Biology subjects

Mihalas, S.

Publications and source records attributed to Mihalas, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Discrete synaptic states and context-modulated readouts support continual learning

Biological neural systems learn new behaviors while retaining earlier ones, whereas sequentially trained artificial networks often overwrite parameters and forget previously learned behaviors. We introduce the Context-modulated Synaptic-states Recurrent Neural Network (CoSyn-RNN), a new continual-learning model inspired by heterogeneous synaptic states and neuromodulatory control. CoSyn-RNN combines sparse recurrent allocation with a discrete protection rule: after each task, recurrent neurons whose synaptic weights cross a fixed threshold form a task-specific neuron group, and their associated parameters enter a protected state during later learning. A neuron-specific gain and a task-cued mask over a fixed base readout control how recurrent activity contributes to behavior. Across various task sequences consisting of many cognitive tasks, training and validation losses and accuracies were maintained throughout learning. The protection rule produced modular, task-ordered recurrent structure compatible with forward reuse of earlier computations. Pairwise recruitment measurements further defined minimum- and maximum-cost curricula that produced markedly different final recurrent structures and different remaining capacity for future learning. Thus, CoSyn-RNN provides a biologically inspired model in which protected synaptic states, sparse allocation, and context-dependent modulation support continual learning.

neuroscience↗

bioRxiv2026

The Unreasonable Effectiveness of Cell Types in Describing Neuronal Physiological Features

Single-cell RNA sequencing (scRNA-seq) captures detailed gene expression profiles at scale, while patch-clamp recordings measure intrinsic neuronal electrophysiological properties. Modeling the relations between these two modalities remains a challenge. Here, we compare how well electrophysiological features can be predicted by traditional transcriptomic cell type classification, representations derived from a foundational model (scGPT) pretrained on large-scale scRNA-seq datasets, ion channel-coding genes, and highly variable genes. Using paired transcriptomic and electrophysiological patch-sequencing data from 495 human neurons from neurosurgical tissue, we find that cluster-level cell type representations generally outperform highly variable gene selection, ion channel gene selection, and context-enriched scGPT embeddings. Notably, performance varies across model architectures and initializations, and the best results are obtained by combining the outputs of separate cell type and scGPT-based models. Together, these findings suggest that traditional discrete cellular classification is highly effective in predicting physiological features. For maximum performance it can be complemented by pretrained transformer models.

neuroscience↗

bioRxiv2026

Learning using switching synaptic plasticity rules

Synapses differ in molecular and ultrastructural state, but the computational role of this heterogeneity remains unclear and computational models do not account for it. Local Hebbian-like plasticity is well supported experimentally, whereas learning complex behaviors most likely requires non-local credit-assignment signals. Recent cortical electron microscopy data suggest that strong and weak synapses occupy distinct structural states, with large synapses often containing a spine apparatus that can alter calcium dynamics and plasticity. Here, we assigned different computational roles to different morphological synaptic types: we asked how networks with state-dependent plasticity operate and whether the mechanisms and structural features they develop are consistent with biological observations. We built a plasticity-switching recurrent neural network (psRNN) in which synaptic state is modeled by recurrent-weight thresholds. Synapses in the low-weight state use a local Hebbian-like rule, whereas synapses in the high-weight state receive an idealized non-local credit-assignment signal implemented with backpropagation through time. psRNNs learned cognitive tasks with temporal and working-memory demands in fewer trials than fully backpropagation-trained RNNs, despite routing BP-based updates to fewer recurrent weights. The advantage arose from three interacting mechanisms: 1) Hebbian-like updates exposed the credit-assignment component to several nearby parameter configurations, 2) the network dynamically built a task-relevant recurrent initialization, and 3) synapses switched into the credit-assignment subset before Hebbian growth destabilized learning. Structurally, trained psRNNs produced more anti-symmetric weights, lower effective rank, and stronger feedforward structure than control RNNs. Our results shed light on the computational role of heterogeneous synapses, which may regulate distinct types of learning signals, and we generate falsifiable predictions for connectomic analyses of recurrent cortical circuits.

