A network that uses few active neurones to code visual input predicts the diverse shapes of cortical receptive fields

Computational models of primary visual cortex have demonstrated that principles of efficient coding and neuronal sparseness can explain the emergence of neurones with localised oriented receptive fields. Yet, existing models have failed to predict the diverse shapes of receptive fields that occur in nature. The existing models used a particular "soft" form of sparseness that limits average neuronal activity. Here we study models of efficient coding in a broader context by comparing soft and "bard" forms of neuronal sparseness. As a result of our analyses, we propose a novel network model for visual cortex. The model forms efficient visual representations in which the number of active neurones, rather than mean neuronal activity, is limited. This form of hard sparseness also economises cortical resources like synaptic memory and metabolic energy. Furthermore, our model accurately predicts the distribution of receptive field shapes found in the primary visual cortex of cat and monkey.     

A sparse coding model with synaptically local plasticity and spiking neurons can account for the diverse shapes of V1 simple cell receptive fields

Sparse coding algorithms trained on natural images can accurately predict the features that excite visual cortical neurons, but it is not known whether such codes can be learned using biologically realistic plasticity rules. We have developed a biophysically motivated spiking network, relying solely on synaptically local information, that can predict the full diversity of V1 simple cell receptive field shapes when trained on natural images. This represents the first demonstration that sparse coding principles, operating within the constraints imposed by cortical architecture, can successfully reproduce these receptive fields. We further prove, mathematically, that sparseness and decorrelation are the key ingredients that allow for synaptically local plasticity rules to optimize a cooperative, linear generative image model formed by the neural representation. Finally, we discuss several interesting emergent properties of our network, with the intent of bridging the gap between theoretical and experimental studies of visual cortex.     

Emergence of Functional Specificity in Balanced Networks with Synaptic Plasticity

In rodent visual cortex, synaptic connections between orientation-selective neurons are unspecific at the time of eye opening, and become to some degree functionally specific only later during development. An explanation for this two-stage process was proposed in terms of Hebbian plasticity based on visual experience that would eventually enhance connections between neurons with similar response features. For this to work, however, two conditions must be satisfied: First, orientation selective neuronal responses must exist before specific recurrent synaptic connections can be established. Second, Hebbian learning must be compatible with the recurrent network dynamics contributing to orientation selectivity, and the resulting specific connectivity must remain stable for unspecific background activity. Previous studies have mainly focused on very simple models, where the receptive fields of neurons were essentially determined by feedforward mechanisms, and where the recurrent network was small, lacking the complex recurrent dynamics of large-scale networks of excitatory and inhibitory neurons. Here we studied the emergence of functionally specific connectivity in large-scale recurrent networks with synaptic plasticity. Our results show that balanced random networks, which already exhibit highly selective responses at eye opening, can develop feature-specific connectivity if appropriate rules of synaptic plasticity are invoked within and between excitatory and inhibitory populations. If these conditions are met, the initial orientation selectivity guides the process of Hebbian learning and, as a result, functionally specific and a surplus of bidirectional connections emerge. Our results thus demonstrate the cooperation of synaptic plasticity and recurrent dynamics in large-scale functional networks with realistic receptive fields, highlight the role of inhibition as a critical element in this process, and paves the road for further computational studies of sensory processing in neocortical network models equipped with synaptic plasticity.     

Functional clustering drives encoding improvement in a developing brain network during awake visual learning

Sensory experience drives dramatic structural and functional plasticity in developing neurons. However, for single-neuron plasticity to optimally improve whole-network encoding of sensory information, changes must be coordinated between neurons to ensure a full range of stimuli is efficiently represented. Using two-photon calcium imaging to monitor evoked activity in over 100 neurons simultaneously, we investigate network-level changes in the developing Xenopus laevis tectum during visual training with motion stimuli. Training causes stimulus-specific changes in neuronal responses and interactions, resulting in improved population encoding. This plasticity is spatially structured, increasing tuning curve similarity and interactions among nearby neurons, and decreasing interactions among distant neurons. Training does not improve encoding by single clusters of similarly responding neurons, but improves encoding across clusters, indicating coordinated plasticity across the network. NMDA receptor blockade prevents coordinated plasticity, reduces clustering, and abolishes whole-network encoding improvement. We conclude that NMDA receptors support experience-dependent network self-organization, allowing efficient population coding of a diverse range of stimuli.     

