Neuromodulators enable overlapping synaptic memory regimes and nonlinear transition dynamics in recurrent neural networks

Mood, arousal, and other internal neural states can drastically alter behavior, even in identical external circumstances - the proverbial glass half full or empty. Neuromodulators are critical in controlling these internal neural states, and aberrations in neuromodulatory processes are linked to various neuropsychiatric disorders. To study how neuromodulators influence neural behavior, we modeled neuromodulation as a multiplicative factor acting on synaptic transmission between neurons in a recurrent neural network. We found this simple mechanism could vastly increase the computational capability and flexibility of a neural network by enabling overlapping storage of synaptic memories able to drive diverse, even diametrically opposed, behaviors. We analyzed how local or cell-type specific neuromodulation changes network activity to support such behaviors and reproduced experimental findings of Drosophila starvation behavior. We revealed that circuits have idiosyncratic, non-linear dose-response properties that can be different for chemical versus electrical modulation. Our findings help explain how neuromodulation "unlocks" specific behaviors with important implications for neuropsychiatric therapeutics.

Download Full-text

Btbd11 is an inhibitory interneuron specific synaptic scaffolding protein that supports excitatory synapse structure and function

10.1101/2021.11.01.466782 ◽

2021 ◽

Author(s):

Alexei M. Bygrave ◽

Ayesha Sengupta ◽

Ella P. Jackert ◽

Mehroz Ahmed ◽

Beloved Adenuga ◽

...

Keyword(s):

Inhibitory Interneuron ◽

Nmda Receptor Antagonist ◽

Specific Protein ◽

Network Activity ◽

Scaffolding Protein ◽

Inhibitory Interneurons ◽

Cell Type ◽

Cell Type Specific

Synapses in the brain exhibit cell–type–specific differences in basal synaptic transmission and plasticity. Here, we evaluated cell–type–specific differences in the composition of glutamatergic synapses, identifying Btbd11, as an inhibitory interneuron–specific synapse–enriched protein. Btbd11 is highly conserved across species and binds to core postsynaptic proteins including Psd–95. Intriguingly, we show that Btbd11 can undergo liquid–liquid phase separation when expressed with Psd–95, supporting the idea that the glutamatergic post synaptic density in synapses in inhibitory and excitatory neurons exist in a phase separated state. Knockout of Btbd11 from inhibitory interneurons decreased glutamatergic signaling onto parvalbumin–positive interneurons. Further, both in vitro and in vivo, we find that Btbd11 knockout disrupts network activity. At the behavioral level, Btbd11 knockout from interneurons sensitizes mice to pharmacologically induced hyperactivity following NMDA receptor antagonist challenge. Our findings identify a cell–type–specific protein that supports glutamatergic synapse function in inhibitory interneurons–with implication for circuit function and animal behavior.

Download Full-text

SeqEnhDL: sequence-based classification of cell type-specific enhancers using deep learning models

10.1101/2020.05.13.093997 ◽

2020 ◽

Author(s):

Yupeng Wang ◽

Rosario B. Jaime-Lara ◽

Abhrarup Roy ◽

Ying Sun ◽

Xinyue Liu ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Dna Sequences ◽

Cell Types ◽

Learning Models ◽

Cell Type ◽

Coding Sequences ◽

Sequence Features ◽

Cell Type Specific ◽

Different Cell Types

AbstractWe propose SeqEnhDL, a deep learning framework for classifying cell type-specific enhancers based on sequence features. DNA sequences of “strong enhancer” chromatin states in nine cell types from the ENCODE project were retrieved to build and test enhancer classifiers. For any DNA sequence, sequential k-mer (k=5, 7, 9 and 11) fold changes relative to randomly selected non-coding sequences were used as features for deep learning models. Three deep learning models were implemented, including multi-layer perceptron (MLP), Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN). All models in SeqEnhDL outperform state-of-the-art enhancer classifiers including gkm-SVM and DanQ, with regard to distinguishing cell type-specific enhancers from randomly selected non-coding sequences. Moreover, SeqEnhDL is able to directly discriminate enhancers from different cell types, which has not been achieved by other enhancer classifiers. Our analysis suggests that both enhancers and their tissue-specificity can be accurately identified according to their sequence features. SeqEnhDL is publicly available at https://github.com/wyp1125/SeqEnhDL.

