Spatial learning and action planning in a prefrontal cortical network model

The interplay between hippocampus and prefrontal cortex (PFC) is fundamental to spatial cognition. Complementing hippocampal place coding, prefrontal representations provide more abstract and hierarchically organized memories suitable for decision making. We model a prefrontal network mediating distributed information processing for spatial learning and action planning. Specific connectivity and synaptic adaptation principles shape the recurrent dynamics of the network arranged in cortical minicolumns. We show how the PFC columnar organization is suitable for learning sparse topological-metrical representations from redundant hippocampal inputs. The recurrent nature of the network supports multilevel spatial processing, allowing structural features of the environment to be encoded. An activation diffusion mechanism spreads the neural activity through the column population leading to trajectory planning. The model provides a functional framework for interpreting the activity of PFC neurons recorded during navigation tasks. We illustrate the link from single unit activity to behavioral responses. The results suggest plausible neural mechanisms subserving the cognitive "insight" capability originally attributed to rodents by Tolman & Honzik. Our time course analysis of neural responses shows how the interaction between hippocampus and PFC can yield the encoding of manifold information pertinent to spatial planning, including prospective coding and distance-to-goal correlates.     

Top-down control-signal dynamics in anterior cingulate and prefrontal cortex neurons following task switching

The prefrontal cortex (PFC) and anterior cingulate cortex (ACC) have both been implicated in cognitive control, but their relative roles remain unclear. Here we recorded the activity of single neurons in both areas while monkeys performed a task that required them to switch between trials in which they had to look toward a flashed stimulus (prosaccades) and trials in which they had to look away from the stimulus (antisaccades). We found that ACC neurons had a higher level of task selectivity than PFC neurons during the preparatory period on trials immediately following a task switch. In ACC neurons, task selectivity was strongest after the task switch and declined throughout the task block, whereas task selectivity remained constant in the PFC. These results demonstrate that the ACC is recruited when cognitive demands increase and suggest a role for both areas in task maintenance and the implementation of top-down control.     

Modelling memory functions with recurrent neural networks consisting of input compensation units: I. Static situations

Humans are able to form internal representations of the information they process—a capability which enables them to perform many different memory tasks. Therefore, the neural system has to learn somehow to represent aspects of the environmental situation; this process is assumed to be based on synaptic changes. The situations to be represented are various as for example different types of static patterns but also dynamic scenes. How are neural networks consisting of mutually connected neurons capable of performing such tasks? Here we propose a new neuronal structure for artificial neurons. This structure allows one to disentangle the dynamics of the recurrent connectivity from the dynamics induced by synaptic changes due to the learning processes. The error signal is computed locally within the individual neuron. Thus, online learning is possible without any additional structures. Recurrent neural networks equipped with these computational units cope with different memory tasks. Examples illustrate how information is extracted from environmental situations comprising fixed patterns to produce sustained activity and to deal with simple algebraic relations.     

Functional consequences of correlated excitatory and inhibitory conductances in cortical networks

Neurons in the neocortex receive a large number of excitatory and inhibitory synaptic inputs. Excitation and inhibition dynamically balance each other, with inhibition lagging excitation by only few milliseconds. To characterize the functional consequences of such correlated excitation and inhibition, we studied models in which this correlation structure is induced by feedforward inhibition (FFI). Simple circuits show that an effective FFI changes the integrative behavior of neurons such that only synchronous inputs can elicit spikes, causing the responses to be sparse and precise. Further, effective FFI increases the selectivity for propagation of synchrony through a feedforward network, thereby increasing the stability to background activity. Last, we show that recurrent random networks with effective inhibition are more likely to exhibit dynamical network activity states as have been observed in vivo. Thus, when a feedforward signal path is embedded in such recurrent network, the stabilizing effect of effective inhibition creates an suitable substrate for signal propagation. In conclusion, correlated excitation and inhibition support the notion that synchronous spiking may be important for cortical processing.     

