Changes of mind in an attractor network of decision-making

Attractor networks successfully account for psychophysical and neurophysiological data in various decision-making tasks. Especially their ability to model persistent activity, a property of many neurons involved in decision-making, distinguishes them from other approaches. Stable decision attractors are, however, counterintuitive to changes of mind. Here we demonstrate that a biophysically-realistic attractor network with spiking neurons, in its itinerant transients towards the choice attractors, can replicate changes of mind observed recently during a two-alternative random-dot motion (RDM) task. Based on the assumption that the brain continues to evaluate available evidence after the initiation of a decision, the network predicts neural activity during changes of mind and accurately simulates reaction times, performance and percentage of changes dependent on difficulty. Moreover, the model suggests a low decision threshold and high incoming activity that drives the brain region involved in the decision-making process into a dynamical regime close to a bifurcation, which up to now lacked evidence for physiological relevance. Thereby, we further affirmed the general conformance of attractor networks with higher level neural processes and offer experimental predictions to distinguish nonlinear attractor from linear diffusion models.     

A fluctuation-driven mechanism for slow decision processes in reverberant networks

The spike activity of cells in some cortical areas has been found to be correlated with reaction times and behavioral responses during two-choice decision tasks. These experimental findings have motivated the study of biologically plausible winner-take-all network models, in which strong recurrent excitation and feedback inhibition allow the network to form a categorical choice upon stimulation. Choice formation corresponds in these models to the transition from the spontaneous state of the network to a state where neurons selective for one of the choices fire at a high rate and inhibit the activity of the other neurons. This transition has been traditionally induced by an increase in the external input that destabilizes the spontaneous state of the network and forces its relaxation to a decision state. Here we explore a different mechanism by which the system can undergo such transitions while keeping the spontaneous state stable, based on an escape induced by finite-size noise from the spontaneous state. This decision mechanism naturally arises for low stimulus strengths and leads to exponentially distributed decision times when the amount of noise in the system is small. Furthermore, we show using numerical simulations that mean decision times follow in this regime an exponential dependence on the amplitude of noise. The escape mechanism provides thus a dynamical basis for the wide range and variability of decision times observed experimentally.     

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.     

A Multi-Area Stochastic Model for a Covert Visual Search Task

Decisions typically comprise several elements. For example, attention must be directed towards specific objects, their identities recognized, and a choice made among alternatives. Pairs of competing accumulators and drift-diffusion processes provide good models of evidence integration in two-alternative perceptual choices, but more complex tasks requiring the coordination of attention and decision making involve multistage processing and multiple brain areas. Here we consider a task in which a target is located among distractors and its identity reported by lever release. The data comprise reaction times, accuracies, and single unit recordings from two monkeys’ lateral interparietal area (LIP) neurons. LIP firing rates distinguish between targets and distractors, exhibit stimulus set size effects, and show response-hemifield congruence effects. These data motivate our model, which uses coupled sets of leaky competing accumulators to represent processes hypothesized to occur in feature-selective areas and limb motor and pre-motor areas, together with the visual selection process occurring in LIP. Model simulations capture the electrophysiological and behavioral data, and fitted parameters suggest that different connection weights between LIP and the other cortical areas may account for the observed behavioral differences between the animals.     

Integration of Sensory and Reward Information during Perceptual Decision-Making in Lateral Intraparietal Cortex (LIP) of the Macaque Monkey

Single neurons in cortical area LIP are known to carry information relevant to both sensory and value-based decisions that are reported by eye movements. It is not known, however, how sensory and value information are combined in LIP when individual decisions must be based on a combination of these variables. To investigate this issue, we conducted behavioral and electrophysiological experiments in rhesus monkeys during performance of a two-alternative, forced-choice discrimination of motion direction (sensory component). Monkeys reported each decision by making an eye movement to one of two visual targets associated with the two possible directions of motion. We introduced choice biases to the monkeys'' decision process (value component) by randomly interleaving balanced reward conditions (equal reward value for the two choices) with unbalanced conditions (one alternative worth twice as much as the other). The monkeys'' behavior, as well as that of most LIP neurons, reflected the influence of all relevant variables: the strength of the sensory information, the value of the target in the neuron''s response field, and the value of the target outside the response field. Overall, detailed analysis and computer simulation reveal that our data are consistent with a two-stage drift diffusion model proposed by Diederich and Bussmeyer [1] for the effect of payoffs in the context of sensory discrimination tasks. Initial processing of payoff information strongly influences the starting point for the accumulation of sensory evidence, while exerting little if any effect on the rate of accumulation of sensory evidence.     

