anti-Hebbian Plasticity Research Articles

Cortical development in the structural model and free energy minimization.

A model of neocortical development invoking Friston's Free Energy Principle is applied within the Structural Model of Barbas etal. and the associated functional interpretation advanced by Tucker and Luu. Evolution of a neural field with Hebbian and anti-Hebbian plasticity, maximizing synchrony and minimizing axonal length by apoptotic selection, leads to paired connection systems with mirror symmetry, interacting via Markov blankets along their line of reflection. Applied to development along the radial lines of development in the Structural Model, a primary Markov blanket emerges between the centrifugal synaptic flux in layers 2,3 and 5,6, versus the centripetal flow in layer 4, and axonal orientations in layer 4 give rise to the differing shape and movement sensitivities characteristic of neurons of dorsal and ventral neocortex. Prediction error minimization along the primary blanket integrates limbic and subcortical networks with the neocortex. Synaptic flux bypassing the blanket triggers the arousal response to surprising stimuli, enabling subsequent adaptation. As development progresses ubiquitous mirror systems separated by Markov blankets and enclosed blankets-within-blankets arise throughout neocortex, creating the typical order and response characteristics of columnar and noncolumnar cortex.

Anti-Hebbian plasticity drives sequence learning in striatum

Spatio-temporal activity patterns have been observed in a variety of brain areas in spontaneous activity, prior to or during action, or in response to stimuli. Biological mechanisms endowing neurons with the ability to distinguish between different sequences remain largely unknown. Learning sequences of spikes raises multiple challenges, such as maintaining in memory spike history and discriminating partially overlapping sequences. Here, we show that anti-Hebbian spike-timing dependent plasticity (STDP), as observed at cortico-striatal synapses, can naturally lead to learning spike sequences. We design a spiking model of the striatal output neuron receiving spike patterns defined as sequential input from a fixed set of cortical neurons. We use a simple synaptic plasticity rule that combines anti-Hebbian STDP and non-associative potentiation for a subset of the presented patterns called rewarded patterns. We study the ability of striatal output neurons to discriminate rewarded from non-rewarded patterns by firing only after the presentation of a rewarded pattern. In particular, we show that two biological properties of striatal networks, spiking latency and collateral inhibition, contribute to an increase in accuracy, by allowing a better discrimination of partially overlapping sequences. These results suggest that anti-Hebbian STDP may serve as a biological substrate for learning sequences of spikes.

Theory of Coupled Neuronal-Synaptic Dynamics

In neural circuits, synaptic strengths influence neuronal activity by shaping network dynamics, and neuronal activity influences synaptic strengths through activity-dependent plasticity. Motivated by this fact, we study a recurrent-network model in which neuronal units and synaptic couplings are interacting dynamic variables, with couplings subject to Hebbian modification with decay around quenched random strengths. Rather than assigning a specific role to the plasticity, we use dynamical mean-field theory and other techniques to systematically characterize the neuronal-synaptic dynamics, revealing a rich phase diagram. Adding Hebbian plasticity slows activity in already chaotic networks and can induce chaos in otherwise quiescent networks. Anti-Hebbian plasticity quickens activity and produces an oscillatory component. Analysis of the Jacobian shows that Hebbian and anti-Hebbian plasticity push locally unstable modes toward the real and imaginary axes, respectively, explaining these behaviors. Both random-matrix and Lyapunov analysis show that strong Hebbian plasticity segregates network timescales into two bands, with a slow, synapse-dominated band driving the dynamics, suggesting a flipped view of the network as synapses connected by neurons. For increasing strength, Hebbian plasticity initially raises the complexity of the dynamics, measured by the maximum Lyapunov exponent and attractor dimension, but then decreases these metrics, likely due to the proliferation of stable fixed points. We compute the marginally stable spectra of such fixed points as well as their number, showing exponential growth with network size. Finally, in chaotic states with strong Hebbian plasticity, a stable fixed point of neuronal dynamics is destabilized by synaptic dynamics, allowing any neuronal state to be stored as a stable fixed point by halting the plasticity. This phase of freezable chaos offers a new mechanism for working memory. Published by the American Physical Society 2024

STDP-Driven Development of Attention-Based People Detection in Spiking Neural Networks