Download Full-text

SeqEnhDL: sequence-based classification of cell type-specific enhancers using deep learning models

10.21203/rs.3.rs-94396/v1 ◽

2020 ◽

Author(s):

Yupeng Wang ◽

Rosario Jaime-Lara ◽

Abhrarup Roy ◽

Ying Sun ◽

Xinyue Liu ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Cell Types ◽

Regulatory Elements ◽

Learning Models ◽

Cell Type ◽

Coding Sequences ◽

Sequence Features ◽

A Genome ◽

Cell Type Specific

Abstract ObjectiveComputational identification of cell type-specific regulatory elements on a genome-wide scale is very challenging.ResultsWe propose SeqEnhDL, a deep learning framework for classifying cell type-specific enhancers based on sequence features. DNA sequences of “strong enhancer” chromatin states in nine cell types from the ENCODE project were retrieved to build and test enhancer classifiers. For any DNA sequence, sequential k-mer (k=5, 7, 9 and 11) fold changes relative to randomly selected non-coding sequences were used as features for deep learning models. Three deep learning models were implemented, including multi-layer perceptron (MLP), Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN). All models in SeqEnhDL outperform state-of-the-art enhancer classifiers including gkm-SVM and DanQ, with regard to distinguishing cell type-specific enhancers from randomly selected non-coding sequences. Moreover, SeqEnhDL is able to directly discriminate enhancers from different cell types, which has not been achieved by other enhancer classifiers. Our analysis suggests that both enhancers and their tissue-specificity can be accurately identified according to their sequence features. SeqEnhDL is publicly available at https://github.com/wyp1125/SeqEnhDL.

Download Full-text

Cell-type–specific neuromodulation guides synaptic credit assignment in a spiking neural network

Proceedings of the National Academy of Sciences ◽

10.1073/pnas.2111821118 ◽

2021 ◽

Vol 118 (51) ◽

pp. e2111821118

Author(s):

Yuhan Helena Liu ◽

Stephen Smith ◽

Stefan Mihalas ◽

Eric Shea-Brown ◽

Uygar Sümbül

Keyword(s):

Neural Network ◽

Normative Theory ◽

Synaptic Connections ◽

Neuron Type ◽

Cell Type ◽

Credit Assignment ◽

Neural Network Learning ◽

Network Learning ◽

Cell Type Specific ◽

Synaptic Learning

Brains learn tasks via experience-driven differential adjustment of their myriad individual synaptic connections, but the mechanisms that target appropriate adjustment to particular connections remain deeply enigmatic. While Hebbian synaptic plasticity, synaptic eligibility traces, and top-down feedback signals surely contribute to solving this synaptic credit-assignment problem, alone, they appear to be insufficient. Inspired by new genetic perspectives on neuronal signaling architectures, here, we present a normative theory for synaptic learning, where we predict that neurons communicate their contribution to the learning outcome to nearby neurons via cell-type–specific local neuromodulation. Computational tests suggest that neuron-type diversity and neuron-type–specific local neuromodulation may be critical pieces of the biological credit-assignment puzzle. They also suggest algorithms for improved artificial neural network learning efficiency.

Download Full-text

An interpretable bimodal neural network characterizes the sequence and preexisting chromatin predictors of induced TF binding

10.1101/672790 ◽

2019 ◽

Cited By ~ 3

Author(s):

Divyanshi Srivastava ◽

Begüm Aydin ◽

Esteban O. Mazzoni ◽

Shaun Mahony

Keyword(s):

Neural Network ◽

Dna Binding ◽

Dna Sequences ◽

Binding Specificity ◽

Chromatin Accessibility ◽

Cell Type ◽

Genome Wide ◽

Binding Preference ◽

Cell Type Specific ◽

Insight Into

AbstractTranscription factor (TF) binding specificity is determined via a complex interplay between the TF’s DNA binding preference and cell type-specific chromatin environments. The chromatin features that correlate with TF binding in a given cell type have been well characterized. For instance, the binding sites for a majority of TFs display concurrent chromatin accessibility. However, concurrent chromatin features reflect the binding activities of the TF itself, and thus provide limited insight into how genome-wide TF-DNA binding patterns became established in the first place. To understand the determinants of TF binding specificity, we therefore need to examine how newly activated TFs interact with sequence and preexisting chromatin landscapes.Here, we investigate the sequence and preexisting chromatin predictors of TF-DNA binding by examining the genome-wide occupancy of TFs that have been induced in well-characterized chromatin environments. We develop Bichrom, a bimodal neural network that jointly models sequence and preexisting chromatin data to interpret the genome-wide binding patterns of induced TFs. We find that the preexisting chromatin landscape is a differential global predictor of TF-DNA binding; incorporating preexisting chromatin features improves our ability to explain the binding specificity of some TFs substantially, but not others. Furthermore, by analyzing site-level predictors, we show that TF binding in previously inaccessible chromatin tends to correspond to the presence of more favorable cognate DNA sequences. Bichrom thus provides a framework for modeling, interpreting, and visualizing the joint sequence and chromatin landscapes that determine TF-DNA binding dynamics.

