons more like those in the brain can enable deep learning. It even suggests that our own neurons may have evolved their shape to support this process.

If confirmed, the link between neuronal shape and deep learning could help us develop better brain-computer interfaces. These allow people to use their brain activity to control devices such as artificial limbs. Despite advances in computing, we are still superior to computers when it comes to learning. Understanding how our own brains show deep learning could thus help us develop better, more human-like artificial intelligence in the future.

Differential dendritic remodeling in prelimbic cortex of male and female rats during recovery from chronic stress

Chronic stress produces differential dendritic remodeling of pyramidal neurons in medial prefrontal cortex of male and female rats. In males, this dendritic remodeling is reversible. However, the timeline of recovery, as well as the potential for reversibility in females, is unknown. Here, we examined dendritic recovery of pyramidal neurons in layer II-II of prelimbic cortex in male and female rats following chronic restraint stress (3h/day for 10days). Dendritic morphology and spine density were analyzed immediately following the cessation of stress, or following a 7- or 10-day recovery period. Chronic stress produced apical dendritic retraction in males, which was coupled with a decrease in the density of stubby spine on apical dendrites. Further, following a 10-day recovery period, the morphology of neurons from stressed rats resembled that of unstressed rats. Male rats given a 7-day recovery period had apical dendritic outgrowth compared to unstressed rats. Immediately after cessation of stress, females showed only minimal dendritic remodeling. The morphology of neurons in stressed females resembled those of unstressed rats following only 7days of recovery, at which time there was also a significant increase in stubby spine density. Males and females also showed different changes in baseline corticosterone concentrations during recovery. These findings not only indicate that dendritic remodeling in prelimbic cortex following chronic stress is different between males and females, but also suggest chronic stress induces differential hypothalamic-pituitary-adrenal axis dysregulation in males and females. These differences may have important implications for responses to subsequent stressors.

Alterations in neuronal morphology in infralimbic cortex predict resistance to fear extinction following acute stress

Dysfunction in corticolimbic circuits that mediate the extinction of learned fear responses is thought to underlie the perseveration of fear in stress-related psychopathologies, including post-traumatic stress disorder. Chronic stress produces dendritic hypertrophy in basolateral amygdala (BLA) and dendritic hypotrophy in medial prefrontal cortex, whereas acute stress leads to hypotrophy in both BLA and prelimbic cortex. Additionally, both chronic and acute stress impair extinction retrieval. Here, we examined the effects of a single elevated platform stress on extinction learning and dendritic morphology in infralimbic cortex, a region considered to be critical for extinction. Acute stress produced resistance to extinction, as well as dendritic retraction in infralimbic cortex. Spine density on apical and basilar terminal branches was unaffected by stress. However, animals that underwent conditioning and extinction had decreased spine density on apical terminal branches. Thus, whereas dendritic morphology in infralimbic cortex appears to be particularly sensitive to stress, changes in spines may more sensitively reflect learning. Further, in stressed rats that underwent conditioning and extinction, the level of extinction learning was correlated with spine densities, in that rats with poorer extinction retrieval had more immature spines and fewer thin spines than rats with better extinction retrieval, suggesting that stress may have impaired learning-related spine plasticity. These results may have implications for understanding the role of medial prefrontal cortex in learning deficits associated with stress-related pathologies.

Social play in juvenile hamsters alters dendritic morphology in the medial prefrontal cortex and attenuates effects of social stress in adulthood

Social play is a fundamental aspect of behavioral development in many species. Social play deprivation in rats alters dendritic morphology in the ventromedial prefrontal cortex (vmPFC) and we have shown that this brain region regulates responses to social defeat stress in Syrian hamsters. In this study, we tested whether play deprivation during the juvenile period disrupts dendritic morphology in the prefrontal cortex and potentiates the effects of social defeat stress. At weaning, male hamsters were either group-housed with peers or pair-housed with their mother, with whom they do not play. In adulthood, animals received acute social defeat stress or no-defeat control treatment. The hamsters were then tested for a conditioned defeat response in a social interaction test with a novel intruder, and were also tested for social avoidance of a familiar opponent. Brains were collected for Golgi-Cox staining and analysis of dendritic morphology in the infralimbic (IL), prelimbic (PL), and orbitofrontal cortex (OFC). Play-deprived animals showed an increased conditioned defeat response and elevated avoidance of a familiar opponent compared with play-exposed animals. Furthermore, play-deprived animals showed increased total length and branch points in apical dendrites of pyramidal neurons in the IL and PL cortices, but not in the OFC. These findings suggest that social play deprivation in juvenile hamsters disrupts neuronal development in the vmPFC and increases vulnerability to the effects of social stress in adulthood. Overall, these results suggest that social play is necessary for the natural dendritic pruning process during adolescence and promotes coping with stress in adulthood. (PsycINFO Database Record

