Structured stabilization in recurrent neural circuits through inhibitory synaptic plasticity DOI Creative Commons
Dylan Festa,

Claudia Cusseddu,

Julijana Gjorgjieva

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 12, 2024

Inhibitory interneurons play a dual role in recurrently connected biological circuits: they regulate global neural activity to prevent runaway excitation, and contribute complex computations. While the first can be achieved through unstructured connections tuned for homeostatic rate stabilization, computational tasks often require structured excitatory-inhibitory (E/I) connectivity. Here, we consider broad class of pairwise inhibitory spike-timing dependent plasticity (iSTDP) rules, demonstrating how synapses self-organize both stabilize excitation generate functionally relevant connectivity structures — process call “structured stabilization”. We show that E/I circuit motifs large spiking recurrent networks choice iSTDP rule lead either mutually pairs, or lateral inhibition, where an neuron connects excitatory does not directly connect back it. In one-dimensional ring network, if two populations follow these distinct forms iSTDP, effective within population self-organizes into Mexican-hat-like profile with influence center away from center. This leads emergent dynamical properties such as surround suppression modular spontaneous activity. Our theoretical work introduces family rules retains applicability simplicity spike-timing-based plasticity, while promoting structured, self-organized stabilization. These findings highlight rich interplay between structure, neuronal dynamics, offering framework understanding shapes network function.

Language: Английский

Efficiency and reliability in biological neural network architectures DOI Creative Commons
Daniela Egas Santander, Christoph Pokorny, András Ecker

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: March 17, 2024

Abstract Simplified models of neural networks have demonstrated the importance establishing a reasonable tradeoff between memory capacity and fault-tolerance in cortical coding schemes. The intensity is mediated by level neuronal variability. Indeed, increased redundancy activity enhances robustness code at cost its efficiency. We hypothesized that heterogeneous architecture biological provides substrate to regulate this tradeoff, thereby allowing different subpopulations same network optimize for objectives. To distinguish subpopulations, we developed metric based on mathematical theory simplicial complexes captures complexity their connectivity, contrasting higher-order structure random control. confirm relevance our analyzed several openly available connectomes, revealing they all exhibited wider distributions across than relevant controls. Using biologically detailed model an electron microscopic data set connectivity with co-registered functional data, showed low exhibit efficient activity. Conversely, high play supporting role boosting reliability as whole, softening robustness-efficiency tradeoff. Crucially, found both types can do coexist within single connectome networks, due heterogeneity connectivity. Our work thus suggests avenue resolving seemingly paradoxical previous results assume homogeneous

Language: Английский

Citations

7
Modeling and Simulation of Neocortical Micro- and Mesocircuitry. Part I: Anatomy DOI Creative Commons
Michael W. Reimann, Sirio Bolaños‐Puchet, Jean-Denis Courcol

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2022, Volume and Issue: unknown

Published: Aug. 15, 2022

Abstract The function of the neocortex is fundamentally determined by its repeating microcircuit motif, but also rich, interregional connectivity. We present a data-driven computational model anatomy non-barrel primary somatosensory cortex juvenile rat, integrating whole-brain scale data while providing cellular and subcellular specificity. consists 4.2 million morphologically detailed neurons, placed in digital brain atlas. They are connected 14.2 billion synapses, comprising local, mid-range extrinsic delineated limits determining connectivity from neuron morphology placement, finding that it reproduces targeting Sst+ requires additional specificity to reproduce PV+ VIP+ interneurons. Globally, was characterized local clusters tied together through hub neurons layer 5, demonstrating how interegional complicit, inseparable networks. suitable for simulation-based studies, 211,712 subvolume made openly available community.

Language: Английский

Citations

8
Assemblies, synapse clustering and network topology interact with plasticity to explain structure-function relationships of the cortical connectome DOI Creative Commons
András Ecker, Daniela Egas Santander, Marwan Abdellah

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 7, 2023

Synaptic plasticity underlies the brain's ability to learn and adapt. While experiments in brain slices have revealed mechanisms protocols for induction of between pairs neurons, how these synaptic changes are coordinated biological neuronal networks ensure emergence learning remains poorly understood. Simulation modeling emerged as important tools study plastic networks, but yet achieve a scale that incorporates realistic network structure, active dendrites, multi-synapse interactions, key determinants plasticity. To rise this challenge, we endowed an existing large-scale cortical model, incorporating data-constrained dendritic processing multi-synaptic connections, with calcium-based model functional captures diversity excitatory connections extrapolated vivo-like conditions. This allowed us dendrites structure interact shape stimulus representations at microcircuit level. In our exploratory simulations, acted sparsely specifically, firing rates weight distributions remained stable without additional homeostatic mechanisms. At circuit level, found was driven by co-firing stimulus-evoked assemblies, spatial clustering synapses on topology connectivity. As result changes, became more reliable stimulus-specific responses. We confirmed testable predictions MICrONS datasets, openly available electron microscopic reconstruction large volume tissue. Our results quantify architecture higher-order microcircuits play central role provide foundation elucidating their learning.

Language: Английский

Citations

3
Structured stabilization in recurrent neural circuits through inhibitory synaptic plasticity DOI Creative Commons
Dylan Festa,

Claudia Cusseddu,

Julijana Gjorgjieva

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 12, 2024

Inhibitory interneurons play a dual role in recurrently connected biological circuits: they regulate global neural activity to prevent runaway excitation, and contribute complex computations. While the first can be achieved through unstructured connections tuned for homeostatic rate stabilization, computational tasks often require structured excitatory-inhibitory (E/I) connectivity. Here, we consider broad class of pairwise inhibitory spike-timing dependent plasticity (iSTDP) rules, demonstrating how synapses self-organize both stabilize excitation generate functionally relevant connectivity structures — process call “structured stabilization”. We show that E/I circuit motifs large spiking recurrent networks choice iSTDP rule lead either mutually pairs, or lateral inhibition, where an neuron connects excitatory does not directly connect back it. In one-dimensional ring network, if two populations follow these distinct forms iSTDP, effective within population self-organizes into Mexican-hat-like profile with influence center away from center. This leads emergent dynamical properties such as surround suppression modular spontaneous activity. Our theoretical work introduces family rules retains applicability simplicity spike-timing-based plasticity, while promoting structured, self-organized stabilization. These findings highlight rich interplay between structure, neuronal dynamics, offering framework understanding shapes network function.

Language: Английский

Citations

0