Hierarchical organization of cortical and thalamic connectivity.

The mammalian cortex is a laminar structure containing many areas and cell types that are densely interconnected in complex ways, and for which generalizable principles of organization remain mostly unknown. Here we describe a major expansion of the Allen Mouse Brain Connectivity Atlas resource1, involving around a thousand new tracer experiments in the cortex and its main satellite structure, the thalamus. We used Cre driver lines (mice expressing Cre recombinase) to comprehensively and selectively label brain-wide connections by layer and class of projection neuron. Through observations of axon termination patterns, we have derived a set of generalized anatomical rules to describe corticocortical, thalamocortical and corticothalamic projections. We have built a model to assign connection patterns between areas as either feedforward or feedback, and generated testable predictions of hierarchical positions for individual cortical and thalamic areas and for cortical network modules. Our results show that cell-class-specific connections are organized in a shallow hierarchy within the mouse corticothalamic network.

Cell-type-specific neuromodulation guides synaptic credit assignment in a spiking neural network.

Brains learn tasks via experience-driven differential adjustment of their myriad individual synaptic connections, but the mechanisms that target appropriate adjustment to particular connections remain deeply enigmatic. While Hebbian synaptic plasticity, synaptic eligibility traces, and top-down feedback signals surely contribute to solving this synaptic credit-assignment problem, alone, they appear to be insufficient. Inspired by new genetic perspectives on neuronal signaling architectures, here, we present a normative theory for synaptic learning, where we predict that neurons communicate their contribution to the learning outcome to nearby neurons via cell-type-specific local neuromodulation. Computational tests suggest that neuron-type diversity and neuron-type-specific local neuromodulation may be critical pieces of the biological credit-assignment puzzle. They also suggest algorithms for improved artificial neural network learning efficiency.

Heterogeneity in Neuronal Dynamics Is Learned by Gradient Descent for Temporal Processing Tasks.

Individual neurons in the brain have complex intrinsic dynamics that are highly diverse. We hypothesize that the complex dynamics produced by networks of complex and heterogeneous neurons may contribute to the brain's ability to process and respond to temporally complex data. To study the role of complex and heterogeneous neuronal dynamics in network computation, we develop a rate-based neuronal model, the generalized-leaky-integrate-and-fire-rate (GLIFR) model, which is a rate equivalent of the generalized-leaky-integrate-and-fire model. The GLIFR model has multiple dynamical mechanisms, which add to the complexity of its activity while maintaining differentiability. We focus on the role of after-spike currents, currents induced or modulated by neuronal spikes, in producing rich temporal dynamics. We use machine learning techniques to learn both synaptic weights and parameters underlying intrinsic dynamics to solve temporal tasks. The GLIFR model allows the use of standard gradient descent techniques rather than surrogate gradient descent, which has been used in spiking neural networks. After establishing the ability to optimize parameters using gradient descent in single neurons, we ask how networks of GLIFR neurons learn and perform on temporally challenging tasks, such as sequential MNIST. We find that these networks learn diverse parameters, which gives rise to diversity in neuronal dynamics, as demonstrated by clustering of neuronal parameters. GLIFR networks have mixed performance when compared to vanilla recurrent neural networks, with higher performance in pixel-by-pixel MNIST but lower in line-by-line MNIST. However, they appear to be more robust to random silencing. We find that the ability to learn heterogeneity and the presence of after-spike currents contribute to these gains in performance. Our work demonstrates both the computational robustness of neuronal complexity and diversity in networks and a feasible method of training such models using exact gradients.

The geometry of representational drift in natural and artificial neural networks.