This correspondence paper provides, to the best of our knowledge, a first analysis of how biologically-plausible spiking neural networks (SNNs) equipped with Spike-Timing-Dependent Plasticity (STDP) can learn to detect people on the fly from non-independent and identically distributed (non-i.i.d) streams of retina-inspired, event camera data. Our system works as follows. First, a short sequence of event data capturing a walking human from a flying drone is forwarded in its natural order to an SNN-STDP system, which also receives teacher spiking signals from the neural activity readout block. Then, when the end of the learning sequence is reached, the learned system is assessed on testing sequences. In addition, we also present a new interpretation of anti-Hebbian plasticity as an over-fitting control mechanism, and provide experimental demonstrations of our findings. This work contributes to the study of attention-based development and perception in bio-inspired systems.

A role for cortical interneurons as adversarial discriminators.

The brain learns representations of sensory information from experience, but the algorithms by which it does so remain unknown. One popular theory formalizes representations as inferred factors in a generative model of sensory stimuli, meaning that learning must improve this generative model and inference procedure. This framework underlies many classic computational theories of sensory learning, such as Boltzmann machines, the Wake/Sleep algorithm, and a more recent proposal that the brain learns with an adversarial algorithm that compares waking and dreaming activity. However, in order for such theories to provide insights into the cellular mechanisms of sensory learning, they must be first linked to the cell types in the brain that mediate them. In this study, we examine whether a subtype of cortical interneurons might mediate sensory learning by serving as discriminators, a crucial component in an adversarial algorithm for representation learning. We describe how such interneurons would be characterized by a plasticity rule that switches from Hebbian plasticity during waking states to anti-Hebbian plasticity in dreaming states. Evaluating the computational advantages and disadvantages of this algorithm, we find that it excels at learning representations in networks with recurrent connections but scales poorly with network size. This limitation can be partially addressed if the network also oscillates between evoked activity and generative samples on faster timescales. Consequently, we propose that an adversarial algorithm with interneurons as discriminators is a plausible and testable strategy for sensory learning in biological systems.

Anti-Hebbian plasticity in the motor cortex promotes defensive freezing

Regional brain activity often decreases from baseline levels in response to external events, but how neurons develop such negative responses is unclear. To study this, we leveraged the negative response that develops in the primary motor cortex (M1) after classical fear learning. We trained mice with a fear conditioning paradigm while imaging their brains with standard two-photon microscopy. This enabled monitoring changes in neuronal responses to the tone with synaptic resolution through learning. We found that M1 layer 5 pyramidal neurons (L5 PNs) developed negative tone responses within an hour after conditioning, which depended on the weakening of their dendritic spines that were active during training. Blocking this form of anti-Hebbian plasticity using an optogenetic manipulation of CaMKII activity disrupted negative tone responses and freezing. Therefore, reducing the strength of spines active at the time of memory encoding leads to negative responses of L5 PNs. In turn, these negative responses curb M1's capacity for promoting movement, thereby aiding freezing. Collectively, this work provides a mechanistic understanding of how area-specific negative responses to behaviorally relevant cues can be achieved.

A neural theory for counting memories

Keeping track of the number of times different stimuli have been experienced is a critical computation for behavior. Here, we propose a theoretical two-layer neural circuit that stores counts of stimulus occurrence frequencies. This circuit implements a data structure, called a count sketch, that is commonly used in computer science to maintain item frequencies in streaming data. Our first model implements a count sketch using Hebbian synapses and outputs stimulus-specific frequencies. Our second model uses anti-Hebbian plasticity and only tracks frequencies within four count categories (“1-2-3-many”), which trades-off the number of categories that need to be distinguished with the potential ethological value of those categories. We show how both models can robustly track stimulus occurrence frequencies, thus expanding the traditional novelty-familiarity memory axis from binary to discrete with more than two possible values. Finally, we show that an implementation of the “1-2-3-many” count sketch exists in the insect mushroom body.

