A dopamine gradient controls access to distributed working memory in the large-scale monkey cortex

Visual Attention in the Prefrontal Cortex

Voluntary attention selects behaviorally relevant signals for further processing while filtering out distracter signals. Neural correlates of voluntary visual attention have been reported across multiple areas of the primate visual processing streams, with the earliest and strongest effects isolated in the prefrontal cortex. In this article, I review evidence supporting the hypothesis that signals guiding the allocation of voluntary attention emerge in areas of the prefrontal cortex and reach upstream areas to modulate the processing of incoming visual information according to its behavioral relevance. Areas located anterior and dorsal to the arcuate sulcus and the frontal eye fields produce signals that guide the allocation of spatial attention. Areas located anterior and ventral to the arcuate sulcus produce signals for feature-based attention. Prefrontal microcircuits are particularly suited to supporting voluntary attention because of their ability to generate attentional template signals and implement signal gating and their extensive connectivity with the rest of the brain. Expected final online publication date for the Annual Review of Vision Science, Volume 8 is September 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Training a spiking neuronal network model of visual-motor cortex to play a virtual racket-ball game using reinforcement learning

Recent models of spiking neuronal networks have been trained to perform behaviors in static environments using a variety of learning rules, with varying degrees of biological realism. Most of these models have not been tested in dynamic visual environments where models must make predictions on future states and adjust their behavior accordingly. The models using these learning rules are often treated as black boxes, with little analysis on circuit architectures and learning mechanisms supporting optimal performance. Here we developed visual/motor spiking neuronal network models and trained them to play a virtual racket-ball game using several reinforcement learning algorithms inspired by the dopaminergic reward system. We systematically investigated how different architectures and circuit-motifs (feed-forward, recurrent, feedback) contributed to learning and performance. We also developed a new biologically-inspired learning rule that significantly enhanced performance, while reducing training time. Our models included visual areas encoding game inputs and relaying the information to motor areas, which used this information to learn to move the racket to hit the ball. Neurons in the early visual area relayed information encoding object location and motion direction across the network. Neuronal association areas encoded spatial relationships between objects in the visual scene. Motor populations received inputs from visual and association areas representing the dorsal pathway. Two populations of motor neurons generated commands to move the racket up or down. Model-generated actions updated the environment and triggered reward or punishment signals that adjusted synaptic weights so that the models could learn which actions led to reward. Here we demonstrate that our biologically-plausible learning rules were effective in training spiking neuronal network models to solve problems in dynamic environments. We used our models to dissect the circuit architectures and learning rules most effective for learning. Our model shows that learning mechanisms involving different neural circuits produce similar performance in sensory-motor tasks. In biological networks, all learning mechanisms may complement one another, accelerating the learning capabilities of animals. Furthermore, this also highlights the resilience and redundancy in biological systems.

Extracting Dynamical Understanding From Neural-Mass Models of Mouse Cortex

New brain atlases with high spatial resolution and whole-brain coverage have rapidly advanced our knowledge of the brain's neural architecture, including the systematic variation of excitatory and inhibitory cell densities across the mammalian cortex. But understanding how the brain's microscale physiology shapes brain dynamics at the macroscale has remained a challenge. While physiologically based mathematical models of brain dynamics are well placed to bridge this explanatory gap, their complexity can form a barrier to providing clear mechanistic interpretation of the dynamics they generate. In this work, we develop a neural-mass model of the mouse cortex and show how bifurcation diagrams, which capture local dynamical responses to inputs and their variation across brain regions, can be used to understand the resulting whole-brain dynamics. We show that strong fits to resting-state functional magnetic resonance imaging (fMRI) data can be found in surprisingly simple dynamical regimes-including where all brain regions are confined to a stable fixed point-in which regions are able to respond strongly to variations in their inputs, consistent with direct structural connections providing a strong constraint on functional connectivity in the anesthetized mouse. We also use bifurcation diagrams to show how perturbations to local excitatory and inhibitory coupling strengths across the cortex, constrained by cell-density data, provide spatially dependent constraints on resulting cortical activity, and support a greater diversity of coincident dynamical regimes. Our work illustrates methods for visualizing and interpreting model performance in terms of underlying dynamical mechanisms, an approach that is crucial for building explanatory and physiologically grounded models of the dynamical principles that underpin large-scale brain activity.

