Computational Neuroscience seminars

Research in Neuroscience, as in many scientific disciplines, is undergoing a renaissance based on deep learning. Unique to Neuroscience, deep learning models can be used not only as a tool but interpreted as models of the brain. The central claims of recent deep learning-based models of brain circuits are that they shed light on fundamental functions being optimized or make novel predictions about neural phenomena. We show, through the case-study of grid cells in the entorhinal-hippocampal circuit, that one may get neither. We rigorously examine the claims of deep learning models of grid cells using large-scale hyperparameter sweeps and theory-driven experimentation, and demonstrate that the results of such models are more strongly driven by particular, non-fundamental, and post-hoc implementation choices than fundamental truths about neural circuits or the loss function(s) they might optimize. We discuss why these models cannot be expected to produce accurate models of the brain without the addition of substantial amounts of inductive bias, an informal No Free Lunch result for Neuroscience.

Studies of the mouse visual system have revealed a variety of visual brain areas in a roughly hierarchical arrangement, together with a multitude of behavioral capacities, ranging from stimulus-reward associations, to goal-directed navigation, and object-centric discriminations. However, an overall understanding of the mouse’s visual cortex organization, and how this organization supports visual behaviors, remains unknown. Here, we take a computational approach to help address these questions, providing a high-fidelity quantitative model of mouse visual cortex. By analyzing factors contributing to model fidelity, we identified key principles underlying the organization of mouse visual cortex. Structurally, we find that comparatively low-resolution and shallow structure were both important for model correctness. Functionally, we find that models trained with task-agnostic, unsupervised objective functions, based on the concept of contrastive embeddings were substantially better than models trained with supervised objectives. Finally, the unsupervised objective builds a general-purpose visual representation that enables the system to achieve better transfer on out-of-distribution visual, scene understanding and reward-based navigation tasks. Our results suggest that mouse visual cortex is a low-resolution, shallow network that makes best use of the mouse’s limited resources to create a light-weight, general-purpose visual system – in contrast to the deep, high-resolution, and more task-specific visual system of primates.

A complete neuroscience requires multi-level theories that address phenomena ranging from higher-level cognitive behaviors to activities within a cell. Unfortunately, we don't have cognitive models of behavior whose components can be decomposed into the neural dynamics that give rise to behavior, leaving an explanatory gap. Here, we decompose SUSTAIN, a clustering model of concept learning, into neuron-like units (SUSTAIN-d; decomposed). Instead of abstract constructs (clusters), SUSTAIN-d has a pool of neuron-like units. With millions of units, a key challenge is how to bridge from abstract constructs such as clusters to neurons, whilst retaining high-level behavior. How does the brain coordinate neural activity during learning? Inspired by algorithms that capture flocking behavior in birds, we introduce a neural flocking learning rule to coordinate units that collectively form higher-level mental constructs ("virtual clusters"), neural representations (concept, place and grid cell-like assemblies), and parallels recurrent hippocampal activity. The decomposed model shows how brain-scale neural populations coordinate to form assemblies encoding concept and spatial representations, and why many neurons are required for robust performance. Our account provides a multi-level explanation for how cognition and symbol-like representations are supported by coordinated neural assemblies formed through learning.

Neural computations are currently investigated using two separate approaches: sorting neurons into functional subpopulations or examining the low-dimensional dynamics of collective activity. Whether and how these two aspects interact to shape computations is currently unclear. Using a novel approach to extract computational mechanisms from networks trained on neuroscience tasks, here we show that the dimensionality of the dynamics and subpopulation structure play fundamentally com- plementary roles. Although various tasks can be implemented by increasing the dimensionality in networks with fully random population structure, flexible input–output mappings instead require a non-random population structure that can be described in terms of multiple subpopulations. Our analyses revealed that such a subpopulation structure enables flexible computations through a mechanism based on gain-controlled modulations that flexibly shape the collective dynamics. Our results lead to task-specific predictions for the structure of neural selectivity, for inactivation experiments and for the implication of different neurons in multi-tasking.