AIM: A network model of attention in auditory cortex

Attentional modulation of cortical networks is critical for the cognitive flexibility required to process complex scenes. Current theoretical frameworks for attention are based almost exclusively on studies in visual cortex, where attentional effects are typically modest and excitatory. In contrast, attentional effects in auditory cortex can be large and suppressive. A theoretical framework for explaining attentional effects in auditory cortex is lacking, preventing a broader understanding of cortical mechanisms underlying attention. Here, we present a cortical network model of attention in primary auditory cortex (A1). A key mechanism in our network is attentional inhibitory modulation (AIM) of cortical inhibitory neurons. In this mechanism, top-down inhibitory neurons disinhibit bottom-up cortical circuits, a prominent circuit motif observed in sensory cortex. Our results reveal that the same underlying mechanisms in the AIM network can explain diverse attentional effects on both spatial and frequency tuning in A1. We find that a dominant effect of disinhibition on cortical tuning is suppressive, consistent with experimental observations. Functionally, the AIM network may play a key role in solving the cocktail party problem. We demonstrate how attention can guide the AIM network to monitor an acoustic scene, select a specific target, or switch to a different target, providing flexible outputs for solving the cocktail party problem.     

Attention Narrows Position Tuning of Population Responses in V1

When attention is directed to a region of space, visual resolution at that location flexibly adapts, becoming sharper to resolve fine-scale details or coarser to reflect large-scale texture and surface properties [1]. By what mechanism does attention improve spatial resolution? An improved signal-to-noise ratio (SNR) at the attended location contributes [2], because of retinotopically specific signal gain [3], [4], [5], [6], [7], [8], [9] and [10]. Additionally, attention could sharpen position tuning at the neural population level, so that adjacent objects activate more distinct regions of the visual cortex. A dual mechanism involving both signal gain and sharpened position tuning would be highly efficient at improving visual resolution, but there is no direct evidence that attention can narrow the position tuning of population responses. Here, we compared the spatial spread of the fMRI BOLD response for attended versus ignored stimuli. The activity produced by adjacent stimuli overlapped less when subjects were attending at their locations versus attending elsewhere, despite a stronger peak response with attention. Our results show that even as early as primary visual cortex (V1), spatially directed attention narrows the tuning of population-coded position representations.     

A sparse object coding scheme in area V4

Sparse coding has long been recognized as a primary goal of image transformation in the visual system. Sparse coding in early visual cortex is achieved by abstracting local oriented spatial frequencies and by excitatory/inhibitory surround modulation. Object responses are thought to be sparse at subsequent processing stages, but neural mechanisms for higher-level sparsification are not known. Here, convergent results from macaque area V4 neural recording and simulated V4 populations trained on natural object contours suggest that sparse coding is achieved in midlevel visual cortex by emphasizing representation of acute convex and concave curvature. We studied 165 V4 neurons with a random, adaptive stimulus strategy to minimize bias and explore an unlimited range of contour shapes. V4 responses were strongly weighted toward contours containing acute convex or concave curvature. In contrast, the tuning distribution in nonsparse simulated V4 populations was strongly weighted toward low curvature. But as sparseness constraints increased, the simulated tuning distribution shifted progressively toward more acute convex and concave curvature, matching the neural recording results. These findings indicate a sparse object coding scheme in midlevel visual cortex based on uncommon but diagnostic regions of acute contour curvature.     