Using the Change Manager Model for the Hippocampal System to Predict Connectivity and Neurophysiological Parameters in the Perirhinal Cortex

Theoretical arguments demonstrate that practical considerations, including the needs to limit physiological resources and to learn without interference with prior learning, severely constrain the anatomical architecture of the brain. These arguments identify the hippocampal system as the change manager for the cortex, with the role of selecting the most appropriate locations for cortical receptive field changes at each point in time and driving those changes. This role results in the hippocampal system recording the identities of groups of cortical receptive fields that changed at the same time. These types of records can also be used to reactivate the receptive fields active during individual unique past events, providing mechanisms for episodic memory retrieval. Our theoretical arguments identify the perirhinal cortex as one important focal point both for driving changes and for recording and retrieving episodic memories. The retrieval of episodic memories must not drive unnecessary receptive field changes, and this consideration places strong constraints on neuron properties and connectivity within and between the perirhinal cortex and regular cortex. Hence the model predicts a number of such properties and connectivity. Experimental test of these falsifiable predictions would clarify how change is managed in the cortex and how episodic memories are retrieved.

Functional Improvement after Photothrombotic Stroke in Rats Is Associated with Different Patterns of Dendritic Plasticity after G-CSF Treatment and G-CSF Treatment Combined with Concomitant or Sequential Constraint-Induced Movement Therapy

We have previously shown that granulocyte-colony stimulating factor (G-CSF) treatment alone, or in combination with constraint movement therapy (CIMT) either sequentially or concomitantly, results in significantly improved sensorimotor recovery after photothrombotic stroke in rats in comparison to untreated control animals. CIMT alone did not result in any significant differences compared to the control group (Diederich et al., Stroke, 2012;43:185-192). Using a subset of rat brains from this former experiment the present study was designed to evaluate whether dendritic plasticity would parallel improved functional outcomes. Five treatment groups were analyzed (n = 6 each) (i) ischemic control (saline); (ii) CIMT (CIMT between post-stroke days 2 and 11); (iii) G-CSF (10 μg/kg G-CSF daily between post-stroke days 2 and 11); (iv) combined concurrent group (CIMT plus G-CSF) and (v) combined sequential group (CIMT between post-stroke days 2 and 11; 10 μg/kg G-CSF daily between post-stroke days 12 and 21, respectively). After impregnation of rat brains with a modified Golgi-Cox protocol layer V pyramidal neurons in the peri-infarct cortex as well as the corresponding contralateral cortex were analyzed. Surprisingly, animals with a similar degree of behavioral recovery exhibited quite different patterns of dendritic plasticity in both peri-lesional and contralesional areas. The cause for these patterns is not easily to explain but puts the simple assumption that increased dendritic complexity after stroke necessarily results in increased functional outcome into perspective.

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.

Topological characterization of neuronal arbor morphology via sequence representation: I--motif analysis

BACKGROUND

The morphology of neurons offers many insights into developmental processes and signal processing. Numerous reports have focused on metrics at the level of individual branches or whole arbors; however, no studies have attempted to quantify repeated morphological patterns within neuronal trees. We introduce a novel sequential encoding of neurite branching suitable to explore topological patterns.

RESULTS

Using all possible branching topologies for comparison we show that the relative abundance of short patterns of up to three bifurcations, together with overall tree size, effectively capture the local branching patterns of neurons. Dendrites and axons display broadly similar topological motifs (over-represented patterns) and anti-motifs (under-represented patterns), differing most in their proportions of bifurcations with one terminal branch and in select sub-sequences of three bifurcations. In addition, pyramidal apical dendrites reveal a distinct motif profile.