A long standing challenge in biological and artificial intelligence is to understand how new knowledge can be constructed from known building blocks in a way that is amenable for computation by neuronal circuits. Here we focus on the task of storage and recall of structured knowledge in long-term memory. Specifically, we ask how recurrent neuronal networks can store and retrieve multiple knowledge structures. We model each structure as a set of binary relations between events and attributes (attributes may represent e.g., temporal order, spatial location, role in semantic structure), and map each structure to a distributed neuronal activity pattern using a vector symbolic architecture (VSA) scheme. We then use associative memory plasticity rules to store the binarized patterns as fixed points in a recurrent network. By a combination of signal-to-noise analysis and numerical simulations, we demonstrate that our model allows for efficient storage of these knowledge structures, such that the memorized structures as well as their individual building blocks (e.g., events and attributes) can be subsequently retrieved from partial retrieving cues. We show that long-term memory of structured knowledge relies on a new principle of computation beyond the memory basins. Finally, we show that our model can be extended to store sequences of memories as single attractors.

Our perception of time isn’t like a clock; it varies depending on other aspects of experience, such as what we see and hear in that moment. However, in everyday life, the properties of these simple features can change frequently, presenting a challenge to understanding real-world time perception based on simple lab experiments. We developed a computational model of human time perception based on tracking changes in neural activity across brain regions involved in sensory processing, using fMRI. By measuring changes in brain activity patterns across these regions, our approach accommodates the different and changing feature combinations present in natural scenarios, such as walking on a busy street. Our model reproduces people’s duration reports for natural videos (up to almost half a minute long) and, most importantly, predicts whether a person reports a scene as relatively shorter or longer–the biases in time perception that reflect how natural experience of time deviates from clock time

The predictive map hypothesis provides a promising framework to model representations in the hippocampal formation. I will introduce a tractable implementation of a predictive map called the successor representation (SR), before presenting data showing that rats and humans display SR-like navigational choices on a novel open-field maze. Next, I will show how such a predictive map could be implemented using spatial representations found in the hippocampal formation, before finally presenting how such learning might be well approximated by phenomena that exist in the spatial memory system - namely spike-timing dependent plasticity and theta phase precession.

How seizures emerge from the abnormal dynamics of neural networks within the epileptogenic tissue remains an enigma. Are seizures random events, or do detectable changes in brain dynamics precede them? Are mechanisms of seizure emergence identical at the onset and later stages of epilepsy? Is the risk of seizure occurrence stable, or does it change over time? A myriad of questions about seizure genesis remains to be answered to understand the core principles governing seizure genesis. The last decade has brought unprecedented insights into the complex nature of seizure emergence. It is now believed that seizure onset represents the product of the interactions between the process of a transition to seizure, long-term fluctuations in seizure susceptibility, epileptogenesis, and disease progression. During the lecture, we will review the latest observations about mechanisms of ictogenesis operating at multiple temporal scales. We will show how the latest observations contribute to the formation of a comprehensive theory of seizure genesis, and challenge the traditional perspectives on ictogenesis. Finally, we will discuss how combining conventional approaches with computational modeling, modern techniques of in vivo imaging, and genetic manipulation open prospects for exploration of yet hidden mechanisms of seizure genesis.

Research in the brain and cognitive sciences attempts to uncover the neural mechanisms underlying intelligent behavior in domains such as vision. Due to the complexities of brain processing, studies necessarily had to start with a narrow scope of experimental investigation and computational modeling. I argue that it is time for our field to take the next step: build system models that capture a range of visual intelligence behaviors along with the underlying neural mechanisms. To make progress on system models, we propose integrative benchmarking – integrating experimental results from many laboratories into suites of benchmarks that guide and constrain those models at multiple stages and scales. We show-case this approach by developing Brain-Score benchmark suites for neural (spike rates) and behavioral experiments in the primate visual ventral stream. By systematically evaluating a wide variety of model candidates, we not only identify models beginning to match a range of brain data (~50% explained variance), but also discover that models’ brain scores are predicted by their object categorization performance (up to 70% ImageNet accuracy). Using the integrative benchmarks, we develop improved state-of-the-art system models that more closely match shallow recurrent neuroanatomy and early visual processing to predict primate temporal processing and become more robust, and require fewer supervised synaptic updates. Taken together, these integrative benchmarks and system models are first steps to modeling the complexities of brain processing in an entire domain of intelligence.