Development of neural circuitry for precise temporal sequences through spontaneous activity, axon remodeling, and synaptic plasticity

Temporally precise sequences of neuronal spikes that span hundreds of milliseconds are observed in many brain areas, including songbird premotor nucleus, cat visual cortex, and primary motor cortex. Synfire chains-networks in which groups of neurons are connected via excitatory synapses into a unidirectional chain-are thought to underlie the generation of such sequences. It is unknown, however, how synfire chains can form in local neural circuits, especially for long chains. Here, we show through computer simulation that long synfire chains can develop through spike-time dependent synaptic plasticity and axon remodeling-the pruning of prolific weak connections that follows the emergence of a finite number of strong connections. The formation process begins with a random network. A subset of neurons, called training neurons, intermittently receive superthreshold external input. Gradually, a synfire chain emerges through a recruiting process, in which neurons within the network connect to the tail of the chain started by the training neurons. The model is robust to varying parameters, as well as natural events like neuronal turnover and massive lesions. Our model suggests that long synfire chain can form during the development through self-organization, and axon remodeling, ubiquitous in developing neural circuits, is essential in the process.     

Attention to stimulus features shifts spectral tuning of V4 neurons during natural vision

Previous neurophysiological studies suggest that attention can alter the baseline or gain of neurons in extrastriate visual areas but that it cannot change tuning. This suggests that neurons in visual cortex function as labeled lines whose meaning does not depend on task demands. To test this common assumption, we used a system identification approach to measure spatial frequency and orientation tuning in area V4 during two attentionally demanding visual search tasks, one that required fixation and one that allowed free viewing during search. We found that spatial attention modulates response baseline and gain but does not alter tuning, consistent with previous reports. In contrast, feature-based attention often shifts neuronal tuning. These tuning shifts are inconsistent with the labeled-line model and tend to enhance responses to stimulus features that distinguish the search target. Our data suggest that V4 neurons behave as matched filters that are dynamically tuned to optimize visual search.     

Molecular regulation of cognitive functions and developmental plasticity: impact of GABAA receptors

By controlling spike timing and sculpting neuronal rhythms, inhibitory interneurons play a key role in regulating neuronal circuits and behavior. The pronounced diversity of GABAergic (gamma-aminobutyric acid) interneurons is paralleled by an extensive diversity of GABAA receptor subtypes. The region- and domain-specific location of these receptor subtypes offers the opportunity to gain functional insights into the role of defined neuronal circuits. These developments are reviewed with regard to the regulation of sleep, anxiety, memory, sensorimotor processing and post-natal developmental plasticity.     

Origin of orientation tuning in the visual cortex.

The results of recent experiments have thrown new light on the neuronal connections underlying orientation-selective responses in the primary visual cortex of adult animals. The pattern of afferent input from the lateral geniculate nucleus to the cortex appears to be specific for orientation, while intracortical inhibitory connections appear to be non-specific in this respect. Experimental and theoretical studies have suggested that the development of cortical cell orientation tuning is an activity-dependent process.     

Bias-free estimation of information content in temporally sparse neuronal activity

Applying information theoretic measures to neuronal activity data enables the quantification of neuronal encoding quality. However, when the sample size is limited, a naïve estimation of the information content typically contains a systematic overestimation (upward bias), which may lead to misinterpretation of coding characteristics. This bias is exacerbated in Ca²⁺ imaging because of the temporal sparsity of elevated Ca²⁺ signals. Here, we introduce methods to correct for the bias in the naïve estimation of information content from limited sample sizes and temporally sparse neuronal activity. We demonstrate the higher accuracy of our methods over previous ones, when applied to Ca²⁺ imaging data recorded from the mouse hippocampus and primary visual cortex, as well as to simulated data with matching tuning properties and firing statistics. Our bias-correction methods allowed an accurate estimation of the information place cells carry about the animal’s position (spatial information) and uncovered the spatial resolution of hippocampal coding. Furthermore, using our methods, we found that cells with higher peak firing rates carry higher spatial information per spike and exposed differences between distinct hippocampal subfields in the long-term evolution of the spatial code. These results could be masked by the bias when applying the commonly used naïve calculation of information content. Thus, a bias-free estimation of information content can uncover otherwise overlooked properties of the neural code.     