How to Find Decision Makers in Neural Networks

Nervous systems often face the problem of classifying stimuli and making decisions based on these classifications. The neurons involved in these tasks can be characterized as sensory or motor, according to their correlation with sensory stimulus or motor response. In this study we define a third class of neurons responsible for making perceptual decisions. Our mathematical formalism enables the weighting of neuronal units according to their contribution to decision making, thus narrowing the field for more detailed studies of underlying mechanisms. We develop two definitions of a contribution to decision making. The first definition states that decision making activity can be found at the points of emergence for behavioral correlations in the system. The second definition involves the study of propagation of noise in the network. The latter definition is shown to be equivalent to the first one in the cases when they can be compared. Our results suggest a new approach to analyzing decision making networks An erratum to this article can be found at     

One-dimensional dynamics of attention and decision making in LIP

Where we allocate our visual spatial attention depends upon a continual competition between internally generated goals and external distractions. Recently it was shown that single neurons in the macaque lateral intraparietal area (LIP) can predict the amount of time a distractor can shift the locus of spatial attention away from a goal. We propose that this remarkable dynamical correspondence between single neurons and attention can be explained by a network model in which generically high-dimensional firing-rate vectors rapidly decay to a single mode. We find direct experimental evidence for this model, not only in the original attentional task, but also in a very different task involving perceptual decision making. These results confirm a theoretical prediction that slowly varying activity patterns are proportional to spontaneous activity, pose constraints on models of persistent activity, and suggest a network mechanism for the emergence of robust behavioral timing from heterogeneous neuronal populations.     

Modeling convergent ON and OFF pathways in the early visual system

For understanding the computation and function of single neurons in sensory systems, one needs to investigate how sensory stimuli are related to a neuron’s response and which biological mechanisms underlie this relationship. Mathematical models of the stimulus–response relationship have proved very useful in approaching these issues in a systematic, quantitative way. A starting point for many such analyses has been provided by phenomenological “linear–nonlinear” (LN) models, which comprise a linear filter followed by a static nonlinear transformation. The linear filter is often associated with the neuron’s receptive field. However, the structure of the receptive field is generally a result of inputs from many presynaptic neurons, which may form parallel signal processing pathways. In the retina, for example, certain ganglion cells receive excitatory inputs from ON-type as well as OFF-type bipolar cells. Recent experiments have shown that the convergence of these pathways leads to intriguing response characteristics that cannot be captured by a single linear filter. One approach to adjust the LN model to the biological circuit structure is to use multiple parallel filters that capture ON and OFF bipolar inputs. Here, we review these new developments in modeling neuronal responses in the early visual system and provide details about one particular technique for obtaining the required sets of parallel filters from experimental data.     

Oscillations and Synchrony in Large-scale Cortical Network Models

Intrinsic neuronal and circuit properties control the responses of large ensembles of neurons by creating spatiotemporal patterns of activity that are used for sensory processing, memory formation, and other cognitive tasks. The modeling of such systems requires computationally efficient single-neuron models capable of displaying realistic response properties. We developed a set of reduced models based on difference equations (map-based models) to simulate the intrinsic dynamics of biological neurons. These phenomenological models were designed to capture the main response properties of specific types of neurons while ensuring realistic model behavior across a sufficient dynamic range of inputs. This approach allows for fast simulations and efficient parameter space analysis of networks containing hundreds of thousands of neurons of different types using a conventional workstation. Drawing on results obtained using large-scale networks of map-based neurons, we discuss spatiotemporal cortical network dynamics as a function of parameters that affect synaptic interactions and intrinsic states of the neurons.     

Adaptive Sampling of Information in Perceptual Decision-Making

In many perceptual and cognitive decision-making problems, humans sample multiple noisy information sources serially, and integrate the sampled information to make an overall decision. We derive the optimal decision procedure for two-alternative choice tasks in which the different options are sampled one at a time, sources vary in the quality of the information they provide, and the available time is fixed. To maximize accuracy, the optimal observer allocates time to sampling different information sources in proportion to their noise levels. We tested human observers in a corresponding perceptual decision-making task. Observers compared the direction of two random dot motion patterns that were triggered only when fixated. Observers allocated more time to the noisier pattern, in a manner that correlated with their sensory uncertainty about the direction of the patterns. There were several differences between the optimal observer predictions and human behaviour. These differences point to a number of other factors, beyond the quality of the currently available sources of information, that influences the sampling strategy.     