Electrophysiological recordings during perceptual decision tasks in monkeys suggest that the degree of confidence in a decision is based on a simple neural signal produced by the neural decision process. Attractor neural networks provide an appropriate biophysical modeling framework, and account for the experimental results very well. However, it remains unclear whether attractor neural networks can account for confidence reports in humans. We present the results from an experiment in which participants are asked to perform an orientation discrimination task, followed by a confidence judgment. Here we show that an attractor neural network model quantitatively reproduces, for each participant, the relations between accuracy, response times and confidence. We show that the attractor neural network also accounts for confidence-specific sequential effects observed in the experiment (participants are faster on trials following high confidence trials), as well as non confidence-specific sequential effects. Remarkably, this is obtained as an inevitable outcome of the network dynamics, without any feedback specific to the previous decision (that would result in, e.g., a change in the model parameters before the onset of the next trial). Our results thus suggest that a metacognitive process such as confidence in one’s decision is linked to the intrinsically nonlinear dynamics of the decision-making neural network.

Efficient navigation requires animals to track their position, velocity and heading direction (HD). Some animals’ behavior suggests that they also track uncertainties about these navigational variables, and make strategic use of these uncertainties, in line with a Bayesian computation. Ring-attractor networks have been proposed to estimate and track these navigational variables, for instance in the HD system of the fruit fly Drosophila. However, such networks are not designed to incorporate a notion of uncertainty, and therefore seem unsuited to implement dynamic Bayesian inference. Here, we close this gap by showing that specifically tuned ring-attractor networks can track both a HD estimate and its associated uncertainty, thereby approximating a circular Kalman filter. We identified the network motifs required to integrate angular velocity observations, e.g., through self-initiated turns, and absolute HD observations, e.g., visual landmark inputs, according to their respective reliabilities, and show that these network motifs are present in the connectome of the Drosophila HD system. Specifically, our network encodes uncertainty in the amplitude of a localized bump of neural activity, thereby generalizing standard ring attractor models. In contrast to such standard attractors, however, proper Bayesian inference requires the network dynamics to operate in a regime away from the attractor state. More generally, we show that near-Bayesian integration is inherent in generic ring attractor networks, and that their amplitude dynamics can account for close-to-optimal reliability weighting of external evidence for a wide range of network parameters. This only holds, however, if their connection strengths allow the network to sufficiently deviate from the attractor state. Overall, our work offers a novel interpretation of ring attractor networks as implementing dynamic Bayesian integrators. We further provide a principled theoretical foundation for the suggestion that the Drosophila HD system may implement Bayesian HD tracking via ring attractor dynamics.

Animals exhibit remarkable learning capabilities with little direct supervision. Likewise, self-supervised learning is an emergent paradigm in artificial intelligence, closing the performance gap to supervised learning. In the context of biology, self-supervised learning corresponds to a setting where one sense or specific stimulus may serve as a supervisory signal for another. After learning, the latter can be used to predict the former. On the implementation level, it has been demonstrated that such predictive learning can occur at the single neuron level, in compartmentalized neurons that separate and associate information from different streams. We demonstrate the power such self-supervised learning over unsupervised (Hebb-like) learning rules, which depend heavily on stimulus statistics, in two examples: First, in the context of animal navigation where predictive learning can associate internal self-motion information always available to the animal with external visual landmark information, leading to accurate path-integration in the dark. We focus on the well-characterized fly head direction system and show that our setting learns a connectivity strikingly similar to the one reported in experiments. The mature network is a quasi-continuous attractor and reproduces key experiments in which optogenetic stimulation controls the internal representation of heading, and where the network remaps to integrate with different gains. Second, we show that incorporating global gating by reward prediction errors allows the same setting to learn conditioning at the neuronal level with mixed selectivity. At its core, conditioning entails associating a neural activity pattern induced by an unconditioned stimulus (US) with the pattern arising in response to a conditioned stimulus (CS). Solving the generic problem of pattern-to-pattern associations naturally leads to emergent cognitive phenomena like blocking, overshadowing, saliency effects, extinction, interstimulus interval effects etc. Surprisingly, we find that the same network offers a reductionist mechanism for causal inference by resolving the post hoc, ergo propter hoc fallacy.