CONCLUSIONS

The quantitative characterization of topological motifs in neuronal arbors provides a thorough description of local features and detailed boundaries for growth mechanisms and hypothesized computational functions.

Feedback to distal dendrites links fMRI signals to neural receptive fields in a spiking network model of the visual cortex

The blood oxygenation level-dependent (BOLD) response has been strongly associated with neuronal activity in the brain. However, some neuronal tuning properties are consistently different from the BOLD response. We studied the spatial extent of neural and hemodynamic responses in the primary visual cortex, where the BOLD responses spread and interact over much longer distances than the small receptive fields of individual neurons would predict. Our model shows that a feedforward-feedback loop between V1 and a higher visual area can account for the observed spread of the BOLD response. In particular, anisotropic landing of inputs to compartmental neurons were necessary to account for the BOLD signal spread, while retaining realistic spiking responses. Our work shows that simple dendrites can separate tuning at the synapses and at the action potential output, thus bridging the BOLD signal to the neural receptive fields with high fidelity.

D-serine deficiency attenuates the behavioral and cellular effects induced by the hallucinogenic 5-HT(2A) receptor agonist DOI

Both the serotonin and glutamate systems have been implicated in the pathophysiology of schizophrenia, as well as in the mechanism of action of antipsychotic drugs. Psychedelic drugs act through the serotonin 2A receptor (5-HT2AR), and elicit a head-twitch response (HTR) in mice, which directly correlates to 5-HT2AR activation and is absent in 5-HT2AR knockout mice. The precise mechanism of this response remains unclear, but both an intrinsic cortico-cortical pathway and a thalamo-cortical pathway involving glutamate release have been proposed. Here, we used a genetic model of NMDAR hypofunction, the serine racemase knockout (SRKO) mouse, to explore the role of glutamatergic transmission in regulating 5-HT2AR-mediated cellular and behavioral responses. SRKO mice treated with the 5-HT2AR agonist (±)-2,5-dimethoxy-4-iodoamphetamine (DOI) showed a clearly diminished HTR and lower induction of c-fos mRNA. These altered functional responses in SRKO mice were not associated with changes in cortical or hippocampal 5-HT levels or in 5-HT2AR and metabotropic glutamate-2 receptor (mGluR2) mRNA and protein expression. Together, these findings suggest that D-serine-dependent NMDAR activity is involved in mediating the cellular and behavioral effects of 5-HT2AR activation.

Maternal exercise during pregnancy ameliorates the postnatal neuronal impairments induced by prenatal restraint stress in mice

Clinical and preclinical studies have demonstrated that prenatal stress (PS) induces neuronal and behavioral disturbances in the offspring. In the present study, we determined whether maternal voluntary wheel running (VWR) during pregnancy could reverse the putative deleterious effects of PS on the neurodevelopment and behavior of the offspring. Pregnant CF-1 mice were randomly assigned to control, restraint stressed or restraint stressed+VWR groups. Dams of the stressed group were subjected to restraint stress between gestational days 14 and delivery, while control pregnant dams remained undisturbed in their home cages. Dams of the restraint stressed+VWR group were subjected to exercise between gestational days 1 and 17. On postnatal day 23 (P23), male pups were assigned to one of the following experimental groups: mice born from control dams, stressed dams or stressed+VWR dams. Locomotor behavior and pyramidal neuronal morphology were evaluated at P23. Animals were then sacrificed, and Golgi-impregnated pyramidal neurons of the parietal cortex were morphometrically analyzed. Here, we present two major findings: first, PS produced significantly diminished dendritic growth of parietal neurons without altered locomotor behavior of the offspring; and second, maternal VWR significantly offset morphological impairments.

A cellular mechanism for cortical associations: an organizing principle for the cerebral cortex

A basic feature of intelligent systems such as the cerebral cortex is the ability to freely associate aspects of perceived experience with an internal representation of the world and make predictions about the future. Here, a hypothesis is presented that the extraordinary performance of the cortex derives from an associative mechanism built in at the cellular level to the basic cortical neuronal unit: the pyramidal cell. The mechanism is robustly triggered by coincident input to opposite poles of the neuron, is exquisitely matched to the large- and fine-scale architecture of the cortex, and is tightly controlled by local microcircuits of inhibitory neurons targeting subcellular compartments. This article explores the experimental evidence and the implications for how the cortex operates.