Meta-learning synaptic plasticity and memory addressing for continual familiarity detection

Over the course of a lifetime, we process a continual stream of information. Extracted from this stream, memories must be efficiently encoded and stored in an addressable manner for retrieval. To explore potential mechanisms, we consider a familiarity detection task in which a subject reports whether an image has been previously encountered. We design a feedforward network endowed with synaptic plasticity and an addressing matrix, meta-learned to optimize familiarity detection over long intervals. We find that anti-Hebbian plasticity leads to better performance than Hebbian plasticity and replicates experimental results such as repetition suppression. A combinatorial addressing function emerges, selecting a unique neuron as an index into the synaptic memory matrix for storage or retrieval. Unlike previous models, this network operates continuously and generalizes to intervals it has not been trained on. Our work suggests a biologically plausible mechanism for continual learning and demonstrates an effective application of machine learning for neuroscience discovery.

Contrastive Similarity Matching for Supervised Learning

We propose a novel biologically plausible solution to the credit assignment problem motivated by observations in the ventral visual pathway and trained deep neural networks. In both, representations of objects in the same category become progressively more similar, while objects belonging to different categories become less similar. We use this observation to motivate a layer-specific learning goal in a deep network: each layer aims to learn a representational similarity matrix that interpolates between previous and later layers. We formulate this idea using a contrastive similarity matching objective function and derive from it deep neural networks with feedforward, lateral, and feedback connections and neurons that exhibit biologically plausible Hebbian and anti-Hebbian plasticity. Contrastive similarity matching can be interpreted as an energy-based learning algorithm, but with significant differences from others in how a contrastive function is constructed.

Learning receptive field properties of complex cells in V1.

There are two distinct classes of cells in the primary visual cortex (V1): simple cells and complex cells. One defining feature of complex cells is their spatial phase invariance; they respond strongly to oriented grating stimuli with a preferred orientation but with a wide range of spatial phases. A classical model of complete spatial phase invariance in complex cells is the energy model, in which the responses are the sum of the squared outputs of two linear spatially phase-shifted filters. However, recent experimental studies have shown that complex cells have a diverse range of spatial phase invariance and only a subset can be characterized by the energy model. While several models have been proposed to explain how complex cells could learn to be selective to orientation but invariant to spatial phase, most existing models overlook many biologically important details. We propose a biologically plausible model for complex cells that learns to pool inputs from simple cells based on the presentation of natural scene stimuli. The model is a three-layer network with rate-based neurons that describes the activities of LGN cells (layer 1), V1 simple cells (layer 2), and V1 complex cells (layer 3). The first two layers implement a recently proposed simple cell model that is biologically plausible and accounts for many experimental phenomena. The neural dynamics of the complex cells is modeled as the integration of simple cells inputs along with response normalization. Connections between LGN and simple cells are learned using Hebbian and anti-Hebbian plasticity. Connections between simple and complex cells are learned using a modified version of the Bienenstock, Cooper, and Munro (BCM) rule. Our results demonstrate that the learning rule can describe a diversity of complex cells, similar to those observed experimentally.

Achieving stable dynamics in neural circuits.

The brain consists of many interconnected networks with time-varying, partially autonomous activity. There are multiple sources of noise and variation yet activity has to eventually converge to a stable, reproducible state (or sequence of states) for its computations to make sense. We approached this problem from a control-theory perspective by applying contraction analysis to recurrent neural networks. This allowed us to find mechanisms for achieving stability in multiple connected networks with biologically realistic dynamics, including synaptic plasticity and time-varying inputs. These mechanisms included inhibitory Hebbian plasticity, excitatory anti-Hebbian plasticity, synaptic sparsity and excitatory-inhibitory balance. Our findings shed light on how stable computations might be achieved despite biological complexity. Crucially, our analysis is not limited to analyzing the stability of fixed geometric objects in state space (e.g points, lines, planes), but rather the stability of state trajectories which may be complex and time-varying.

How Neural Circuits Achieve and Use Stable Dynamics

The brain consists of many interconnected networks with time-varying activity. There are multiple sources of noise and variation yet activity has to eventually converge to a stable state for its computations to make sense. We approached this from a control-theory perspective by applying contraction analysis to recurrent neural networks. This allowed us to find mechanisms for achieving stability in multiple connected networks with biologically realistic dynamics, including synaptic plasticity and time-varying inputs. These mechanisms included anti-Hebbian plasticity, synaptic sparsity and excitatory-inhibitory balance. We leveraged these findings to construct networks that could perform functionally relevant computations in the presence of noise and disturbance. Our work provides a blueprint for how to construct stable plastic and distributed networks.

Why Do Similarity Matching Objectives Lead to Hebbian/Anti-Hebbian Networks?