Alternative time representation in dopamine models

Dopaminergic neuron activity has been modeled during learning and appetitive behavior, most commonly using the temporal-difference (TD) algorithm. However, a proper representation of elapsed time and of the exact task is usually required for the model to work. Most models use timing elements such as delay-line representations of time that are not biologically realistic for intervals in the range of seconds. The interval-timing literature provides several alternatives. One of them is that timing could emerge from general network dynamics, instead of coming from a dedicated circuit. Here, we present a general rate-based learning model based on long short-term memory (LSTM) networks that learns a time representation when needed. Using a naïve network learning its environment in conjunction with TD, we reproduce dopamine activity in appetitive trace conditioning with a constant CS-US interval, including probe trials with unexpected delays. The proposed model learns a representation of the environment dynamics in an adaptive biologically plausible framework, without recourse to delay lines or other special-purpose circuits. Instead, the model predicts that the task-dependent representation of time is learned by experience, is encoded in ramp-like changes in single-neuron activity distributed across small neural networks, and reflects a temporal integration mechanism resulting from the inherent dynamics of recurrent loops within the network. The model also reproduces the known finding that trace conditioning is more difficult than delay conditioning and that the learned representation of the task can be highly dependent on the types of trials experienced during training. Finally, it suggests that the phasic dopaminergic signal could facilitate learning in the cortex.     

Experience-dependent adult cortical plasticity requires cognitive association between sensation and reward

We tested the involvement of cognition in adult experience-dependent neuroplasticity using primate cortical implants. In a prior study, learning an operant sensory discrimination increased cortical excitability and target selectivity. Here, the prior task was separated into three behavioral phases. First, naive animals were exposed to stimulus-reward pairings from the prior study. These yoked animals did not have to discriminate to be rewarded and did not learn the discrimination. The plasticity observed in the prior study did not occur. Second, the animals were classically conditioned to discriminate the same stimuli in a simplified format. Learning was accompanied by increased sensory response strength and an increased range of sensory inputs eliciting responses. The third study recreated the original operant discrimination, and selectivity for task targets increased. These studies demonstrate that cognitive association between sensory stimuli and reinforcers accompanies adult experience-dependent cortical plasticity and suggest that selectivity in representation and action are linked.     

A network of spiking neurons that can represent interval timing: mean field analysis

Despite the vital importance of our ability to accurately process and encode temporal information, the underlying neural mechanisms are largely unknown. We have previously described a theoretical framework that explains how temporal representations, similar to those reported in the visual cortex, can form in locally recurrent cortical networks as a function of reward modulated synaptic plasticity. This framework allows networks of both linear and spiking neurons to learn the temporal interval between a stimulus and paired reward signal presented during training. Here we use a mean field approach to analyze the dynamics of non-linear stochastic spiking neurons in a network trained to encode specific time intervals. This analysis explains how recurrent excitatory feedback allows a network structure to encode temporal representations.     

Narrative event segmentation in the cortical reservoir

Recent research has revealed that during continuous perception of movies or stories, humans display cortical activity patterns that reveal hierarchical segmentation of event structure. Thus, sensory areas like auditory cortex display high frequency segmentation related to the stimulus, while semantic areas like posterior middle cortex display a lower frequency segmentation related to transitions between events. These hierarchical levels of segmentation are associated with different time constants for processing. Likewise, when two groups of participants heard the same sentence in a narrative, preceded by different contexts, neural responses for the groups were initially different and then gradually aligned. The time constant for alignment followed the segmentation hierarchy: sensory cortices aligned most quickly, followed by mid-level regions, while some higher-order cortical regions took more than 10 seconds to align. These hierarchical segmentation phenomena can be considered in the context of processing related to comprehension. In a recently described model of discourse comprehension word meanings are modeled by a language model pre-trained on a billion word corpus. During discourse comprehension, word meanings are continuously integrated in a recurrent cortical network. The model demonstrates novel discourse and inference processing, in part because of two fundamental characteristics: real-world event semantics are represented in the word embeddings, and these are integrated in a reservoir network which has an inherent gradient of functional time constants due to the recurrent connections. Here we demonstrate how this model displays hierarchical narrative event segmentation properties beyond the embeddings alone, or their linear integration. The reservoir produces activation patterns that are segmented by a hidden Markov model (HMM) in a manner that is comparable to that of humans. Context construction displays a continuum of time constants across reservoir neuron subsets, while context forgetting has a fixed time constant across these subsets. Importantly, virtual areas formed by subgroups of reservoir neurons with faster time constants segmented with shorter events, while those with longer time constants preferred longer events. This neurocomputational recurrent neural network simulates narrative event processing as revealed by the fMRI event segmentation algorithm provides a novel explanation of the asymmetry in narrative forgetting and construction. The model extends the characterization of online integration processes in discourse to more extended narrative, and demonstrates how reservoir computing provides a useful model of cortical processing of narrative structure.     