Neurons in sensory areas encode/represent stimuli. Surprisingly, recent studies have suggested that, even during persistent performance, these representations are not stable and change over the course of days and weeks. We examine stimulus representations from fluorescence recordings across hundreds of neurons in the visual cortex using in vivo two-photon calcium imaging and we corroborate previous studies finding that such representations change as experimental trials are repeated across days. This phenomenon has been termed "representational drift". In this study we geometrically characterize the properties of representational drift in the primary visual cortex of mice in two open datasets from the Allen Institute and propose a potential mechanism behind such drift. We observe representational drift both for passively presented stimuli, as well as for stimuli which are behaviorally relevant. Across experiments, the drift differs from in-session variance and most often occurs along directions that have the most in-class variance, leading to a significant turnover in the neurons used for a given representation. Interestingly, despite this significant change due to drift, linear classifiers trained to distinguish neuronal representations show little to no degradation in performance across days. The features we observe in the neural data are similar to properties of artificial neural networks where representations are updated by continual learning in the presence of dropout, i.e. a random masking of nodes/weights, but not other types of noise. Therefore, we conclude that a potential reason for the representational drift in biological networks is driven by an underlying dropout-like noise while continuously learning and that such a mechanism may be computational advantageous for the brain in the same way it is for artificial neural networks, e.g. preventing overfitting.

MouseNet: A biologically constrained convolutional neural network model for the mouse visual cortex.

Convolutional neural networks trained on object recognition derive inspiration from the neural architecture of the visual system in mammals, and have been used as models of the feedforward computation performed in the primate ventral stream. In contrast to the deep hierarchical organization of primates, the visual system of the mouse has a shallower arrangement. Since mice and primates are both capable of visually guided behavior, this raises questions about the role of architecture in neural computation. In this work, we introduce a novel framework for building a biologically constrained convolutional neural network model of the mouse visual cortex. The architecture and structural parameters of the network are derived from experimental measurements, specifically the 100-micrometer resolution interareal connectome, the estimates of numbers of neurons in each area and cortical layer, and the statistics of connections between cortical layers. This network is constructed to support detailed task-optimized models of mouse visual cortex, with neural populations that can be compared to specific corresponding populations in the mouse brain. Using a well-studied image classification task as our working example, we demonstrate the computational capability of this mouse-sized network. Given its relatively small size, MouseNet achieves roughly 2/3rds the performance level on ImageNet as VGG16. In combination with the large scale Allen Brain Observatory Visual Coding dataset, we use representational similarity analysis to quantify the extent to which MouseNet recapitulates the neural representation in mouse visual cortex. Importantly, we provide evidence that optimizing for task performance does not improve similarity to the corresponding biological system beyond a certain point. We demonstrate that the distributions of some physiological quantities are closer to the observed distributions in the mouse brain after task training. We encourage the use of the MouseNet architecture by making the code freely available.

Single Circuit in V1 Capable of Switching Contexts During Movement Using an Inhibitory Population as a Switch.

As animals adapt to their environments, their brains are tasked with processing stimuli in different sensory contexts. Whether these computations are context dependent or independent, they are all implemented in the same neural tissue. A crucial question is what neural architectures can respond flexibly to a range of stimulus conditions and switch between them. This is a particular case of flexible architecture that permits multiple related computations within a single circuit. Here, we address this question in the specific case of the visual system circuitry, focusing on context integration, defined as the integration of feedforward and surround information across visual space. We show that a biologically inspired microcircuit with multiple inhibitory cell types can switch between visual processing of the static context and the moving context. In our model, the VIP population acts as the switch and modulates the visual circuit through a disinhibitory motif. Moreover, the VIP population is efficient, requiring only a relatively small number of neurons to switch contexts. This circuit eliminates noise in videos by using appropriate lateral connections for contextual spatiotemporal surround modulation, having superior denoising performance compared to circuits where only one context is learned. Our findings shed light on a minimally complex architecture that is capable of switching between two naturalistic contexts using few switching units.

Adaptation supports short-term memory in a visual change detection task.