neuroscience↗

bioRxiv2026

Functional reorganization of motor cortex connectivity during learning

Learning new tasks requires the brain to reshape the flow of neural activity, but how these changes arise from dynamic neural connectivity remains unclear. Here, we used two-photon photostimulation and calcium imaging to map learning-related changes in connectivity in layer 2/3 of mouse motor cortex, induced by learning of an optical brain-computer interface (BCI) task. Mice rapidly (within minutes) learned to change activity in a conditioned neuron to earn rewards. Activity changes were sparse; the conditioned neuron increased activity more than surrounding neurons. Mapping connectivity before and after learning revealed changes in motor cortex connectivity, enriched in neurons that were active before trial initiation, analogous to motor cortex populations that are active preceding movement. Motor cortex plasticity reroutes preparatory activity to neurons that are active later and control the conditioned neuron. Our findings show how rapid learning can be achieved through structured changes in motor cortex connectivity.

neuroscience↗

bioRxiv2025

A dataset of differentiable biologically-derived single neuron models

Biological neural networks contain diverse cell types with heterogeneous electrophysiological properties. Artificial neural networks (ANNs) model computational aspects of biology but use homogeneous neurons, limiting realism. A key obstacle to building bio-inspired ANNs is the absence of a database of neuron models that both fit biological data and remain differentiable for gradient-based learning. We present such a database: over 1,000 differentiable single-neuron firing-rate models with accompanying PyTorch code for integration with machine learning. The models belong to the linear-nonlinear (LN) class, which, for ease of reference, we term Generalized Firing Rate (GFR) neurons. Each GFR neuron uses input and firing-rate history filters across multiple timescales, followed by a nonlinearity to generate firing rates. Models are fit to patch-clamp recordings from mouse and human central nervous system slices. Parameter clustering reflects electrophysiological diversity and partially aligns with transgenic lines. This resource will enable development of bio-inspired ANNs built from biologically grounded, differentiable single-neuron models. HighlightsO_LIResource of over 1,000 differentiable single-neuron models fit to mouse and human patch-clamp recordings C_LIO_LIOpen source PyTorch code and reproducible pipelines support model fitting, evaluation, and network integration C_LIO_LIFitted model parameters capture electrophysiological diversity, with parameter clusters partially aligning to inhibitory/excitatory types and Cre driver lines C_LI

neuroscience↗

bioRxiv2025

Slow Transition to Low-Dimensional Chaos in Heavy-Tailed Recurrent Neural Networks

Growing evidence suggests that synaptic weights in the brain follow heavy-tailed distributions, yet most theoretical analyses of recurrent neural networks (RNNs) assume Gaussian connectivity. We systematically study the activity of RNNs with random weights drawn from biologically plausible Levy alpha-stable distributions. While mean-field theory for the infinite system predicts that the quiescent state is always unstable--implying ubiquitous chaos--our finite-size analysis reveals a sharp transition between quiescent and chaotic dynamics. We theoretically predict the gain at which the finite system transitions from quiescent to chaotic dynamics, and validate it through simulations. Compared to Gaussian networks, finite heavy-tailed RNNs exhibit a broader gain regime near the edge of chaos, namely, a slow transition to chaos. However, this robustness comes with a tradeoff: heavier tails reduce the Lyapunov dimension of the attractor, indicating lower effective dimensionality. Our results reveal a biologically aligned tradeoff between the robustness of dynamics near the edge of chaos and the richness of high-dimensional neural activity. By analytically characterizing the transition point in finite-size networks-- where mean-field theory breaks down--we provide a tractable framework for understanding dynamics in realistically sized, heavy-tailed neural circuits.*

neuroscience↗

bioRxiv2025

Map of spiking activity underlying change detection in the mouse visual system

Visual behavior requires coordinated activity across hierarchically organized brain circuits. Understanding this complexity demands datasets that are both large-scale (sampling many areas) and dense (recording many neurons in each area). Here we present a database of spiking activity across the mouse visual system--including thalamus, cortex, and midbrain--while mice perform an image change detection task. Using Neuropixels probes, we record from >75,000 high-quality units in 54 mice, mapping area-, cortical layer-, and cell type-specific coding of sensory and motor information. Modulation by task-engagement increased across the thalamocortical hierarchy but was strongest in the midbrain. Novel images modulated cortical (but not thalamic) responses through delayed recurrent activity. Population decoding and optogenetics identified a critical decision window for change detection and revealed that mice use an adaptation-based rather than image-comparison strategy. This comprehensive resource provides a valuable substrate for understanding sensorimotor computations in neural networks.

