A subset of mouse hippocampus CA1 pyramidal neurons learns sparse synaptic input patterns. DOI Creative Commons
Anzal K. Shahul, Upinder S. Bhalla

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Дек. 9, 2024

Abstract Synaptic plasticity in the hippocampus is fundamental to learning and memory, yet few studies examine how pattern occurs across multiple synapses. Such cross-synapse emergent properties of discrimination generalisation, which depend on assumptions about independence linearity summation. We used sparse optogenetic spatio-temporal ‘pattern stimulation CA3 coupled with postsynaptic depolarization elicit CA1 pyramidal neurons, found that ‘trained’ patterns were selectively strengthened, but only a subset cells. Increased resting membrane potential background mini-EPSP rates predictive learner Summation following became more linear learners compared non-learners, consistent observed elevated post-stimulus hyperpolarization non-learner Thus our exploration biologically plausible activity supports pattern-selective learning, heterogeneous manner modulated by both cell-intrinsic network features.

Язык: Английский

A connectome manipulation framework for the systematic and reproducible study of structure function relationships through simulations DOI Creative Commons
Christoph Pokorny, Omar Awile, James B. Isbister

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Май 26, 2024

ABSTRACT Synaptic connectivity at the neuronal level is characterized by highly non-random features. Hypotheses about their role can be developed correlating structural metrics to functional But prove causation, manipula- tions of would have studied. However, fine-grained scale which trends are expressed makes this approach challenging pursue experimentally. Simulations networks provide an alternative route study arbitrarily complex manipulations in morphologically and biophysically detailed models. Here, we present Connectome-Manipulator, a Python framework for rapid connectome large- network models SONATA format. In addition creating or manipulating model, it provides tools fit parameters stochastic against existing connectomes. This enables replacement any with equivalent connectomes different levels complexity, transplantation features from one another, systematic study. We employed model rat somatosensory cortex two exemplary use cases: transplanting interneuron electron microscopy data simplified excitatory connectivity. ran series simulations found diverse shifts activity individual neuron populations causally linked these manipulations.

Язык: Английский

Процитировано

5
On the validity of electric brain signal predictions based on population firing rates DOI Creative Commons
Torbjørn V. Ness, Tom Tetzlaff, Gaute T. Einevoll

и другие.

PLoS Computational Biology, Год журнала: 2025, Номер 21(4), С. e1012303 - e1012303

Опубликована: Апрель 14, 2025

Neural activity at the population level is commonly studied experimentally through measurements of electric brain signals like local field potentials (LFPs), or electroencephalography (EEG) signals. To allow for comparison between observed and simulated neural it therefore important that simulations can accurately predict these Simulations often rely on point-neuron network models firing-rate models. While simplified representations are computationally efficient, they lack explicit spatial information needed calculating LFP/EEG Different heuristic approaches have been suggested overcoming this limitation, but accuracy has not fully assessed. One such approach, so-called kernel method, previously applied with promising results additional advantage being well-grounded in biophysics underlying signal generation. It based rate-to-LFP/EEG kernels each synaptic pathway a model, after which be obtained directly from firing rates. This amounts to massive reduction computational effort because calculated instead neuron. Here, we investigate how when method expected work, present theoretical framework predicting its accuracy. We show relative error predictions function single-cell heterogeneity spike-train correlations. Finally, demonstrate most accurate contributions also dominating signals: spatially clustered correlated input large populations pyramidal cells. thereby further establish as approach large-scale simulations.

Язык: Английский

Процитировано

0
A connectome manipulation framework for the systematic and reproducible study of structure–function relationships through simulations DOI Creative Commons
Christoph Pokorny, Omar Awile, James B. Isbister

и другие.

Network Neuroscience, Год журнала: 2024, Номер 9(1), С. 207 - 236

Опубликована: Дек. 2, 2024

Synaptic connectivity at the neuronal level is characterized by highly nonrandom features. Hypotheses about their role can be developed correlating structural metrics to functional But, prove causation, manipulations of would have studied. However, fine-grained scale which trends are expressed makes this approach challenging pursue experimentally. Simulations networks provide an alternative route study arbitrarily complex in morphologically and biophysically detailed models. Here, we present Connectome-Manipulator, a Python framework for rapid connectome large-scale network models Scalable Open Network Architecture TemplAte (SONATA) format. In addition creating or manipulating model, it provides tools fit parameters stochastic against existing connectomes. This enables replacement any with equivalent connectomes different levels complexity, transplantation features from one another, systematic study. We employed model rat somatosensory cortex two exemplary use cases: transplanting interneuron electron microscopy data simplified excitatory connectivity. ran series simulations found diverse shifts activity individual neuron populations causally linked these manipulations.

Язык: Английский

Процитировано

3
Gather your neurons and model together: Community times ahead DOI Creative Commons
Maria Diamantaki, Athanasia Papoutsi

PLoS Biology, Год журнала: 2024, Номер 22(11), С. e3002839 - e3002839

Опубликована: Ноя. 6, 2024

Bottom-up, data-driven, large-scale models provide a mechanistic understanding of neuronal functions. A new study in PLOS Biology builds biologically realistic model the rodent CA1 region that aims to become an accessible tool for whole hippocampal community.

Язык: Английский

Процитировано

0
A subset of mouse hippocampus CA1 pyramidal neurons learns sparse synaptic input patterns. DOI Creative Commons
Anzal K. Shahul, Upinder S. Bhalla

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Дек. 9, 2024

Abstract Synaptic plasticity in the hippocampus is fundamental to learning and memory, yet few studies examine how pattern occurs across multiple synapses. Such cross-synapse emergent properties of discrimination generalisation, which depend on assumptions about independence linearity summation. We used sparse optogenetic spatio-temporal ‘pattern stimulation CA3 coupled with postsynaptic depolarization elicit CA1 pyramidal neurons, found that ‘trained’ patterns were selectively strengthened, but only a subset cells. Increased resting membrane potential background mini-EPSP rates predictive learner Summation following became more linear learners compared non-learners, consistent observed elevated post-stimulus hyperpolarization non-learner Thus our exploration biologically plausible activity supports pattern-selective learning, heterogeneous manner modulated by both cell-intrinsic network features.

Язык: Английский

Процитировано

0