Unsupervised learning for robust working memory.

Working memory is a core component of critical cognitive functions such as planning and decision-making. Persistent activity that lasts long after the stimulus offset has been considered a neural substrate for working memory. Attractor dynamics based on network interactions can successfully reproduce such persistent activity. However, it requires a fine-tuning of network connectivity, in particular, to form continuous attractors which were suggested for encoding continuous signals in working memory. Here, we investigate whether a specific form of synaptic plasticity rules can mitigate such tuning problems in two representative working memory models, namely, rate-coded and location-coded persistent activity. We consider two prominent types of plasticity rules, differential plasticity correcting the rapid activity changes and homeostatic plasticity regularizing the long-term average of activity, both of which have been proposed to fine-tune the weights in an unsupervised manner. Consistent with the findings of previous works, differential plasticity alone was enough to recover a graded-level persistent activity after perturbations in the connectivity. For the location-coded memory, differential plasticity could also recover persistent activity. However, its pattern can be irregular for different stimulus locations under slow learning speed or large perturbation in the connectivity. On the other hand, homeostatic plasticity shows a robust recovery of smooth spatial patterns under particular types of synaptic perturbations, such as perturbations in incoming synapses onto the entire or local populations. However, homeostatic plasticity was not effective against perturbations in outgoing synapses from local populations. Instead, combining it with differential plasticity recovers location-coded persistent activity for a broader range of perturbations, suggesting compensation between two plasticity rules.

Visual Familiarity Induced 5-Hz Oscillations and Improved Orientation and Direction Selectivities in V1.

Neural oscillations play critical roles in information processing, communication between brain areas, learning, and memory. We have recently discovered that familiar visual stimuli can robustly induce 5-Hz oscillations in the primary visual cortex (V1) of awake mice after the visual experience. To gain more mechanistic insight into this phenomenon, we used in vivo patch-clamp recordings to monitor the subthreshold activity of individual neurons during these oscillations. We analyzed the visual tuning properties of V1 neurons in naive and experienced mice to assess the effect of visual experience on the orientation and direction selectivity. Using optogenetic stimulation through the patch pipette in vivo, we measured the synaptic strength of specific intracortical and thalamocortical projections in vivo in the visual cortex before and after the visual experience. We found 5-Hz oscillations in membrane potential (Vm) and firing rates evoked in single neurons in response to the familiar stimulus, consistent with previous studies. Following the visual experience, the average firing rates of visual responses were reduced while the orientation and direction selectivities were increased. Light-evoked EPSCs were significantly increased for layer 5 (L5) projections to other layers of V1 after the visual experience, while the thalamocortical synaptic strength was decreased. In addition, we developed a computational model that could reproduce 5-Hz oscillations with enhanced neuronal selectivity following synaptic plasticity within the recurrent network and decreased feedforward input.SIGNIFICANCE STATEMENT Neural oscillations at around 5 Hz are involved in visual working memory and temporal expectations in primary visual cortex (V1). However, how the oscillations modulate the visual response properties of neurons in V1 and their underlying mechanism is poorly understood. Here, we show that these oscillations may alter the orientation and direction selectivity of the layer 2/3 (L2/3) neurons and correlate with the synaptic plasticity within V1. Our computational recurrent network model reproduces all these observations and provides a mechanistic framework for studying the role of 5-Hz oscillations in visual familiarity.

Balanced cortical microcircuitry for spatial working memory based on corrective feedback control.

A hallmark of working memory is the ability to maintain graded representations of both the spatial location and amplitude of a memorized stimulus. Previous work has identified a neural correlate of spatial working memory in the persistent maintenance of spatially specific patterns of neural activity. How such activity is maintained by neocortical circuits remains unknown. Traditional models of working memory maintain analog representations of either the spatial location or the amplitude of a stimulus, but not both. Furthermore, although most previous models require local excitation and lateral inhibition to maintain spatially localized persistent activity stably, the substrate for lateral inhibitory feedback pathways is unclear. Here, we suggest an alternative model for spatial working memory that is capable of maintaining analog representations of both the spatial location and amplitude of a stimulus, and that does not rely on long-range feedback inhibition. The model consists of a functionally columnar network of recurrently connected excitatory and inhibitory neural populations. When excitation and inhibition are balanced in strength but offset in time, drifts in activity trigger spatially specific negative feedback that corrects memory decay. The resulting networks can temporally integrate inputs at any spatial location, are robust against many commonly considered perturbations in network parameters, and, when implemented in a spiking model, generate irregular neural firing characteristic of that observed experimentally during persistent activity. This work suggests balanced excitatory-inhibitory memory circuits implementing corrective negative feedback as a substrate for spatial working memory.

Efficient inference of synaptic plasticity rule with Gaussian process regression.

Finding the form of synaptic plasticity is critical to understanding its functions underlying learning and memory. We investigated an efficient method to infer synaptic plasticity rules in various experimental settings. We considered biologically plausible models fitting a wide range of in-vitro studies and examined the recovery of their firing-rate dependence from sparse and noisy data. Among the methods assuming low-rankness or smoothness of plasticity rules, Gaussian process regression (GPR), a nonparametric Bayesian approach, performs the best. Under the conditions measuring changes in synaptic weights directly or measuring changes in neural activities as indirect observables of synaptic plasticity, which leads to different inference problems, GPR performs well. Also, GPR could simultaneously recover multiple plasticity rules and robustly perform under various plasticity rules and noise levels. Such flexibility and efficiency, particularly at the low sampling regime, make GPR suitable for recent experimental developments and inferring a broader class of plasticity models.

Noise tolerance of attractor and feedforward memory models.