neuroscience↗

bioRxiv2025

Predicting Neural Activity from Connectome Embedding Spaces

Understanding how structured patterns of neural activity emerge from the underlying connectivity is fundamental to elucidating brain function. While cortical connectomes are intrinsically high-dimensional, population activity typically resides in a much lower-dimensional subspace. Consequently, only a small fraction of the information encoded in the connectome appears relevant for shaping activity. Can we identify low-dimensional features of the connectome features be that reliably predict neural activity? Leveraging the MICrONS dataset, which combines millimeter-scale, nanometer-resolution connectivity with simultaneously recorded in-vivo activity, we demonstrate a statistically significant alignment between morphological and functional similarity, quantified by subspace angles and centered kernel alignment. Topological analyses further reveal that the representation spaces of both the connectome and neural activity share a low-dimensional hyperbolic geometry with exponential scaling. These parallels motivated the hypothesis that embedding anatomical affinities into an appropriate geometric space can isolate the functionally relevant features of the connectome. We therefore applied multidimensional scaling to generate such embeddings and trained a simple linear model to reconstruct neuronal activity. Remarkably, the embedded connectome explained 68% in activity similarity, surpassing models that had direct access to activity similarity itself and outperforming similarly simple models that used the full high-dimensional connectome (56%). Our findings uncover a robust structure-function coupling: geometry-aware dimensionality reduction discards much of the connectomes microscopic detail yet yields superior predictions of neural activity. This suggests that synaptic wiring implicitly encodes an abstract, low-dimensional organization that underlies the observed low-dimensional cortical activity.

neuroscience↗

bioRxiv2025

Integrated Classification of Cortical Cells and Quantitative Projectomic Mapping Unveil Organizational Principles of Brain-Wide Connectomes at Single Cell Level

Molecularly defined cortical cell types have recently been linked to whole neuronal morphology (WNM), particularly those characterized by whole-brain-wide projections, such as intratelencephalic (IT), extratelencephalic (ET), and corticothalamic (CT) neurons. In contrast, classical morphological classifications (e.g., tufted TPC, small tufted SPC, and stellate SSC) are based primarily on local dendrosomatic and axonal structures, especially apical dendrites. This study bridges these perspectives by establishing a new neuronal taxonomy, analyzing the connectomes of defined cortical cell types, and comparing them with those obtained from bulk anterograde injections. Neurons were sparsely labeled via tamoxifen-inducible Cre lines with GFP reporters, and 1,419 WNM cells were comprehensively reconstructed with Vaa3D-TeraVR from [~]15 areas across six functional regions of molecularly labeled brains imaged with 2p-fMOST. These cells were newly classified by integrating current molecular-WNM and classical morphological perspectives, with sample size augmented by 1,455 publicly available WNM cells reconstructed from the Mouse-Light project and CEBSIT. This effort defined ten combined molecular-WNM-classical morphological cell types: L5ET_TPC, L6CT_NPC, L6b_HPC, and seven IT types--L2/3IT_TPC, L4IT_SSC, L4IT_UPC, L4IT_TPC, L5IT_SPC, L6IT_IPC, and L6IT_car3PC. Clustering, quantitative analyses and random Forest classifier objectively validated these types and revealed their distinct connectomes, along with convergent, topographic, and hierarchical organizations across their projection brain regions. At the single-cell level, multiple organizational principles governing cortico-cortical (C-C) and cortico-subcortical (C-subC) connectomes emerged with unprecedented detail, offering a precise GPS-like tool for in vivo recordings and robust datasets for neuronal network modeling. Comparisons with bulk anterograde injection data underscored the limitations of traditional methods in identifying projection targets. Overall, our approach provides significant insights into cortical circuitry and elucidates the complex interplay between neuronal molecular identity, whole morphology, and classical morphological classification.