The time window for generation of dendritic spikes by coincidence of action potentials and EPSPs is layer specific in somatosensory cortex

The precise timing of events in the brain has consequences for intracellular processes, synaptic plasticity, integration and network behaviour. Pyramidal neurons, the most widespread excitatory neuron of the neocortex have multiple spike initiation zones, which interact via dendritic and somatic spikes actively propagating in all directions within the dendritic tree. For these neurons, therefore, both the location and timing of synaptic inputs are critical. The time window for which the backpropagating action potential can influence dendritic spike generation has been extensively studied in layer 5 neocortical pyramidal neurons of rat somatosensory cortex. Here, we re-examine this coincidence detection window for pyramidal cell types across the rat somatosensory cortex in layers 2/3, 5 and 6. We find that the time-window for optimal interaction is widest and shifted in layer 5 pyramidal neurons relative to cells in layers 6 and 2/3. Inputs arriving at the same time and locations will therefore differentially affect spike-timing dependent processes in the different classes of pyramidal neurons.

Unsupervised learning of generative and discriminative weights encoding elementary image components in a predictive coding model of cortical function

A method is presented for learning the reciprocal feedforward and feedback connections required by the predictive coding model of cortical function. When this method is used, feedforward and feedback connections are learned simultaneously and independently in a biologically plausible manner. The performance of the proposed algorithm is evaluated by applying it to learning the elementary components of artificial and natural images. For artificial images, the bars problem is employed, and the proposed algorithm is shown to produce state-of-the-art performance on this task. For natural images, components resembling Gabor functions are learned in the first processing stage, and neurons responsive to corners are learned in the second processing stage. The properties of these learned representations are in good agreement with neurophysiological data from V1 and V2. The proposed algorithm demonstrates for the first time that a single computational theory can explain the formation of cortical RFs and also the response properties of cortical neurons once those RFs have been learned.

A single functional model accounts for the distinct properties of suppression in cortical area V1

Cross-orientation suppression and surround suppression have been extensively studied in primary visual cortex (V1). These two forms of suppression have some distinct properties which has led to the suggestion that they are generated by different underlying mechanisms. Furthermore, it has been suggested that mechanisms other than intracortical inhibition may be central to both forms of suppression. A simple computational model (PC/BC), in which intracortical inhibition is fundamental, is shown to simulate the distinct properties of cross-orientation and surround suppression. The same model has previously been shown to account for a large range of V1 response properties including orientation-tuning, spatial and temporal frequency tuning, facilitation and inhibition by flankers and textured surrounds as well as a range of other experimental results on cross-orientation suppression and surround suppression. The current results thus provide additional support for the PC/BC model of V1 and for the proposal that the diverse range of response properties observed in V1 neurons have a single computational explanation. Furthermore, these results demonstrate that current neurophysiological evidence is insufficient to discount intracortical inhibition as a central mechanism underlying both forms of suppression.

PSD-95 is essential for hallucinogen and atypical antipsychotic drug actions at serotonin receptors

Here, we report that postsynaptic density protein of 95 kDa (PSD-95), a postsynaptic density scaffolding protein, classically conceptualized as being essential for the regulation of ionotropic glutamatergic signaling at the postsynaptic membrane, plays an unanticipated and essential role in mediating the actions of hallucinogens and atypical antipsychotic drugs at 5-HT(2A) and 5-HT(2C) serotonergic G-protein-coupled receptors. We show that PSD-95 is crucial for normal 5-HT(2A) and 5-HT(2C) expression in vivo and that PSD-95 maintains normal receptor expression by promoting apical dendritic targeting and stabilizing receptor turnover in vivo. Significantly, 5-HT(2A)- and 5-HT(2C)-mediated downstream signaling is impaired in PSD-95(null) mice, and the 5-HT(2A)-mediated head-twitch response is abnormal. Furthermore, the ability of 5-HT(2A) inverse agonists to normalize behavioral changes induced by glutamate receptor antagonists is abolished in the absence of PSD-95 in vivo. These results demonstrate that PSD-95, in addition to the well known role it plays in scaffolding macromolecular glutamatergic signaling complexes, profoundly modulates metabotropic 5-HT(2A) and 5-HT(2C) receptor function.