The maintenance of short-term memories is critical for survival in a dynamically changing world. Previous studies suggest that this memory can be stored in the form of persistent neural activity or using a synaptic mechanism, such as with short-term plasticity. Here, we compare the predictions of these two mechanisms to neural and behavioral measurements in a visual change detection task. Mice were trained to respond to changes in a repeated sequence of natural images while neural activity was recorded using two-photon calcium imaging. We also trained two types of artificial neural networks on the same change detection task as the mice. Following fixed pre-processing using a pretrained convolutional neural network, either a recurrent neural network (RNN) or a feedforward neural network with short-term synaptic depression (STPNet) was trained to the same level of performance as the mice. While both networks are able to learn the task, the STPNet model contains units whose activity are more similar to the in vivo data and produces errors which are more similar to the mice. When images are omitted, an unexpected perturbation which was absent during training, mice often do not respond to the omission but are more likely to respond to the subsequent image. Unlike the RNN model, STPNet produces a similar pattern of behavior. These results suggest that simple neural adaptation mechanisms may serve as an important bottom-up memory signal in this task, which can be used by downstream areas in the decision-making process.

Brain Modeling ToolKit: An open source software suite for multiscale modeling of brain circuits.

Experimental studies in neuroscience are producing data at a rapidly increasing rate, providing exciting opportunities and formidable challenges to existing theoretical and modeling approaches. To turn massive datasets into predictive quantitative frameworks, the field needs software solutions for systematic integration of data into realistic, multiscale models. Here we describe the Brain Modeling ToolKit (BMTK), a software suite for building models and performing simulations at multiple levels of resolution, from biophysically detailed multi-compartmental, to point-neuron, to population-statistical approaches. Leveraging the SONATA file format and existing software such as NEURON, NEST, and others, BMTK offers a consistent user experience across multiple levels of resolution. It permits highly sophisticated simulations to be set up with little coding required, thus lowering entry barriers to new users. We illustrate successful applications of BMTK to large-scale simulations of a cortical area. BMTK is an open-source package provided as a resource supporting modeling-based discovery in the community.

Synchronization dependent on spatial structures of a mesoscopic whole-brain network.

Complex structural connectivity of the mammalian brain is believed to underlie the versatility of neural computations. Many previous studies have investigated properties of small subsystems or coarse connectivity among large brain regions that are often binarized and lack spatial information. Yet little is known about spatial embedding of the detailed whole-brain connectivity and its functional implications. We focus on closing this gap by analyzing how spatially-constrained neural connectivity shapes synchronization of the brain dynamics based on a system of coupled phase oscillators on a mammalian whole-brain network at the mesoscopic level. This was made possible by the recent development of the Allen Mouse Brain Connectivity Atlas constructed from viral tracing experiments together with a new mapping algorithm. We investigated whether the network can be compactly represented based on the spatial dependence of the network topology. We found that the connectivity has a significant spatial dependence, with spatially close brain regions strongly connected and distal regions weakly connected, following a power law. However, there are a number of residuals above the power-law fit, indicating connections between brain regions that are stronger than predicted by the power-law relationship. By measuring the sensitivity of the network order parameter, we show how these strong connections dispersed across multiple spatial scales of the network promote rapid transitions between partial synchronization and more global synchronization as the global coupling coefficient changes. We further demonstrate the significance of the locations of the residual connections, suggesting a possible link between the network complexity and the brain's exceptional ability to swiftly switch computational states depending on stimulus and behavioral context.

A mesoscale connectome of the mouse brain.

Comprehensive knowledge of the brain's wiring diagram is fundamental for understanding how the nervous system processes information at both local and global scales. However, with the singular exception of the C. elegans microscale connectome, there are no complete connectivity data sets in other species. Here we report a brain-wide, cellular-level, mesoscale connectome for the mouse. The Allen Mouse Brain Connectivity Atlas uses enhanced green fluorescent protein (EGFP)-expressing adeno-associated viral vectors to trace axonal projections from defined regions and cell types, and high-throughput serial two-photon tomography to image the EGFP-labelled axons throughout the brain. This systematic and standardized approach allows spatial registration of individual experiments into a common three dimensional (3D) reference space, resulting in a whole-brain connectivity matrix. A computational model yields insights into connectional strength distribution, symmetry and other network properties. Virtual tractography illustrates 3D topography among interconnected regions. Cortico-thalamic pathway analysis demonstrates segregation and integration of parallel pathways. The Allen Mouse Brain Connectivity Atlas is a freely available, foundational resource for structural and functional investigations into the neural circuits that support behavioural and cognitive processes in health and disease.