Mechanisms of distributed working memory in a large-scale network of macaque neocortex

Neural activity underlying working memory is not a local phenomenon but distributed across multiple brain regions. To elucidate the circuit mechanism of such distributed activity, we developed an anatomically constrained computational model of large-scale macaque cortex. We found that mnemonic internal states may emerge from inter-areal reverberation, even in a regime where none of the isolated areas is capable of generating self-sustained activity. The mnemonic activity pattern along the cortical hierarchy indicates a transition in space, separating areas engaged in working memory and those which do not. A host of spatially distinct attractor states is found, potentially subserving various internal processes. The model yields testable predictions, including the idea of counterstream inhibitory bias, the role of prefrontal areas in controlling distributed attractors, and the resilience of distributed activity to lesions or inactivation. This work provides a theoretical framework for identifying large-scale brain mechanisms and computational principles of distributed cognitive processes.

Differential gene expression in layer 3 pyramidal neurons across 3 regions of the human cortical visual spatial working memory network

Visual spatial working memory (vsWM) is mediated by a distributed cortical network composed of multiple nodes, including primary visual (V1), posterior parietal (PPC), and dorsolateral prefrontal (DLPFC) cortices. Feedforward and feedback information is transferred among these nodes via projections furnished by pyramidal neurons (PNs) located primarily in cortical layer 3. Morphological and electrophysiological differences among layer 3 PNs across these nodes have been reported; however, the transcriptional signatures underlying these differences have not been examined in the human brain. Here we interrogated the transcriptomes of layer 3 PNs from 39 neurotypical human subjects across 3 critical nodes of the vsWM network. Over 8,000 differentially expressed genes were detected, with more than 6,000 transcriptional differences present between layer 3 PNs in V1 and those in PPC and DLPFC. Additionally, over 600 other genes differed in expression along the rostral-to-caudal hierarchy formed by these 3 nodes. Moreover, pathway analysis revealed enrichment of genes in V1 related to circadian rhythms and in DLPFC of genes involved in synaptic plasticity. Overall, these results show robust regional differences in the transcriptome of layer 3 PNs, which likely contribute to regional specialization in their morphological and physiological features and thus in their functional contributions to vsWM.

Big contributions of the little brain for precision psychiatry

Our previous work using 3T functional Magnetic Resonance Imaging (fMRI) parcellated the human dentate nuclei (DN), the primary output of the cerebellum, to three distinct functional zones each contributing uniquely to default-mode, salience-motor, and visual brain networks. In this perspective piece, we highlight the possibility to target specific functional territories within the cerebellum using non-invasive brain stimulation, potentially leading to the refinement of cerebellar-based therapeutics for precision psychiatry. Significant knowledge gap exists in our functional understanding of cerebellar systems. Intervening early, gauging severity of illness, developing intervention strategies and assessing treatment response, are all dependent on our understanding of the cerebello-cerebral networks underlying the pathology of psychotic disorders. A promising yet under-examined avenue for biomarker discovery is disruptions in cerebellar output circuitry. This is primarily because most 3T MRI studies in the past had to exclude cerebellum from the field of view due to limitations in spatiotemporal resolutions. Using recent technological advances in 7T MRI (e.g., parallel transmit head coils) to identify functional territories of the DN, with a focus on dentato-cerebello-thalamo-cortical (CTC) circuitry can lead to better characterization of brain-behavioral correlations and assessments of co-morbidities. Such an improved mechanistic understanding of psychiatric illnesses can reveal aspects of CTC circuitry that can aid in neuroprognosis, identification of subtypes, and generate testable hypothesis for future studies.