Current SNNs studies frequently ignore dendrites, the thin membranous extensions of biological neurons that receive and preprocess nearly all synaptic inputs in the brain. However, decades of experimental and theoretical research suggest that dendrites possess compelling computational capabilities that greatly influence neuronal and circuit functions. Notably, standard point-neuron networks cannot adequately capture most hallmark dendritic properties. Meanwhile, biophysically detailed neuron models are impractical for large-network simulations due to their complexity, and high computational cost. For this reason, we introduce Dendrify, a new theoretical framework combined with an open-source Python package (compatible with Brian2) that facilitates the development of bioinspired SNNs. Dendrify, through simple commands, can generate reduced compartmental neuron models with simplified yet biologically relevant dendritic and synaptic integrative properties. Such models strike a good balance between flexibility, performance, and biological accuracy, allowing us to explore dendritic contributions to network-level functions while paving the way for developing more realistic neuromorphic systems.

Neuronal network computation and computation by avalanche supporting networks are of interest to the fields of physics, computer science (computation theory as well as statistical or machine learning) and neuroscience. Here we show that computation of complex Boolean functions arises spontaneously in threshold networks as a function of connectivity and antagonism (inhibition), computed by logic automata (motifs) in the form of computational cascades. We explain the emergent inverse relationship between the computational complexity of the motifs and their rank-ordering by function probabilities due to motifs, and its relationship to symmetry in function space. We also show that the optimal fraction of inhibition observed here supports results in computational neuroscience, relating to optimal information processing.

To move efficiently, animals must continuously work out their x,y,z positions with respect to real-world objects, and many animals have a pair of eyes to achieve this. How photoreceptors actively sample the eyes’ optical image disparity is not understood because this fundamental information-limiting step has not been investigated in vivo over the eyes’ whole sampling matrix. This integrative multiscale study will advance our current understanding of stereopsis from static image disparity comparison to a morphodynamic active sampling theory. It shows how photomechanical photoreceptor microsaccades enable Drosophila superresolution three-dimensional vision and proposes neural computations for accurately predicting these flies’ depth-perception dynamics, limits, and visual behaviors.

An object’s colour, brightness and pattern are all influenced by its surroundings, and a number of visual phenomena and “illusions” have been discovered that highlight these often dramatic effects. Explanations for these phenomena range from low-level neural mechanisms to high-level processes that incorporate contextual information or prior knowledge. Importantly, few of these phenomena can currently be accounted for when measuring an object’s perceived colour. Here we ask to what extent colour appearance is predicted by a model based on the principle of coding efficiency. The model assumes that the image is encoded by noisy spatio-chromatic filters at one octave separations, which are either circularly symmetrical or oriented. Each spatial band’s lower threshold is set by the contrast sensitivity function, and the dynamic range of the band is a fixed multiple of this threshold, above which the response saturates. Filter outputs are then reweighted to give equal power in each channel for natural images. We demonstrate that the model fits human behavioural performance in psychophysics experiments, and also primate retinal ganglion responses. Next we systematically test the model’s ability to qualitatively predict over 35 brightness and colour phenomena, with almost complete success. This implies that contrary to high-level processing explanations, much of colour appearance is potentially attributable to simple mechanisms evolved for efficient coding of natural images, and is a basis for modelling the vision of humans and other animals.