Thalamic input to distal apical dendrites in neocortical layer 1 is massive and highly convergent

Input to apical dendritic tufts is now deemed crucial for associative learning, attention, and similar "feedback" interactions in the cerebral cortex. Excitatory input to apical tufts in neocortical layer 1 has been traditionally assumed to be predominantly cortical, as thalamic pathways directed to this layer were regarded relatively scant and diffuse. However, the sensitive tracing methods used in the present study show that, throughout the rat neocortex, large numbers (mean approximately 4500/mm(2)) of thalamocortical neurons converge in layer 1 and that this convergence gives rise to a very high local density of thalamic terminals. Moreover, we show that the layer 1-projecting neurons are present in large numbers in most, but not all, motor, association, limbic, and sensory nuclei of the rodent thalamus. Some layer 1-projecting axons branch to innervate large swaths of the cerebral hemisphere, whereas others arborize within only a single cortical area. Present data imply that realistic modeling of cortical circuitry should factor in a dense axonal canopy carrying highly convergent thalamocortical input to pyramidal cell apical tufts. In addition, they are consistent with the notion that layer 1-projecting axons may be a robust anatomical substrate for extensive "feedback" interactions between cortical areas via the thalamus.

A model of non-linear interactions between cortical top-down and horizontal connections explains the attentional gating of collinear facilitation

Past physiological and psychophysical experiments have shown that attention can modulate the effects of contextual information appearing outside the classical receptive field of a cortical neuron. Specifically, it has been suggested that attention, operating via cortical feedback connections, gates the effects of long-range horizontal connections underlying collinear facilitation in cortical area V1. This article proposes a novel mechanism, based on the computations performed within the dendrites of cortical pyramidal cells, that can account for these observations. Furthermore, it is shown that the top-down gating signal into V1 can result from a process of biased competition occurring in extrastriate cortex. A model based on these two assumptions is used to replicate the results of physiological and psychophysical experiments on collinear facilitation and attentional modulation.

Neural mass model of human multisensory integration

A neural mass model of interacting macro-columns is stimulated to reproduce unisensory, auditory and visually evoked potentials and multisensory (concurrent audiovisual) evoked potentials. These were elicited from patients conducting a reaction response task and recorded from intracranial electrodes placed on the parietal lobe. Important features of this model include inhibitory and excitatory feedback connections to pyramidal cells and extrinsic input to the stellate cell pool, with provision for hierarchical positioning depending on extrinsic connections. Both auditory and visually evoked potentials were best fit using a top-down paradigm. The multisensory response reconstructed from its constituent models was then compared to the actual multisensory EP. Fitting of the multisensory response from constituent models to the actual response required no significant changes to the architecture but did require a decrease in top-down feedback delay. This suggests that multisensory integration, and its related improvement in reaction behavior is not an automatic process but instead controlled by a central executive functioning.

Cue-guided search: a computational model of selective attention

Selective visual attention in a natural environment can be seen as the interaction between the external visual stimulus and task specific knowledge of the required behavior. This interaction between the bottom-up stimulus and the top-down, task-related knowledge is crucial for what is selected in the space and time within the scene. In this paper, we propose a computational model for selective attention for a visual search task. We go beyond simple saliency-based attention models to model selective attention guided by top-down visual cues, which are dynamically integrated with the bottom-up information. In this way, selection of a location is accomplished by interaction between bottom-up and top-down information. First, the general structure of our model is briefly introduced and followed by a description of the top-down processing of task-relevant cues. This is then followed by a description of the processing of the external images to give three feature maps that are combined to give an overall bottom-up map. Second, the development of the formalism for our novel interactive spiking neural network (ISNN) is given, with the interactive activation rule that calculates the integration map. The learning rule for both bottom-up and top-down weight parameters are given, together with some further analysis of the properties of the resulting ISNN. Third, the model is applied to a face detection task to search for the location of a specific face that is cued. The results show that the trajectories of attention are dramatically changed by interaction of information and variations of cues, giving an appropriate, task-relevant search pattern. Finally, we discuss ways in which these results can be seen as compatible with existing psychological evidence.