Neural Substrates of Visual Perception and Working Memory: Two Sides of the Same Coin or Two Different Coins?

Visual perception occurs when a set of physical signals emanating from the environment enter the visual system and the brain interprets such signals as a percept. Visual working memory occurs when the brain produces and maintains a mental representation of a percept while the physical signals corresponding to that percept are not available. Early studies in humans and non-human primates demonstrated that lesions of the prefrontal cortex impair performance during visual working memory tasks but not during perceptual tasks. These studies attributed a fundamental role in working memory and a lesser role in visual perception to the prefrontal cortex. Indeed, single cell recording studies have found that neurons in the lateral prefrontal cortex of macaques encode working memory representations via persistent firing, validating the results of lesion studies. However, other studies have reported that neurons in some areas of the parietal and temporal lobe-classically associated with visual perception-similarly encode working memory representations via persistent firing. This prompted a line of enquiry about the role of the prefrontal and other associative cortices in working memory and perception. Here, we review evidence from single neuron studies in macaque monkeys examining working memory representations across different areas of the visual hierarchy and link them to studies examining the role of the same areas in visual perception. We conclude that neurons in early visual areas of both ventral (V1-V2-V4) and dorsal (V1-V3-MT) visual pathways of macaques mainly encode perceptual signals. On the other hand, areas downstream from V4 and MT contain subpopulations of neurons that encode both perceptual and/or working memory signals. Differences in cortical architecture (neuronal types, layer composition, and synaptic density and distribution) may be linked to the differential encoding of perceptual and working memory signals between early visual areas and higher association areas.

50 years of mnemonic persistent activity: quo vadis?

Half a century ago persistent spiking activity in the neocortex was discovered to be a neural substrate of working memory. Since then scientists have sought to understand this core cognitive function across biological and computational levels. Studies are reviewed here that cumulatively lend support to a synaptic theory of recurrent circuits for mnemonic persistent activity that depends on various cellular and network substrates and is mathematically described by a multiple-attractor network model. Crucially, a mnemonic attractor state of the brain is consistent with temporal variations and heterogeneity across neurons in a subspace of population activity. Persistent activity should be broadly understood as a contrast to decaying transients. Mechanisms in the absence of neural firing ('activity-silent state') are suitable for passive short-term memory but not for working memory - which is characterized by executive control for filtering out distractors, limited capacity, and internal manipulation of information.

Across-Area Synchronization Supports Feature Integration in a Biophysical Network Model of Working Memory

Working memory function is severely limited. One key limitation that constrains the ability to maintain multiple items in working memory simultaneously is so-called swap errors. These errors occur when an inaccurate response is in fact accurate relative to a non-target stimulus, reflecting the failure to maintain the appropriate association or "binding" between the features that define one object (e.g., color and location). The mechanisms underlying feature binding in working memory remain unknown. Here, we tested the hypothesis that features are bound in memory through synchrony across feature-specific neural assemblies. We built a biophysical neural network model composed of two one-dimensional attractor networks - one for color and one for location - simulating feature storage in different cortical areas. Within each area, gamma oscillations were induced during bump attractor activity through the interplay of fast recurrent excitation and slower feedback inhibition. As a result, different memorized items were held at different phases of the network's oscillation. These two areas were then reciprocally connected via weak cortico-cortical excitation, accomplishing binding between color and location through the synchronization of pairs of bumps across the two areas. Encoding and decoding of color-location associations was accomplished through rate coding, overcoming a long-standing limitation of binding through synchrony. In some simulations, swap errors arose: "color bumps" abruptly changed their phase relationship with "location bumps." This model, which leverages the explanatory power of similar attractor models, specifies a plausible mechanism for feature binding and makes specific predictions about swap errors that are testable at behavioral and neurophysiological levels.