neuroscience↗

bioRxiv2025

Hebb's Vision: The Structural Underpinnings of Hebbian Assemblies

1In 1949, Donald Hebb proposed that groups of neurons that activate stereotypically form the organizational building blocks of perception, cognition, and behavior. Finding the structural underpinning of such assemblies has been technically challenging, due to a lack of large-scale structure-activity maps. Here, we analyze this relation using a novel dataset that links in vivo optical physiology to connectivity using postmortem elec-tron microscopy (EM). From the fluorescence traces, we extract neural assemblies from higher-order correlations in neural activity. Physiologically, we show that these assemblies exhibit properties consistent with Hebbs theory, including more reliable responses to repeated natural movie inputs than size-matched random ensembles and superior decoding of visual stimuli. Structurally, we find that neurons that participate in assemblies are significantly more integrated into the structural network than those that do not. Contrary to Hebbs original prediction, we do not observe a marked increase in the strength of monosynaptic excitatory connections between cells participating in the same assembly. However, we find significantly stronger indirect feed-forward inhibitory connections targeting cells in other assemblies. These results show that assemblies can be useful components of perception, and, surprisingly, they are delineated by mutual inhibition.

neuroscience↗

bioRxiv2024

Spatial organization of multisensory convergence in mouse isocortex

1The diverse functions of different cortical areas are thought to arise from their distinct groups of inputs. However, additional organizing principles may exist in the spatial structure of converging inputs. We investigated spatial convergence patterns of projections from primary sensory areas to other areas throughout the mouse isocortex. We used a large tract tracing dataset to estimate the dimension of the space into which topographical connections from multiple modalities converged within each other cortical area. We call this measure the topography dimension (TD). TD is higher for areas that receive inputs of similar strength from multiple sensory modalities, and lower when multiple inputs terminate in register with one another. Across the isocortex, TD varied by a factor of 4. TD was positively correlated with hierarchy score, an independent measure that is based on laminar connection patterns. Furthermore, TD (an anatomical measure) was significantly related to several measures of neural activity. In particular, higher TD was associated with higher neural activity dimension, lower population sparseness, and lower lifetime sparseness of spontaneous activity, independent of an areas hierarchical position. Finally, we analyzed factors that limited TD and found that linear correlations among projections from different areas typically had little impact, while diversity of connection strengths, both between different projections onto the same area, and within projections across different parts of an area, limited TD substantially. This analysis revealed additional intricacy of cortical networks, beyond areas sets of connections and hierarchical organization. We propose a means of approximating this organization in deep-network models.

neuroscience↗

bioRxiv2024

Communication subspaces align with training in ANNs

Communication subspaces have recently been identified as a promising mechanism for selectively routing information between brain areas. In this study, we explored whether communication sub-spaces develop with training in artificial neural networks (ANNs) and explored differences across connection types. Specifically, we analyzed the subspace angles between activations and weights in ResNet-50 before and after training. We found that activations were more aligned to the weight layers after training, although this effect decreased in deeper layers. We also analyzed the angles between pairs of weight layers. We found that for all branching, direct, and skip connections, weight layer pairs were more geometrically aligned in trained versus untrained models throughout the entire network. These findings indicate that such alignment is essential for the proper functioning of deep networks and highlights the potential to enhance training efficiency through pre-alignment. In biological data, our results motivate further exploration into whether learning induces similar subspace alignment.

neuroscience↗

bioRxiv2024

Complex Properties of Training Stimuli Affect Brain Alignment in a Deep Network Model of Mouse Visual Cortex

Deep convolutional neural networks are important models of the visual cortex that ac-count relatively well for brain activity and are able to perform ethologically relevant functions. However, it is unknown which combination of factors, such as network ar-chitecture, training objectives, and data best align this family of models with the brain. Here we investigate the statistics of training data. We hypothesized that stimuli that are naturalistic for mice would lead to higher similarity between deep network models and activity in mouse visual cortex. We used a video-game engine to create training datasets in which we varied the naturalism of the environment, the movement statis-tics, and the optics of the modelled eye. The naturalistic environment substantially and consistently led to greater brain similarity, while the other factors had more subtle and area-specific effects. We then hypothesized that differences in brain similarity between the two environments arose due to differences in spatial frequency spectra, distribu-tions of color and orientation, and/or temporal autocorrelations. To test this, we created abstract environments, composed of cubes and spheres, that resembled the naturalis-tic and non-naturalistic environments in these respects. Contrary to our expectations, these factors accounted poorly for differences in brain similarity due to the naturalis-tic and non-naturalistic environments. This suggests that the higher brain similarities we observed after training with the naturalistic environment were due to more complex factors.