Adaptation-induced plasticity of orientation tuning in adult visual cortex

A key emergent property of the primary visual cortex (V1) is the orientation selectivity of its neurons. The extent to which adult visual cortical neurons can exhibit changes in orientation selectivity is unknown. Here we use single-unit recording and intrinsic signal imaging in V1 of adult cats to demonstrate systematic repulsive shifts in orientation preference following short-term exposure (adaptation) to one stimulus orientation. In contrast to the common view of adaptation as a passive process by which responses around the adapting orientation are reduced, we show that changes in orientation tuning also occur due to response increases at orientations away from the adapting stimulus. Adaptation-induced orientation plasticity is thus an active time-dependent process that involves network interactions and includes both response depression and enhancement.     

Learning-induced improvement in encoding and decoding of specific movement directions by neurons in the primary motor cortex

Many recent studies describe learning-related changes in sensory and motor areas, but few have directly probed for improvement in neuronal coding after learning. We used information theory to analyze single-cell activity from the primary motor cortex of monkeys, before and after learning a local rotational visuomotor task. We show that after learning, neurons in the primary motor cortex conveyed more information about the direction of movement and did so with relation to their directional sensitivity. Similar to recent findings in sensory systems, this specific improvement in encoding is correlated with an increase in the slope of the neurons' tuning curve. We further demonstrate that the improved information after learning enables a more accurate reconstruction of movement direction from neuronal populations. Our results suggest that similar mechanisms govern learning in sensory and motor areas and provide further evidence for a tight relationship between the locality of learning and the properties of neurons; namely, cells only show plasticity if their preferred direction is near the training one. The results also suggest that simple learning tasks can enhance the performance of brain–machine interfaces.     

Rapid adaptation and efficient coding

Rapid adaptation is a prominent feature of biological neuronal systems. From a functional perspective the adaptation of neuronal properties, namely the input-output relation of sensory neurons, is usually interpreted as an adaptation of the sensory system to changing environments as characterized by their stimulus statistics. Here we argue that this interpretation is only applicable as long as the adaptation processes are slower than the time-scale at which the stimulus statistics change. We present a definition of optimality of a neuronal code which still captures the idea of efficient coding, but which can also explain rapid adaptation without referring to an adaptation to different sensory environments. Finally, we apply our new idea to a simple model of an orientation hypercolumn in the primary visual cortex and predict that the interactions between orientation columns should adapt at the time-scale of a single stimulus presentation.     

Balancing feed-forward excitation and inhibition via Hebbian inhibitory synaptic plasticity

It has been suggested that excitatory and inhibitory inputs to cortical cells are balanced, and that this balance is important for the highly irregular firing observed in the cortex. There are two hypotheses as to the origin of this balance. One assumes that it results from a stable solution of the recurrent neuronal dynamics. This model can account for a balance of steady state excitation and inhibition without fine tuning of parameters, but not for transient inputs. The second hypothesis suggests that the feed forward excitatory and inhibitory inputs to a postsynaptic cell are already balanced. This latter hypothesis thus does account for the balance of transient inputs. However, it remains unclear what mechanism underlies the fine tuning required for balancing feed forward excitatory and inhibitory inputs. Here we investigated whether inhibitory synaptic plasticity is responsible for the balance of transient feed forward excitation and inhibition. We address this issue in the framework of a model characterizing the stochastic dynamics of temporally anti-symmetric Hebbian spike timing dependent plasticity of feed forward excitatory and inhibitory synaptic inputs to a single post-synaptic cell. Our analysis shows that inhibitory Hebbian plasticity generates 'negative feedback' that balances excitation and inhibition, which contrasts with the 'positive feedback' of excitatory Hebbian synaptic plasticity. As a result, this balance may increase the sensitivity of the learning dynamics to the correlation structure of the excitatory inputs.     