Visual physiology of the layer 4 cortical circuit in silico.

Despite advances in experimental techniques and accumulation of large datasets concerning the composition and properties of the cortex, quantitative modeling of cortical circuits under in-vivo-like conditions remains challenging. Here we report and publicly release a biophysically detailed circuit model of layer 4 in the mouse primary visual cortex, receiving thalamo-cortical visual inputs. The 45,000-neuron model was subjected to a battery of visual stimuli, and results were compared to published work and new in vivo experiments. Simulations reproduced a variety of observations, including effects of optogenetic perturbations. Critical to the agreement between responses in silico and in vivo were the rules of functional synaptic connectivity between neurons. Interestingly, after extreme simplification the model still performed satisfactorily on many measurements, although quantitative agreement with experiments suffered. These results emphasize the importance of functional rules of cortical wiring and enable a next generation of data-driven models of in vivo neural activity and computations.

Rationalizing spatial exploration patterns of wild animals and humans through a temporal discounting framework.

Understanding the exploration patterns of foragers in the wild provides fundamental insight into animal behavior. Recent experimental evidence has demonstrated that path lengths (distances between consecutive turns) taken by foragers are well fitted by a power law distribution. Numerous theoretical contributions have posited that "Lévy random walks"-which can produce power law path length distributions-are optimal for memoryless agents searching a sparse reward landscape. It is unclear, however, whether such a strategy is efficient for cognitively complex agents, from wild animals to humans. Here, we developed a model to explain the emergence of apparent power law path length distributions in animals that can learn about their environments. In our model, the agent's goal during search is to build an internal model of the distribution of rewards in space that takes into account the cost of time to reach distant locations (i.e., temporally discounting rewards). For an agent with such a goal, we find that an optimal model of exploration in fact produces hyperbolic path lengths, which are well approximated by power laws. We then provide support for our model by showing that humans in a laboratory spatial exploration task search space systematically and modify their search patterns under a cost of time. In addition, we find that path length distributions in a large dataset obtained from free-ranging marine vertebrates are well described by our hyperbolic model. Thus, we provide a general theoretical framework for understanding spatial exploration patterns of cognitively complex foragers.

Inferring cortical function in the mouse visual system through large-scale systems neuroscience.

The scientific mission of the Project MindScope is to understand neocortex, the part of the mammalian brain that gives rise to perception, memory, intelligence, and consciousness. We seek to quantitatively evaluate the hypothesis that neocortex is a relatively homogeneous tissue, with smaller functional modules that perform a common computational function replicated across regions. We here focus on the mouse as a mammalian model organism with genetics, physiology, and behavior that can be readily studied and manipulated in the laboratory. We seek to describe the operation of cortical circuitry at the computational level by comprehensively cataloging and characterizing its cellular building blocks along with their dynamics and their cell type-specific connectivities. The project is also building large-scale experimental platforms (i.e., brain observatories) to record the activity of large populations of cortical neurons in behaving mice subject to visual stimuli. A primary goal is to understand the series of operations from visual input in the retina to behavior by observing and modeling the physical transformations of signals in the corticothalamic system. We here focus on the contribution that computer modeling and theory make to this long-term effort.

A Comparison of Visual Response Properties in the Lateral Geniculate Nucleus and Primary Visual Cortex of Awake and Anesthetized Mice.