Slow manifolds within network dynamics encode working memory efficiently and robustly

Working memory is a cognitive function involving the storage and manipulation of latent information over brief intervals of time, thus making it crucial for context-dependent computation. Here, we use a top-down modeling approach to examine network-level mechanisms of working memory, an enigmatic issue and central topic of study in neuroscience. We optimize thousands of recurrent rate-based neural networks on a working memory task and then perform dynamical systems analysis on the ensuing optimized networks, wherein we find that four distinct dynamical mechanisms can emerge. In particular, we show the prevalence of a mechanism in which memories are encoded along slow stable manifolds in the network state space, leading to a phasic neuronal activation profile during memory periods. In contrast to mechanisms in which memories are directly encoded at stable attractors, these networks naturally forget stimuli over time. Despite this seeming functional disadvantage, they are more efficient in terms of how they leverage their attractor landscape and paradoxically, are considerably more robust to noise. Our results provide new hypotheses regarding how working memory function may be encoded within the dynamics of neural circuits.     

Emergence of symmetric, modular, and reciprocal connections in recurrent networks with Hebbian learning

While learning and development are well characterized in feedforward networks, these features are more difficult to analyze in recurrent networks due to the increased complexity of dual dynamics – the rapid dynamics arising from activation states and the slow dynamics arising from learning or developmental plasticity. We present analytical and numerical results that consider dual dynamics in a recurrent network undergoing Hebbian learning with either constant weight decay or weight normalization. Starting from initially random connections, the recurrent network develops symmetric or near-symmetric connections through Hebbian learning. Reciprocity and modularity arise naturally through correlations in the activation states. Additionally, weight normalization may be better than constant weight decay for the development of multiple attractor states that allow a diverse representation of the inputs. These results suggest a natural mechanism by which synaptic plasticity in recurrent networks such as cortical and brainstem premotor circuits could enhance neural computation and the generation of motor programs. Received: 27 April 1998 / Accepted in revised form: 16 March 1999     

Dynamics of orientation selectivity in the primary visual cortex and the importance of cortical inhibition

To test theories of orientation selectivity in primary visual cortex (V1), we have done experiments to measure the dynamics of orientation tuning of single neurons in the V1 cortex of macaque monkeys. Based on our dynamics results, we propose that a V1 cell's orientation selectivity is generated mainly by both tuned enhancement and global suppression. Enhancement near the preferred orientation is probably caused by feed-forward input from LGN (plus amplification by cortical-cortical interaction). Global suppression could be supplied by cortical inhibition. Additionally, in about 1/3 of V1 neurons (usually the most sharply tuned) there is tuned suppression, centered near the cell's preferred orientation but broader than tuned enhancement. These mechanisms also can explain important features of steady-state selectivity in the V1 neuron population. Furthermore, similar neuronal mechanisms may be used generally throughout the cerebral cortex.     

Inhibitory synchrony as a mechanism for attentional gain modulation

Recordings from area V4 of monkeys have revealed that when the focus of attention is on a visual stimulus within the receptive field of a cortical neuron, two distinct changes can occur: The firing rate of the neuron can change and there can be an increase in the coherence between spikes and the local field potential (LFP) in the gamma-frequency range (30-50 Hz). The hypothesis explored here is that these observed effects of attention could be a consequence of changes in the synchrony of local interneuron networks. We performed computer simulations of a Hodgkin-Huxley type neuron driven by a constant depolarizing current, I, representing visual stimulation and a modulatory inhibitory input representing the effects of attention via local interneuron networks. We observed that the neuron's firing rate and the coherence of its output spike train with the synaptic inputs was modulated by the degree of synchrony of the inhibitory inputs. When inhibitory synchrony increased, the coherence of spiking model neurons with the synaptic input increased, but the firing rate either increased or remained the same. The mean number of synchronous inhibitory inputs was a key determinant of the shape of the firing rate versus current (f-I) curves. For a large number of inhibitory inputs (approximately 50), the f-I curve saturated for large I and an increase in input synchrony resulted in a shift of sensitivity-the model neuron responded to weaker inputs I. For a small number (approximately 10), the f-I curves were non-saturating and an increase in input synchrony led to an increase in the gain of the response-the firing rate in response to the same input was multiplied by an approximately constant factor. The firing rate modulation with inhibitory synchrony was highest when the input network oscillated in the gamma frequency range. Thus, the observed changes in firing rate and coherence of neurons in the visual cortex could be controlled by top-down inputs that regulated the coherence in the activity of a local inhibitory network discharging at gamma frequencies.     