Local non-linear interactions in the visual cortex may reflect global decorrelation

The classical receptive field in the primary visual cortex have been successfully explained by sparse activation of relatively independent units, whose tuning properties reflect the statistical dependencies in the natural environment. Robust surround modulation, emerging from stimulation beyond the classical receptive field, has been associated with increase of lifetime sparseness in the V1, but the system-wide modulation of response strength have currently no theoretical explanation. We measured fMRI responses from human visual cortex and quantified the contextual modulation with a decorrelation coefficient (d), derived from a subtractive normalization model. All active cortical areas demonstrated local non-linear summation of responses, which were in line with hypothesis of global decorrelation of voxels responses. In addition, we found sensitivity to surrounding stimulus structure across the ventral stream, and large-scale sensitivity to the number of simultaneous objects. Response sparseness across voxel population increased consistently with larger stimuli. These data suggest that contextual modulation for a stimulus event reflect optimization of the code and perhaps increase in energy efficiency throughout the ventral stream hierarchy. Our model provides a novel prediction that average suppression of response amplitude for simultaneous stimuli across the cortical network is a monotonic function of similarity of response strengths in the network when the stimuli are presented alone.     

The Role of Thalamic Population Synchrony in the Emergence of Cortical Feature Selectivity

In a wide range of studies, the emergence of orientation selectivity in primary visual cortex has been attributed to a complex interaction between feed-forward thalamic input and inhibitory mechanisms at the level of cortex. Although it is well known that layer 4 cortical neurons are highly sensitive to the timing of thalamic inputs, the role of the stimulus-driven timing of thalamic inputs in cortical orientation selectivity is not well understood. Here we show that the synchronization of thalamic firing contributes directly to the orientation tuned responses of primary visual cortex in a way that optimizes the stimulus information per cortical spike. From the recorded responses of geniculate X-cells in the anesthetized cat, we synthesized thalamic sub-populations that would likely serve as the synaptic input to a common layer 4 cortical neuron based on anatomical constraints. We used this synchronized input as the driving input to an integrate-and-fire model of cortical responses and demonstrated that the tuning properties match closely to those measured in primary visual cortex. By modulating the overall level of synchronization at the preferred orientation, we show that efficiency of information transmission in the cortex is maximized for levels of synchronization which match those reported in thalamic recordings in response to naturalistic stimuli, a property which is relatively invariant to the orientation tuning width. These findings indicate evidence for a more prominent role of the feed-forward thalamic input in cortical feature selectivity based on thalamic synchronization.     

Slow Feature Analysis on Retinal Waves Leads to V1 Complex Cells

The developing visual system of many mammalian species is partially structured and organized even before the onset of vision. Spontaneous neural activity, which spreads in waves across the retina, has been suggested to play a major role in these prenatal structuring processes. Recently, it has been shown that when employing an efficient coding strategy, such as sparse coding, these retinal activity patterns lead to basis functions that resemble optimal stimuli of simple cells in primary visual cortex (V1). Here we present the results of applying a coding strategy that optimizes for temporal slowness, namely Slow Feature Analysis (SFA), to a biologically plausible model of retinal waves. Previously, SFA has been successfully applied to model parts of the visual system, most notably in reproducing a rich set of complex-cell features by training SFA with quasi-natural image sequences. In the present work, we obtain SFA units that share a number of properties with cortical complex-cells by training on simulated retinal waves. The emergence of two distinct properties of the SFA units (phase invariance and orientation tuning) is thoroughly investigated via control experiments and mathematical analysis of the input-output functions found by SFA. The results support the idea that retinal waves share relevant temporal and spatial properties with natural visual input. Hence, retinal waves seem suitable training stimuli to learn invariances and thereby shape the developing early visual system such that it is best prepared for coding input from the natural world.