The cerebral cortex of the mouse has become one of the most important systems for studying information processing and the neural correlates of behavior. Multiple studies have examined the first stages of visual cortical processing: primary visual cortex (V1) and its thalamic inputs from the dorsal lateral geniculate nucleus (dLGN), but more rarely in the lateral posterior nucleus (LP) in mice. Multiple single-unit surveys of dLGN and V1, both with electrophysiology and two-photon calcium imaging, have described receptive fields in anesthetized animals. Increasingly, awake animals are being used in physiological studies, so it is important to compare neuronal responses between awake and anesthetized state. We have performed a comprehensive survey of spatial and temporal response properties in V1, dLGN, and lateral posterior nucleus of both anesthetized and awake animals, using a common set of stimuli: drifting sine-wave gratings spanning a broad range of spatial and temporal parameters, and sparse noise stimuli consisting of flashed light and dark squares. Most qualitative receptive field parameters were found to be unchanged between the two states, such as most aspects of spatial processing, but there were significant differences in several parameters, most notably in temporal processing. Compared with anesthetized animals, the temporal frequency that evoked the peak response was shifted toward higher values in the dLGN of awake mice and responses were more sustained. Further, the peak response to a flashed stimulus was earlier in all three areas. Overall, however, receptive field properties in the anesthetized animal remain a good model for those in the awake animal. SIGNIFICANCE STATEMENT: The primary visual cortex (V1) of the mouse and its inputs from visual thalamus (dLGN), have become a dominant model for studying information processing in the brain. Early surveys of visual response properties (receptive fields) were performed in anesthetized animals. Although most recent studies of V1 have been performed in awake animals to examine links between vision and behavior, there have been few comprehensive studies of receptive field properties in the awake mouse, especially in dLGN and lateral posterior nucleus. We have performed a comparative survey of receptive fields in dLGN, lateral posterior nucleus, and V1 in anesthetized and awake mice. We found multiple differences in processing of time-varying stimuli, whereas the spatial aspects of receptive fields remain comparatively unchanged.

Temporal correlation mechanisms and their role in feature selection: a single-unit study in primate somatosensory cortex.

Studies in vision show that attention enhances the firing rates of cells when it is directed towards their preferred stimulus feature. However, it is unknown whether other sensory systems employ this mechanism to mediate feature selection within their modalities. Moreover, whether feature-based attention modulates the correlated activity of a population is unclear. Indeed, temporal correlation codes such as spike-synchrony and spike-count correlations (r(sc)) are believed to play a role in stimulus selection by increasing the signal and reducing the noise in a population, respectively. Here, we investigate (1) whether feature-based attention biases the correlated activity between neurons when attention is directed towards their common preferred feature, (2) the interplay between spike-synchrony and rsc during feature selection, and (3) whether feature attention effects are common across the visual and tactile systems. Single-unit recordings were made in secondary somatosensory cortex of three non-human primates while animals engaged in tactile feature (orientation and frequency) and visual discrimination tasks. We found that both firing rate and spike-synchrony between neurons with similar feature selectivity were enhanced when attention was directed towards their preferred feature. However, attention effects on spike-synchrony were twice as large as those on firing rate, and had a tighter relationship with behavioral performance. Further, we observed increased r(sc) when attention was directed towards the visual modality (i.e., away from touch). These data suggest that similar feature selection mechanisms are employed in vision and touch, and that temporal correlation codes such as spike-synchrony play a role in mediating feature selection. We posit that feature-based selection operates by implementing multiple mechanisms that reduce the overall noise levels in the neural population and synchronize activity across subpopulations that encode the relevant features of sensory stimuli.

The Computational Properties of a Simplified Cortical Column Model.

The mammalian neocortex has a repetitious, laminar structure and performs functions integral to higher cognitive processes, including sensory perception, memory, and coordinated motor output. What computations does this circuitry subserve that link these unique structural elements to their function? Potjans and Diesmann (2014) parameterized a four-layer, two cell type (i.e. excitatory and inhibitory) model of a cortical column with homogeneous populations and cell type dependent connection probabilities. We implement a version of their model using a displacement integro-partial differential equation (DiPDE) population density model. This approach, exact in the limit of large homogeneous populations, provides a fast numerical method to solve equations describing the full probability density distribution of neuronal membrane potentials. It lends itself to quickly analyzing the mean response properties of population-scale firing rate dynamics. We use this strategy to examine the input-output relationship of the Potjans and Diesmann cortical column model to understand its computational properties. When inputs are constrained to jointly and equally target excitatory and inhibitory neurons, we find a large linear regime where the effect of a multi-layer input signal can be reduced to a linear combination of component signals. One of these, a simple subtractive operation, can act as an error signal passed between hierarchical processing stages.

Large-scale topology and the default mode network in the mouse connectome.