A learning theory for reward-modulated spike-timing-dependent plasticity with application to biofeedback

Reward-modulated spike-timing-dependent plasticity (STDP) has recently emerged as a candidate for a learning rule that could explain how behaviorally relevant adaptive changes in complex networks of spiking neurons could be achieved in a self-organizing manner through local synaptic plasticity. However, the capabilities and limitations of this learning rule could so far only be tested through computer simulations. This article provides tools for an analytic treatment of reward-modulated STDP, which allows us to predict under which conditions reward-modulated STDP will achieve a desired learning effect. These analytical results imply that neurons can learn through reward-modulated STDP to classify not only spatial but also temporal firing patterns of presynaptic neurons. They also can learn to respond to specific presynaptic firing patterns with particular spike patterns. Finally, the resulting learning theory predicts that even difficult credit-assignment problems, where it is very hard to tell which synaptic weights should be modified in order to increase the global reward for the system, can be solved in a self-organizing manner through reward-modulated STDP. This yields an explanation for a fundamental experimental result on biofeedback in monkeys by Fetz and Baker. In this experiment monkeys were rewarded for increasing the firing rate of a particular neuron in the cortex and were able to solve this extremely difficult credit assignment problem. Our model for this experiment relies on a combination of reward-modulated STDP with variable spontaneous firing activity. Hence it also provides a possible functional explanation for trial-to-trial variability, which is characteristic for cortical networks of neurons but has no analogue in currently existing artificial computing systems. In addition our model demonstrates that reward-modulated STDP can be applied to all synapses in a large recurrent neural network without endangering the stability of the network dynamics.     

Learning to Attend: Modeling the Shaping of Selectivity in Infero-temporal Cortex in a Categorization Task

Recent experiments on behaving monkeys have shown that learning a visual categorization task makes the neurons in infero-temporal cortex (ITC) more selective to the task-relevant features of the stimuli (Sigala and Logothetis in Nature 415 318–320, 2002). We hypothesize that such a selectivity modulation emerges from the interaction between ITC and other cortical area, presumably the prefrontal cortex (PFC), where the previously learned stimulus categories are encoded. We propose a biologically inspired model of excitatory and inhibitory spiking neurons with plastic synapses, modified according to a reward based Hebbian learning rule, to explain the experimental results and test the validity of our hypothesis. We assume that the ITC neurons, receiving feature selective inputs, form stronger connections with the category specific neurons to which they are consistently associated in rewarded trials. After learning, the top-down influence of PFC neurons enhances the selectivity of the ITC neurons encoding the behaviorally relevant features of the stimuli, as observed in the experiments. We conclude that the perceptual representation in visual areas like ITC can be strongly affected by the interaction with other areas which are devoted to higher cognitive functions. Electronic Supplementary Material: Supplementary material is available in the online: version of this article at http://dx.doi.org/10.007/s00422-006-0054-z     

Dynamical features of higher-order correlation events: impact on cortical cells

Cortical neurons receive signals from thousands of other neurons. The statistical properties of the input spike trains substantially shape the output response properties of each neuron. Experimental and theoretical investigations have mostly focused on the second order statistical features of the input spike trains (mean firing rates and pairwise correlations). Little is known of how higher order correlations affect the integration and firing behavior of a cell independently of the second order statistics. To address this issue, we simulated the dynamics of a population of 5000 neurons, controlling both their second order and higher-order correlation properties to reflect physiological data. We then used these ensemble dynamics as the input stage to morphologically reconstructed cortical cells (layer 5 pyramidal, layer 4 spiny stellate cell), and to an integrate and fire neuron. Our results show that changes done solely to the higher-order correlation properties of the network’s dynamics significantly affect the response properties of a target neuron, both in terms of output rate and spike timing. Moreover, the neuronal morphology and voltage dependent mechanisms of the target neuron considerably modulate the quantitative aspects of these effects. Finally, we show how these results affect sparseness of neuronal representations, tuning properties, and feature selectivity of cortical cells. An erratum to this article can be found at     