neuroscience↗

bioRxiv2024

Transition to chaos separates learning regimes and relates to measure of consciousness in recurrent neural networks

AbstractRecurrent neural networks exhibit chaotic dynamics when the variance in their connection strengths exceed a critical value. Recent work indicates connection variance also modulates learning strategies; networks learn "rich" representations when initialized with low coupling and "lazier"solutions with larger variance. Using Watts-Strogatz networks of varying sparsity, structure, and hidden weight variance, we find that the critical coupling strength dividing chaotic from ordered dynamics also differentiates rich and lazy learning strategies. Training moves both stable and chaotic networks closer to the edge of chaos, with networks learning richer representations before the transition to chaos. In contrast, biologically realistic connectivity structures foster stability over a wide range of variances. The transition to chaos is also reflected in a measure that clinically discriminates levels of consciousness, the perturbational complexity index (PCIst). Networks with high values of PCIst exhibit stable dynamics and rich learning, suggesting a consciousness prior may promote rich learning. The results suggest a clear relationship between critical dynamics, learning regimes and complexity-based measures of consciousness.

neuroscience↗

bioRxiv2023

Familiarity-modulated synapses model visual cortical circuit novelty responses

Since environments are constantly in flux, the brains ability to identify novel stimuli that fall outside its own internal representation of the world is crucial for an organisms survival. Within the mammalian neocortex, inhibitory microcircuits are proposed to regulate activity in an experience-dependent manner and different inhibitory neuron subtypes exhibit distinct novelty responses. Discerning the function of diverse neural circuits and their modulation by experience can be daunting unless one has a biologically plausible mechanism to detect and learn from novel experiences that is both understandable and flexible. Here we introduce a learning mechanism, familiarity modulated synapses (FMSs), through which a network response that encodes novelty emerges from unsupervised multiplicative synaptic modifications depending only on the presynaptic or both the pre- and postsynaptic activity. FMSs stand apart from other familiarity mechanisms in their simplicity: they operate under continual learning, do not require specialized architecture, and can distinguish novelty rapidly without requiring feedback. Implementing FMSs within an experimentally-constrained model of a visual cortical circuit, we demonstrate the generalizability of FMSs by reproducing three distinct novelty effects recently observed in experiments: absolute, contextual (or oddball), and omission novelty. Additionally, our model reproduces functional diversity within cell subpopulations, leading to experimentally testable predictions about connectivity and synaptic dynamics that can produce both population-level novelty responses and heterogeneous individual neuron signals. Altogether, our findings demonstrate how simple plasticity mechanisms within the cortical circuit structure can give rise to qualitatively distinct novelty responses. The flexibility of FMSs opens the door to computationally and theoretically investigating how distinct synapse modulations can lead to a variety of experience-dependent responses in a simple, understandable, and biologically plausible setup.

neuroscience↗

bioRxiv2023

Modeling the cell-type specific mesoscale murine connectome with anterograde tracing experiments