In short-term memory networks, transient stimuli are represented by patterns of neural activity that persist long after stimulus offset. Here, we compare the performance of two prominent classes of memory networks, feedback-based attractor networks and feedforward networks, in conveying information about the amplitude of a briefly presented stimulus in the presence of gaussian noise. Using Fisher information as a metric of memory performance, we find that the optimal form of network architecture depends strongly on assumptions about the forms of nonlinearities in the network. For purely linear networks, we find that feedforward networks outperform attractor networks because noise is continually removed from feedforward networks when signals exit the network; as a result, feedforward networks can amplify signals they receive faster than noise accumulates over time. By contrast, attractor networks must operate in a signal-attenuating regime to avoid the buildup of noise. However, if the amplification of signals is limited by a finite dynamic range of neuronal responses or if noise is reset at the time of signal arrival, as suggested by recent experiments, we find that attractor networks can outperform feedforward ones. Under a simple model in which neurons have a finite dynamic range, we find that the optimal attractor networks are forgetful if there is no mechanism for noise reduction with signal arrival but nonforgetful (perfect integrators) in the presence of a strong reset mechanism. Furthermore, we find that the maximal Fisher information for the feedforward and attractor networks exhibits power law decay as a function of time and scales linearly with the number of neurons. These results highlight prominent factors that lead to trade-offs in the memory performance of networks with different architectures and constraints, and suggest conditions under which attractor or feedforward networks may be best suited to storing information about previous stimuli.

Noise-induced transitions in slow wave neuronal dynamics.

Many neuronal systems exhibit slow random alternations and sudden switches in activity states. Models with noisy relaxation dynamics (oscillatory, excitable or bistable) account for these temporal, slow wave, patterns and the fluctuations within states. The noise-induced transitions in a relaxation dynamics are analogous to escape by a particle in a slowly changing double-well potential. In this formalism, we obtain semi-analytically the first and second order statistical properties: the distributions of the slow process at the transitions and the temporal correlations of successive switching events. We find that the temporal correlations can be used to help distinguish among biophysical mechanisms for the slow negative feedback, such as divisive or subtractive. We develop our results in the context of models for cellular pacemaker neurons; they also apply to mean-field models for spontaneously active networks with slow wave dynamics.

Mechanisms underlying sharpening of visual response dynamics with familiarity.

Experience-dependent modifications of synaptic connections are thought to change patterns of network activities and stimulus tuning with learning. However, only a few studies explored how synaptic plasticity shapes the response dynamics of cortical circuits. Here, we investigated the mechanism underlying sharpening of both stimulus selectivity and response dynamics with familiarity observed in monkey inferotemporal cortex. Broadening the distribution of activities and stronger oscillations in the response dynamics after learning provide evidence for synaptic plasticity in recurrent connections modifying the strength of positive feedback. Its interplay with slow negative feedback via firing rate adaptation is critical in sharpening response dynamics. Analysis of changes in temporal patterns also enables us to disentangle recurrent and feedforward synaptic plasticity and provides a measure for the strengths of recurrent synaptic plasticity. Overall, this work highlights the importance of analyzing changes in dynamics as well as network patterns to further reveal the mechanisms of visual learning.

Population-scale organization of cerebellar granule neuron signaling during a visuomotor behavior.

Granule cells at the input layer of the cerebellum comprise over half the neurons in the human brain and are thought to be critical for learning. However, little is known about granule neuron signaling at the population scale during behavior. We used calcium imaging in awake zebrafish during optokinetic behavior to record transgenically identified granule neurons throughout a cerebellar population. A significant fraction of the population was responsive at any given time. In contrast to core precerebellar populations, granule neuron responses were relatively heterogeneous, with variation in the degree of rectification and the balance of positive versus negative changes in activity. Functional correlations were strongest for nearby cells, with weak spatial gradients in the degree of rectification and the average sign of response. These data open a new window upon cerebellar function and suggest granule layer signals represent elementary building blocks under-represented in core sensorimotor pathways, thereby enabling the construction of novel patterns of activity for learning.

Inferring learning rules from distributions of firing rates in cortical neurons.

Information about external stimuli is thought to be stored in cortical circuits through experience-dependent modifications of synaptic connectivity. These modifications of network connectivity should lead to changes in neuronal activity as a particular stimulus is repeatedly encountered. Here we ask what plasticity rules are consistent with the differences in the statistics of the visual response to novel and familiar stimuli in inferior temporal cortex, an area underlying visual object recognition. We introduce a method that allows one to infer the dependence of the presumptive learning rule on postsynaptic firing rate, and we show that the inferred learning rule exhibits depression for low postsynaptic rates and potentiation for high rates. The threshold separating depression from potentiation is strongly correlated with both mean and s.d. of the firing rate distribution. Finally, we show that network models implementing a rule extracted from data show stable learning dynamics and lead to sparser representations of stimuli.

Balanced cortical microcircuitry for maintaining information in working memory.

Persistent neural activity in the absence of a stimulus has been identified as a neural correlate of working memory, but how such activity is maintained by neocortical circuits remains unknown. We used a computational approach to show that the inhibitory and excitatory microcircuitry of neocortical memory-storing regions is sufficient to implement a corrective feedback mechanism that enables persistent activity to be maintained stably for prolonged durations. When recurrent excitatory and inhibitory inputs to memory neurons were balanced in strength and offset in time, drifts in activity triggered a corrective signal that counteracted memory decay. Circuits containing this mechanism temporally integrated their inputs, generated the irregular neural firing observed during persistent activity and were robust against common perturbations that severely disrupted previous models of short-term memory storage. These results reveal a mechanism for the accumulation and storage of memories in neocortical circuits based on principles of corrective negative feedback that are widely used in engineering applications.