Modeling self-organization of neural networks for unsupervised learning using Hebbian and anti-Hebbian plasticity has a long history in neuroscience. Yet derivations of single-layer networks with such local learning rules from principled optimization objectives became possible only recently, with the introduction of similarity matching objectives. What explains the success of similarity matching objectives in deriving neural networks with local learning rules? Here, using dimensionality reduction as an example, we introduce several variable substitutions that illuminate the success of similarity matching. We show that the full network objective may be optimized separately for each synapse using local learning rules in both the offline and online settings. We formalize the long-standing intuition of the rivalry between Hebbian and anti-Hebbian rules by formulating a min-max optimization problem. We introduce a novel dimensionality reduction objective using fractional matrix exponents. To illustrate the generality of our approach, we apply it to a novel formulation of dimensionality reduction combined with whitening. We confirm numerically that the networks with learning rules derived from principled objectives perform better than those with heuristic learning rules.

Transcutaneous induction of stimulus-timing-dependent plasticity in dorsal cochlear nucleus.

The cochlear nucleus (CN) is the first site of multisensory integration in the ascending auditory pathway. The principal output neurons of the dorsal cochlear nucleus (DCN), fusiform cells, receive somatosensory information relayed by the CN granule cells from the trigeminal and dorsal column pathways. Integration of somatosensory and auditory inputs results in long-term enhancement or suppression in a stimulus-timing-dependent manner. Here, we demonstrate that stimulus-timing-dependent plasticity (STDP) can be induced in DCN fusiform cells using paired auditory and transcutaneous electrical stimulation of the face and neck to activate trigeminal and dorsal column pathways to the CN, respectively. Long-lasting changes in fusiform cell firing rates persisted for up to 2 h after this bimodal stimulation, and followed Hebbian or anti-Hebbian rules, depending on tone duration, but not somatosensory stimulation location: 50 ms paired tones evoked predominantly Hebbian, while 10 ms paired tones evoked predominantly anti-Hebbian plasticity. The tone-duration-dependent STDP was strongly correlated with first inter-spike intervals, implicating intrinsic cellular properties as determinants of STDP. This study demonstrates that transcutaneous stimulation with precise auditory–somatosensory timing parameters can non-invasively induce fusiform cell long-term modulation, which could be harnessed in the future to moderate tinnitus-related hyperactivity in DCN.

A role for mixed corollary discharge and proprioceptive signals in predicting the sensory consequences of movements.

Animals must distinguish behaviorally relevant patterns of sensory stimulation from those that are attributable to their own movements. In principle, this distinction could be made based on internal signals related to motor commands, known as corollary discharge (CD), sensory feedback, or some combination of both. Here we use an advantageous model system--the electrosensory lobe (ELL) of weakly electric mormyrid fish--to directly examine how CD and proprioceptive feedback signals are transformed into negative images of the predictable electrosensory consequences of the fish's motor commands and/or movements. In vivo recordings from ELL neurons and theoretical modeling suggest that negative images are formed via anti-Hebbian plasticity acting on random, nonlinear mixtures of CD and proprioception. In support of this, we find that CD and proprioception are randomly mixed in spinal mossy fibers and that properties of granule cells are consistent with a nonlinear recoding of these signals. The mechanistic account provided here may be relevant to understanding how internal models of movement consequences are implemented in other systems in which similar components (e.g., mixed sensory and motor signals and synaptic plasticity) are found.

Plastic corollary discharge predicts sensory consequences of movements in a cerebellum-like circuit.

The capacity to predict the sensory consequences of movements is critical for sensory, motor, and cognitive function. Though it is hypothesized that internal signals related to motor commands, known as corollary discharge, serve to generate such predictions, this process remains poorly understood at the neural circuit level. Here we demonstrate that neurons in the electrosensory lobe (ELL) of weakly electric mormyrid fish generate negative images of the sensory consequences of the fish's own movements based on ascending spinal corollary discharge signals. These results generalize previous findings describing mechanisms for generating negative images of the effects of the fish's specialized electric organ discharge (EOD) and suggest that a cerebellum-like circuit endowed with associative synaptic plasticity acting on corollary discharge can solve the complex and ubiquitous problem of predicting sensory consequences of movements.