The Allen Mouse Brain Connectivity Atlas (MCA) consists of anterograde tracing experiments targeting diverse structures and classes of projecting neurons. Beyond regional anterograde tracing done in C57BL/6 wild type mice, a large fraction of experiments are performed using transgenic Cre-lines. This allows access to cell-class specific whole brain connectivity information, with class defined by the transgenic lines. However, even though the number of experiments is large, it does not come close to covering all existing cell classes in every area where they exist. Here, we study how much we can fill in these gaps and estimate the cell-class specific connectivity function given the simplifying assumptions that nearby voxels have smoothly varying projections, but that these projection tensors can change sharply depending on the region and class of the projecting cells. This paper describes the conversion of Cre-line tracer experiments into class-specific connectivity matrices representing the connection strengths between source and target structures. We introduce and validate a novel statistical model for creation of connectivity matrices. We extend the Nadaraya-Watson kernel learning method which we previously used to fill in spatial gaps to also fill in a gaps in cell-class connectivity information. To do this, we construct a "cell-class space" based on class-specific averaged regionalized projections and combine smoothing in 3D space as well as in this abstract space to share information between similar neuron classes. Using this method we construct a set of connectivity matrices using multiple levels of resolution at which discontinuities in connectivity are assumed. We show that the connectivities obtained from this model display expected cell-type and structure specific connectivities. We also show that the wild type connectivity matrix can be factored using a sparse set of factors, and analyze the informativeness of this latent variable model. AUTHOR SUMMARYLarge-scale studies have described the connections between areas in multiple mammalian models in ever expanding detail. Standard connectivity studies focus on the connection strength between areas. However, when describing functions at a local circuit level, there is an increasing focus on cell types. We have recently described the importance of connection types in the cortico-thalamic system, which allows an unsupervised discovery of its hierarchical organization. In this study we focus on adding a dimension of connection type for a brain-wide mesoscopic connectivity model. Even with our relatively massive dataset, the data in the cell type direction for connectivity is quite sparse, and we had to develop methods to more reliably extrapolate in such directions, and to estimate when such extrapolations are impossible. This allows us to fill in such a connection type specific inter-areal connectivity matrix to the extent our data allows. While analyzing this complex connectivity, we observed that it can be described via a small set of factors. While not complete, this connectivity matrix represents a a categorical and quantitative improvement in mouse mesoscale connectivity models.

neuroscience↗

bioRxiv2022

Neural Population Dynamics of Computing with Synaptic Modulations

In addition to long-time scale rewiring, synapses in the brain are subject to significant modulation that occurs at much shorter time scales and allows them to process short-term information. Despite this, models of the brain like recurrent neural networks (RNNs) often have their weights frozen after training, relying on an internal state stored in neuron activity to process temporal information. Although networks with dynamical synapses have been explored previously, often said dynamics are added to networks that also have recurrent connections and thus the short-time scale computational capabilities of synapse modulation alone remain unclear. In this work, we analyze the dynamics of a network that relies solely on synaptic modulations to process short-time scale information, the multi-plasticity network (MPN). We thoroughly examine the neural population dynamics of the MPN trained on integration-based tasks and compare it to known RNN dynamics, findings the two to have fundamentally different behavior and attractor structure. We find said differences in dynamics allow the MPN to outperform its RNN counterparts on several neuroscience-relevant tasks. Of note, the MPN has a significantly simpler attractor structure that allows it to be more flexible in training and sequential-learning settings. Lastly, how the dynamics change for MPNs trained on contextual and continuous integration tasks is also investigated.

neuroscience↗

bioRxiv2022

Heterogeneity in Neuronal Dynamics is Learned by Gradient Descent for Temporal Processing Tasks

Individual neurons in the brain have complex intrinsic dynamics that are highly diverse. We hypothesize that the complex dynamics produced by networks of complex and heterogeneous neurons may contribute to the brains ability to process and respond to temporally complex data. To study the role of complex and heterogeneous neuronal dynamics in network computation, we develop a rate-based neuronal model, the generalized-leaky-integrate-and-firing-rate (GLIFR) model, which is a rate-equivalent of the generalized-leaky-integrate-and-fire model. The GLIFR model has multiple dynamical mechanisms which add to the complexity of its activity while maintaining differentiability. We focus on the role of after-spike currents, currents induced or modulated by neuronal spikes, in producing rich temporal dynamics. We use machine learning techniques to learn both synaptic weights and parameters underlying intrinsic dynamics to solve temporal tasks. The GLIFR model allows us to use standard gradient descent techniques rather than surrogate gradient descent, which has been utilized in spiking neural networks. After establishing the ability to optimize parameters using gradient descent in single neurons, we ask how networks of GLIFR neurons learn and perform on temporally challenging tasks, such as sinusoidal pattern generation and sequential MNIST. We find that these networks learn a diversity of parameters, which gives rise to diversity in neuronal dynamics. We also observe that training networks on the sequential MNIST task leads to formation of cell classes based on the clustering of neuronal parameters. GLIFR networks have mixed performance when compared to vanilla recurrent neural networks but appear to be more robust to random silencing. When we explore these performance gains further, we find that both the ability to learn heterogeneity and the presence of after-spike currents contribute. Our work both demonstrates the computational robustness of neuronal complexity and diversity in networks and demonstrates a feasible method of training such models using exact gradients.

neuroscience↗