Which nodes in a brain network causally influence one another, and how do such interactions utilize the underlying structural connectivity? One of the fundamental goals of neuroscience is to pinpoint such causal relations. Conventionally, these relationships are established by manipulating a node while tracking changes in another node. A causal role is then assigned to the first node if this intervention led to a significant change in the state of the tracked node. In this presentation, I use a series of intuitive thought experiments to demonstrate the methodological shortcomings of the current ‘causation via manipulation’ framework. Namely, a node might causally influence another node, but how much and through which mechanistic interactions? Therefore, establishing a causal relationship, however reliable, does not provide the proper causal understanding of the system, because there often exists a wide range of causal influences that require to be adequately decomposed. To do so, I introduce a game-theoretical framework called Multi-perturbation Shapley value Analysis (MSA). Then, I present our work in which we employed MSA on an Echo State Network (ESN), quantified how much its nodes were influencing each other, and compared these measures with the underlying synaptic strength. We found that: 1. Even though the network itself was sparse, every node could causally influence other nodes. In this case, a mere elucidation of causal relationships did not provide any useful information. 2. Additionally, the full knowledge of the structural connectome did not provide a complete causal picture of the system either, since nodes frequently influenced each other indirectly, that is, via other intermediate nodes. Our results show that just elucidating causal contributions in complex networks such as the brain is not sufficient to draw mechanistic conclusions. Moreover, quantifying causal interactions requires a systematic and extensive manipulation framework. The framework put forward here benefits from employing neural network models, and in turn, provides explainability for them.

Spiking recurrent neural networks (RNNs) are a promising tool for solving a wide variety of complex cognitive and motor tasks, due to their rich temporal dynamics and sparse processing. However training spiking RNNs on dedicated neuromorphic hardware is still an open challenge. This is due mainly to the lack of local, hardware-friendly learning mechanisms that can solve the temporal credit assignment problem and ensure stable network dynamics, even when the weight resolution is limited. These challenges are further accentuated, if one resorts to using memristive devices for in-memory computing to resolve the von-Neumann bottleneck problem, at the expense of a substantial increase in variability in both the computation and the working memory of the spiking RNNs. In this talk, I will present our recent work where we introduced a PyTorch simulation framework of memristive crossbar arrays that enables accurate investigation of such challenges. I will show that recently proposed e-prop learning rule can be used to train spiking RNNs whose weights are emulated in the presented simulation framework. Although e-prop locally approximates the ideal synaptic updates, it is difficult to implement the updates on the memristive substrate due to substantial device non-idealities. I will mention several widely adapted weight update schemes that primarily aim to cope with these device non-idealities and demonstrate that accumulating gradients can enable online and efficient training of spiking RNN on memristive substrates.

Change is ubiquitous in living beings. In particular, the connectome and neural representations can change. Nevertheless behaviors and memories often persist over long times. In a standard model, associative memories are represented by assemblies of strongly interconnected neurons. For faithful storage these assemblies are assumed to consist of the same neurons over time. We propose a contrasting memory model with complete temporal remodeling of assemblies, based on experimentally observed changes of synapses and neural representations. The assemblies drift freely as noisy autonomous network activity or spontaneous synaptic turnover induce neuron exchange. The exchange can be described analytically by reduced, random walk models derived from spiking neural network dynamics or from first principles. The gradual exchange allows activity-dependent and homeostatic plasticity to conserve the representational structure and keep inputs, outputs and assemblies consistent. This leads to persistent memory. Our findings explain recent experimental results on temporal evolution of fear memory representations and suggest that memory systems need to be understood in their completeness as individual parts may constantly change.

Efficient coding has served as a guiding principle in understanding the neural code. To date, however, it has been explored mainly in the context of peripheral sensory cells with simple tuning curves. By contrast, ‘deeper’ neurons such as grid cells come with more complex tuning properties which imply a different, yet highly efficient, strategy for representing information. I will show that a highly efficient code is not specific to a population of neurons with finely tuned response properties: it emerges robustly in a shallow network with random synapses. Here, the geometry of population responses implies that optimality obtains from a tradeoff between two qualitatively different types of error: ‘local’ errors (common to classical neural population codes) and ‘global’ (or ‘catastrophic’) errors. This tradeoff leads to efficient compression of information from a high-dimensional representation to a low-dimensional one. After describing the theoretical framework, I will use it to re-interpret recordings of motor cortex in behaving monkey. Our framework addresses the encoding of (sensory) information; if time allows, I will comment on ongoing work that focuses on decoding from the perspective of efficient coding.