Multimodal Integration in Granule Cells as a Basis for Associative Plasticity and Sensory Prediction in a Cerebellum-like Circuit

The recoding of diverse sensory and motor signals by granule cells (GCs) is probably critical for the function of cerebellar circuits, yet the nature of these transformations and their significance for cerebellar information processing remain poorly understood. In cerebellum-like structures in fish, anti-Hebbian plasticity at parallel fiber synapses generates "negative images" that act to cancel predictable patterns of electrosensory input. Here I test the hypothesis that GCs enhance the capacity of Purkinje-like cells to generate specific negative images by selectively encoding combinations of sensory and motor signals. Using in vivo whole-cell recordings, I show (1) that a subset of GCs integrate sensory and motor signals conveyed by distinct mossy fiber classes and (2) that Purkinje-like cells exhibit plastic changes specific to the combinations of signals that individual GCs encode. Consistent with influential theories of cerebellar function, these findings suggest that selective GC output enhances the capacity of Purkinje-like cells to acquire selectivity through associative plasticity.

Modulation of LTP/LTD balance in STDP by an activity-dependent feedback mechanism

Spike-timing-dependent plasticity (STDP) has been suggested to play a role in the development of functional neuronal connections. However, for STDP to contribute to the synaptic organization, its learning curve should satisfy a requirement that the magnitude of long-term potentiation (LTP) is approximately the same as that of long-term depression (LTD). Without such balance between LTP and LTD, all the synapses are potentiated toward the upper limit or depressed toward the lower limit. Therefore, in this study, we explore the mechanisms by which the LTP/LTD balance in STDP can be modulated adequately. We examine a plasticity model that incorporates an activity-dependent feedback (ADFB) mechanism, wherein LTP induction is suppressed by higher postsynaptic activity. In this model, strengthening an ADFB function gradually decreases the temporal average of the ratio of the magnitude of LTP to that of LTD, whereas enhancing background inhibition augments this ratio. Additionally, correlated inputs can be strengthened or weakened depending on whether the correlation time is shorter or longer than a threshold value, respectively, suggesting that STDP may lead to either Hebbian or anti-Hebbian plasticity outcomes. At an intermediate range of correlation times, the reversal between the two distinct plasticity regimes can occur by changing the level of ADFB modulation and inhibition, providing a physiological mechanism for neurons to select from functionally different forms of learning rules.

Learning to Encode Timing: Mechanisms of Plasticity in the Auditory Brainstem

Mechanisms of plasticity have traditionally been ascribed to higher-order sensory processing areas such as the cortex, whereas early sensory processing centers have been considered largely hard-wired. In agreement with this view, the auditory brainstem has been viewed as a nonplastic site, important for preserving temporal information and minimizing transmission delays. However, recent groundbreaking results from animal models and human studies have revealed remarkable evidence for cellular and behavioral mechanisms for learning and memory in the auditory brainstem.

Control of neuronal firing by dynamic parallel fiber feedback:implications for electrosensory reafference suppression

The cancellation of self-generated components of sensory inputs is a key function of sensory feedback pathways. In many systems, cerebellar parallel fiber feedback mediates this cancellation through anti-Hebbian plasticity, resulting in the generation of a negative image of the reafferent inputs. Parallel fiber feedback involves direct excitation and disynaptic inhibition as well as synaptic plasticity on multiple time scales. How the dynamics of these processes interact with anti-Hebbian plasticity to shape synaptic inputs and provide a cancellation mechanism remains unclear. In the present study, we investigated the influence of parallel fiber feedback onto pyramidal neurons of the electrosensory lateral line lobe (ELL) in weakly electric fish under open loop conditions. We mimicked naturalistic parallel fiber inputs in an ELL brain slice by implementing an experimentally based model of this synaptic pathway using dynamic clamp. We showed that as parallel fiber activity increases, the effective input to ELL pyramidal neurons changes from net excitation to net inhibition, resulting in a non-monotonic firing response. Using a model neuron, we found that this robust non-monotonic response is due to a shift from balanced excitation and inhibition at low parallel fiber input rates, to dominant inhibition at high input rates. We then showed that this non-monotonic response provides a simple basis for negative image generation. Through changes in the mean activation rate of parallel fibers, the feedback can switch roles between enhancement and suppression of sensory inputs in a manner that is directly determined by the slope of the non-monotonic response curve.