Noninvasive functional imaging holds great promise for serving as a translational bridge between human and animal models of various neurological and psychiatric disorders. However, despite a depth of knowledge of the cellular and molecular underpinnings of atypical processes in mouse models, little is known about the large-scale functional architecture measured by functional brain imaging, limiting translation to human conditions. Here, we provide a robust processing pipeline to generate high-resolution, whole-brain resting-state functional connectivity MRI (rs-fcMRI) images in the mouse. Using a mesoscale structural connectome (i.e., an anterograde tracer mapping of axonal projections across the mouse CNS), we show that rs-fcMRI in the mouse has strong structural underpinnings, validating our procedures. We next directly show that large-scale network properties previously identified in primates are present in rodents, although they differ in several ways. Last, we examine the existence of the so-called default mode network (DMN)--a distributed functional brain system identified in primates as being highly important for social cognition and overall brain function and atypically functionally connected across a multitude of disorders. We show the presence of a potential DMN in the mouse brain both structurally and functionally. Together, these studies confirm the presence of basic network properties and functional networks of high translational importance in structural and functional systems in the mouse brain. This work clears the way for an important bridge measurement between human and rodent models, enabling us to make stronger conclusions about how regionally specific cellular and molecular manipulations in mice relate back to humans.

Analytical Calculation of Errors in Time and Value Perception Due to a Subjective Time Accumulator: A Mechanistic Model and the Generation of Weber's Law.

It has been previously shown (Namboodiri, Mihalas, Marton, & Hussain Shuler, 2014) that an evolutionary theory of decision making and time perception is capable of explaining numerous behavioral observations regarding how humans and animals decide between differently delayed rewards of differing magnitudes and how they perceive time. An implementation of this theory using a stochastic drift-diffusion accumulator model (Namboodiri, Mihalas, & Hussain Shuler, 2014a) showed that errors in time perception and decision making approximately obey Weber's law for a range of parameters. However, prior calculations did not have a clear mechanistic underpinning. Further, these calculations were only approximate, with the range of parameters being limited. In this letter, we provide a full analytical treatment of such an accumulator model, along with a mechanistic implementation, to calculate the expression of these errors for the entirety of the parameter space. In our mechanistic model, Weber's law results from synaptic facilitation and depression within the feedback synapses of the accumulator. Our theory also makes the prediction that the steepness of temporal discounting can be affected by requiring the precise timing of temporal intervals. Thus, by presenting exact quantitative calculations, this work provides falsifiable predictions for future experimental testing.

The influence of synaptic weight distribution on neuronal population dynamics.

The manner in which different distributions of synaptic weights onto cortical neurons shape their spiking activity remains open. To characterize a homogeneous neuronal population, we use the master equation for generalized leaky integrate-and-fire neurons with shot-noise synapses. We develop fast semi-analytic numerical methods to solve this equation for either current or conductance synapses, with and without synaptic depression. We show that its solutions match simulations of equivalent neuronal networks better than those of the Fokker-Planck equation and we compute bounds on the network response to non-instantaneous synapses. We apply these methods to study different synaptic weight distributions in feed-forward networks. We characterize the synaptic amplitude distributions using a set of measures, called tail weight numbers, designed to quantify the preponderance of very strong synapses. Even if synaptic amplitude distributions are equated for both the total current and average synaptic weight, distributions with sparse but strong synapses produce higher responses for small inputs, leading to a larger operating range. Furthermore, despite their small number, such synapses enable the network to respond faster and with more stability in the face of external fluctuations.

Mechanisms of perceptual organization provide auto-zoom and auto-localization for attention to objects.

Visual attention is often understood as a modulatory field acting at early stages of processing, but the mechanisms that direct and fit the field to the attended object are not known. We show that a purely spatial attention field propagating downward in the neuronal network responsible for perceptual organization will be reshaped, repositioned, and sharpened to match the object's shape and scale. Key features of the model are grouping neurons integrating local features into coherent tentative objects, excitatory feedback to the same local feature neurons that caused grouping neuron activation, and inhibition between incompatible interpretations both at the local feature level and at the object representation level.