Theory of Choice in Bandit,Information Sampling and Foraging Tasks

Decision making has been studied with a wide array of tasks. Here we examine the theoretical structure of bandit, information sampling and foraging tasks. These tasks move beyond tasks where the choice in the current trial does not affect future expected rewards. We have modeled these tasks using Markov decision processes (MDPs). MDPs provide a general framework for modeling tasks in which decisions affect the information on which future choices will be made. Under the assumption that agents are maximizing expected rewards, MDPs provide normative solutions. We find that all three classes of tasks pose choices among actions which trade-off immediate and future expected rewards. The tasks drive these trade-offs in unique ways, however. For bandit and information sampling tasks, increasing uncertainty or the time horizon shifts value to actions that pay-off in the future. Correspondingly, decreasing uncertainty increases the relative value of actions that pay-off immediately. For foraging tasks the time-horizon plays the dominant role, as choices do not affect future uncertainty in these tasks.     

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.     

Spatio-temporal credit assignment in neuronal population learning

In learning from trial and error, animals need to relate behavioral decisions to environmental reinforcement even though it may be difficult to assign credit to a particular decision when outcomes are uncertain or subject to delays. When considering the biophysical basis of learning, the credit-assignment problem is compounded because the behavioral decisions themselves result from the spatio-temporal aggregation of many synaptic releases. We present a model of plasticity induction for reinforcement learning in a population of leaky integrate and fire neurons which is based on a cascade of synaptic memory traces. Each synaptic cascade correlates presynaptic input first with postsynaptic events, next with the behavioral decisions and finally with external reinforcement. For operant conditioning, learning succeeds even when reinforcement is delivered with a delay so large that temporal contiguity between decision and pertinent reward is lost due to intervening decisions which are themselves subject to delayed reinforcement. This shows that the model provides a viable mechanism for temporal credit assignment. Further, learning speeds up with increasing population size, so the plasticity cascade simultaneously addresses the spatial problem of assigning credit to synapses in different population neurons. Simulations on other tasks, such as sequential decision making, serve to contrast the performance of the proposed scheme to that of temporal difference-based learning. We argue that, due to their comparative robustness, synaptic plasticity cascades are attractive basic models of reinforcement learning in the brain.     

Temporal integration of olfactory perceptual evidence in human orbitofrontal cortex

Given a noisy sensory world, the nervous system integrates perceptual evidence over time to optimize decision-making. Neurophysiological accumulation of sensory information is well-documented in the animal visual system, but how such mechanisms are instantiated in the human brain remains poorly understood. Here we combined psychophysical techniques, drift-diffusion modeling, and functional magnetic resonance imaging (fMRI) to establish that odor evidence integration in the human olfactory system enhances discrimination on a two-alternative forced-choice task. Model-based measures of fMRI brain activity highlighted a ramp-like increase in orbitofrontal cortex (OFC) that peaked at the time of decision, conforming to predictions derived from an integrator model. Combined behavioral and fMRI data further suggest that decision bounds are not fixed but collapse over time, facilitating choice behavior in the presence of low-quality evidence. These data highlight a key role for the orbitofrontal cortex in resolving sensory uncertainty and provide substantiation for accumulator models of human perceptual decision-making.     

The Correlation Structure of Local Neuronal Networks Intrinsically Results from Recurrent Dynamics

Correlated neuronal activity is a natural consequence of network connectivity and shared inputs to pairs of neurons, but the task-dependent modulation of correlations in relation to behavior also hints at a functional role. Correlations influence the gain of postsynaptic neurons, the amount of information encoded in the population activity and decoded by readout neurons, and synaptic plasticity. Further, it affects the power and spatial reach of extracellular signals like the local-field potential. A theory of correlated neuronal activity accounting for recurrent connectivity as well as fluctuating external sources is currently lacking. In particular, it is unclear how the recently found mechanism of active decorrelation by negative feedback on the population level affects the network response to externally applied correlated stimuli. Here, we present such an extension of the theory of correlations in stochastic binary networks. We show that (1) for homogeneous external input, the structure of correlations is mainly determined by the local recurrent connectivity, (2) homogeneous external inputs provide an additive, unspecific contribution to the correlations, (3) inhibitory feedback effectively decorrelates neuronal activity, even if neurons receive identical external inputs, and (4) identical synaptic input statistics to excitatory and to inhibitory cells increases intrinsically generated fluctuations and pairwise correlations. We further demonstrate how the accuracy of mean-field predictions can be improved by self-consistently including correlations. As a byproduct, we show that the cancellation of correlations between the summed inputs to pairs of neurons does not originate from the fast tracking of external input, but from the suppression of fluctuations on the population level by the local network. This suppression is a necessary constraint, but not sufficient to determine the structure of correlations; specifically, the structure observed at finite network size differs from the prediction based on perfect tracking, even though perfect tracking implies suppression of population fluctuations.     