A fuzzy theory of cortical computation: neuropoietic engrams, fuzzy hypercubes, and the nature of consciousness

We apply fuzzy logic to a theory of memory representation and computation in the human cerebral cortex. The theory termed neuropoiesis is based on the hypothetical transfer of mRNA polyribosomes from the post-synaptic dendritic spine of cortical pyramidal neurons to the presynaptic boutons of connecting axons through a hypothetical process termed retroduction. The net effect of this process is a vast increase in predicted memory storage. The fuzzification of memory engrams permits this expansion in memory capacity and requires multiplex signaling which, in turn, requires the formation of a spike packet whose length is determined by the EEG frequencies generated by the thalamo-cortical rhythm generators. The role of the EEG frequencies is to provide a wavelet-like transform of the multiplex signal which in turn determines the degree of data compression that is required for memory recruitment at a given level of vigilance during cortical computation. In this conceptual model cortical computation is hypothesized to be a form of cluster analysis that can be represented by a fuzzy hypercube in which each dimension of the unit hypercube represents an apical dendrite of a layer 5 pyramidal cell in a cortical fascicle. The tuftal area of the apical dendrite in cortical layer one corresponds to the MIN or zero point of the hypercube's dimension and the cell body in layer 5 corresponds to the MAX or one point of that dimension in the unit hypercube. The neuroanatomical location of synapses on the apical dendrites in the fascicle is mapped onto the fuzzy hypercube. These synapses form clusters composed of both bottom-up and top-down signals represented as metasynaptic fuzzy sets-as-points in the hypercube. Soft winner-take-all gating by inhibitory neurons is proposed to supply the only non-linear operation needed for cortical computation. Feed-forward inhibition is envisioned to play the decisive role of spicing or de-fuzzifying the output signal. Proper transmission of the multiplex signal that carries the fuzzy engram requires synchrony of neuronal firing. For this fuzzy cortical model, synchrony of firing, multiplex signaling, winner-take-all gating, and the known spectrum of EEG frequencies are all derivable from the fundamental mechanism termed synaptopoiesis as described in the theory of neuropoiesis. Finally, this theory predicts that the neural correlate of consciousness must include inhibitory subcortical connections and that its function is largely that of limiting coherence to a narrow range of cortical engrams.

Single-neuron theory of consciousness

By most accounts, the mind arises from the integrated activity of large populations of neurons distributed across multiple brain regions. A contrasting model is presented in the present paper that places the mind/brain interface not at the whole brain level but at the level of single neurons. Specifically, it is proposed that each neuron in the nervous system is independently conscious, with conscious content corresponding to the spatial pattern of a portion of that neuron's dendritic electrical activity. For most neurons, such as those in the hypothalamus or posterior sensory cortices, the conscious activity would be assumed to be simple and unable to directly affect the organism's macroscopic conscious behavior. For a subpopulation of layer 5 pyramidal neurons in the lateral prefrontal cortices, however, an arrangement is proposed to be present such that, at any given moment: (i) the spatial pattern of electrical activity in a portion of the dendritic tree of each neuron in the subpopulation individually manifests a complexity and diversity sufficient to account for the complexity and diversity of conscious experience; (ii) the dendritic trees of the neurons in the subpopulation all contain similar spatial electrical patterns; (iii) the spatial electrical pattern in the dendritic tree of each neuron interacts non-linearly with the remaining ambient dendritic electrical activity to determine the neuron's overall axonal response; (iv) the dendritic spatial pattern is reexpressed at the population level by the spatial pattern exhibited by a synchronously firing subgroup of the conscious neurons, thereby providing a mechanism by which conscious activity at the neuronal level can influence overall behavior. The resulting scheme is one in which conscious behavior appears to be the product of a single macroscopic mind, but is actually the integrated output of a chorus of minds, each associated with a different neuron.

A feedback model of visual attention

Feedback connections are a prominent feature of cortical anatomy and are likely to have a significant functional role in neural information processing. We present a neural network model of cortical feedback that successfully simulates neurophysiological data associated with attention. In this domain, our model can be considered a more detailed, and biologically plausible, implementation of the biased competition model of attention. However, our model is more general as it can also explain a variety of other top-down processes in vision, such as figure/ground segmentation and contextual cueing. This model thus suggests that a common mechanism, involving cortical feedback pathways, is responsible for a range of phenomena and provides a unified account of currently disparate areas of research.