Recurrent Network Models for Perfect Temporal Integration of Fluctuating Correlated Inputs

Temporal integration of input is essential to the accumulation of information in various cognitive and behavioral processes, and gradually increasing neuronal activity, typically occurring within a range of seconds, is considered to reflect such computation by the brain. Some psychological evidence suggests that temporal integration by the brain is nearly perfect, that is, the integration is non-leaky, and the output of a neural integrator is accurately proportional to the strength of input. Neural mechanisms of perfect temporal integration, however, remain largely unknown. Here, we propose a recurrent network model of cortical neurons that perfectly integrates partially correlated, irregular input spike trains. We demonstrate that the rate of this temporal integration changes proportionately to the probability of spike coincidences in synaptic inputs. We analytically prove that this highly accurate integration of synaptic inputs emerges from integration of the variance of the fluctuating synaptic inputs, when their mean component is kept constant. Highly irregular neuronal firing and spike coincidences are the major features of cortical activity, but they have been separately addressed so far. Our results suggest that the efficient protocol of information integration by cortical networks essentially requires both features and hence is heterotic.     

Emergence of network structure due to spike-timing-dependent plasticity in recurrent neuronal networks. II. Input selectivity—symmetry breaking

Spike-timing-dependent plasticity (STDP) is believed to structure neuronal networks by slowly changing the strengths (or weights) of the synaptic connections between neurons depending upon their spiking activity, which in turn modifies the neuronal firing dynamics. In this paper, we investigate the change in synaptic weights induced by STDP in a recurrently connected network in which the input weights are plastic but the recurrent weights are fixed. The inputs are divided into two pools with identical constant firing rates and equal within-pool spike-time correlations, but with no between-pool correlations. Our analysis uses the Poisson neuron model in order to predict the evolution of the input synaptic weights and focuses on the asymptotic weight distribution that emerges due to STDP. The learning dynamics induces a symmetry breaking for the individual neurons, namely for sufficiently strong within-pool spike-time correlation each neuron specializes to one of the input pools. We show that the presence of fixed excitatory recurrent connections between neurons induces a group symmetry-breaking effect, in which neurons tend to specialize to the same input pool. Consequently STDP generates a functional structure on the input connections of the network.     

The response of cortical neurons to in vivo-like input current: theory and experiment

The study of several aspects of the collective dynamics of interacting neurons can be highly simplified if one assumes that the statistics of the synaptic input is the same for a large population of similarly behaving neurons (mean field approach). In particular, under such an assumption, it is possible to determine and study all the equilibrium points of the network dynamics when the neuronal response to noisy, in vivo-like, synaptic currents is known. The response function can be computed analytically for simple integrate-and-fire neuron models and it can be measured directly in experiments in vitro. Here we review theoretical and experimental results about the neural response to noisy inputs with stationary statistics. These response functions are important to characterize the collective neural dynamics that are proposed to be the neural substrate of working memory, decision making and other cognitive functions. Applications to the case of time-varying inputs are reviewed in a companion paper (Giugliano et al. in Biol Cybern, 2008). We conclude that modified integrate-and-fire neuron models are good enough to reproduce faithfully many of the relevant dynamical aspects of the neuronal response measured in experiments on real neurons in vitro.     

Neuro-robotics study on integrative learning of proactive visual attention and motor behaviors

The current paper proposes a novel model for integrative learning of proactive visual attention and sensory-motor control as inspired by the premotor theory of visual attention. The model is characterized by coupling a slow dynamics network with a fast dynamics network and by inheriting our prior proposed multiple timescales recurrent neural networks model (MTRNN) that may correspond to the fronto-parietal networks in the cortical brains. The neuro-robotics experiments in a task of manipulating multiple objects utilizing the proposed model demonstrated that some degrees of generalization in terms of position and object size variation can be achieved by organizing seamless integration of the proactive object-related visual attention and the related sensory-motor control into a set of action primitives in the distributed neural activities appearing in the fast dynamics network. It was also shown that such action primitives can be combined in compositional ways in acquiring novel actions in the slow dynamics network. The experimental results presented substantiate the premotor theory of visual attention.