Oscillations and oscillatory behavior in small neural circuits

In order to determine the dynamical properties of central pattern generators (CPGs), we have examined the lobster stomatogastric ganglion using the tools of nonlinear dynamics. The lobster pyloric and gastric mill central pattern generators can be analyzed at both the cellular and network levels because they are small, i.e., contain only 25 neurons between them and each neuron and synapse are repeatedly identifiable from animal to animal. We discuss how the biophysical properties of each neuron and synapse in the two circuits act cooperatively to generate two different patterns of sequential activity, how these patterns are altered by neuromodulators and perturbed by noise and sensory inputs. Finally, we show how simplified Hindmarsh–Rose models can be made into analog electronic neurons that mimic the lobster neurons and in addition be incorporated into artificial CPGs with robotic applications.     

Beyond Two-Cell Networks: Experimental Measurement of Neuronal Responses to Multiple Synaptic Inputs

Oscillations of large populations of neurons are thought to be important in the normal functioning of the brain. We have used phase response curve (PRC) methods to characterize the dynamics of single neurons and predict population dynamics. Our past experimental work was limited to special circumstances (e.g., 2-cell networks of periodically firing neurons). Here, we explore the feasibility of extending our methods to predict the synchronization properties of stellate cells (SCs) in the rat entorhinal cortex under broader conditions. In particular, we test the hypothesis that PRCs in SCs scale linearly with changes in synaptic amplitude, and measure how well responses to Poisson process-driven inputs can be predicted in terms of PRCs. Although we see nonlinear responses to excitatory and inhibitory inputs, we find that models based on weak coupling account for scaling and Poisson process-driven inputs reasonably accurately.     

Driving reservoir models with oscillations: a solution to the extreme structural sensitivity of chaotic networks

A large body of experimental and theoretical work on neural coding suggests that the information stored in brain circuits is represented by time-varying patterns of neural activity. Reservoir computing, where the activity of a recurrently connected pool of neurons is read by one or more units that provide an output response, successfully exploits this type of neural activity. However, the question of system robustness to small structural perturbations, such as failing neurons and synapses, has been largely overlooked. This contrasts with well-studied dynamical perturbations that lead to divergent network activity in the presence of chaos, as is the case for many reservoir networks. Here, we distinguish between two types of structural network perturbations, namely local (e.g., individual synaptic or neuronal failure) and global (e.g., network-wide fluctuations). Surprisingly, we show that while global perturbations have a limited impact on the ability of reservoir models to perform various tasks, local perturbations can produce drastic effects. To address this limitation, we introduce a new architecture where the reservoir is driven by a layer of oscillators that generate stable and repeatable trajectories. This model outperforms previous implementations while being resistant to relatively large local and global perturbations. This finding has implications for the design of reservoir models that capture the capacity of brain circuits to perform cognitively and behaviorally relevant tasks while remaining robust to various forms of perturbations. Further, our work proposes a novel role for neuronal oscillations found in cortical circuits, where they may serve as a collection of inputs from which a network can robustly generate complex dynamics and implement rich computations.     

Computation by ensemble synchronization in recurrent networks with synaptic depression

While computation by ensemble synchronization is considered to be a robust and efficient way for information processing in the cortex (C. Von der Malsburg and W. Schneider (1986) Biol. Cybern. 54: 29–40; W. Singer (1994) Inter. Rev. Neuro. 37: 153–183; J.J. Hopfield (1995) Nature 376: 33–36; E. Vaadia et al. (1995) Nature 373: 515–518), the neuronal mechanisms that might be used to achieve it are yet to be uncovered. Here we analyze a neural network model in which the computations are performed by near coincident firing of neurons in response to external inputs. This near coincident firing is enabled by activity dependent depression of inter-neuron connections. We analyze the network behavior by using a mean-field approximation, which allows predicting the network response to various inputs. We demonstrate that the network is very sensitive to temporal aspects of the inputs. In particular, periodically applied inputs of increasing frequency result in different response profiles. Moreover, applying combinations of different stimuli lead to a complex response, which cannot be easily predicted from responses to individual components. These results demonstrate that networks with synaptic depression can perform complex computations on time-dependent inputs utilizing the ability to generate temporally synchronous firing of single neurons.