Download Full-text

SeqEnhDL: sequence-based classification of cell type-specific enhancers using deep learning models

BMC Research Notes ◽

10.1186/s13104-021-05518-7 ◽

2021 ◽

Vol 14 (1) ◽

Author(s):

Yupeng Wang ◽

Rosario B. Jaime-Lara ◽

Abhrarup Roy ◽

Ying Sun ◽

Xinyue Liu ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Cell Types ◽

Regulatory Elements ◽

Learning Models ◽

Cell Type ◽

Coding Sequences ◽

Sequence Features ◽

A Genome ◽

Cell Type Specific

Abstract Objective To address the challenge of computational identification of cell type-specific regulatory elements on a genome-wide scale. Results We propose SeqEnhDL, a deep learning framework for classifying cell type-specific enhancers based on sequence features. DNA sequences of “strong enhancer” chromatin states in nine cell types from the ENCODE project were retrieved to build and test enhancer classifiers. For any DNA sequence, positional k-mer (k = 5, 7, 9 and 11) fold changes relative to randomly selected non-coding sequences across each nucleotide position were used as features for deep learning models. Three deep learning models were implemented, including multi-layer perceptron (MLP), Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN). All models in SeqEnhDL outperform state-of-the-art enhancer classifiers (including gkm-SVM and DanQ) in distinguishing cell type-specific enhancers from randomly selected non-coding sequences. Moreover, SeqEnhDL can directly discriminate enhancers from different cell types, which has not been achieved by other enhancer classifiers. Our analysis suggests that both enhancers and their tissue-specificity can be accurately identified based on their sequence features. SeqEnhDL is publicly available at https://github.com/wyp1125/SeqEnhDL.

Download Full-text

Determining Burst Firing Time Distributions from Multiple Spike Trains

Neural Computation ◽

10.1162/neco.2008.07-07-571 ◽

2009 ◽

Vol 21 (4) ◽

pp. 973-990 ◽

Cited By ~ 3

Author(s):

Luis F. Lago-Fernández ◽

Attila Szücs ◽

Pablo Varona

Keyword(s):

Temporal Structure ◽

Alignment Algorithm ◽

Firing Time ◽

Cell Type ◽

Firing Patterns ◽

Bursting Neurons ◽

Cell Type Specific ◽

Experimental Findings ◽

General Tool

Recent experimental findings have shown the presence of robust and cell-type-specific intraburst firing patterns in bursting neurons. We address the problem of characterizing these patterns under the assumption that the bursts exhibit well-defined firing time distributions. We propose a method for estimating these distributions based on a burst alignment algorithm that minimizes the overlap among the firing time distributions of the different spikes within the burst. This method provides a good approximation to the burst's intrinsic temporal structure as a set of firing time distributions. In addition, the method allows labeling the spikes in any particular burst, establishing a correspondence between each spike and the distribution that best explains it, and identifying missing spikes. Our results on both simulated and experimental data from the lobster stomatogastric ganglion show that the proposed method provides a reliable characterization of the intraburst firing patterns and avoids the errors derived from missing spikes. This method can also be applied to nonbursting neurons as a general tool for the study and the interpretation of firing time distributions as part of a temporal neural code.

Download Full-text

An alternative splicing switch shapes neurexin repertoires in principal neurons versus interneurons in the mouse hippocampus

eLife ◽

10.7554/elife.22757 ◽

2016 ◽

Vol 5 ◽

Cited By ~ 30

Author(s):

Thi-Minh Nguyen ◽

Dietmar Schreiner ◽

Le Xiao ◽

Lisa Traunmüller ◽

Caroline Bornmann ◽

...

Keyword(s):

Alternative Splicing ◽

Rna Binding ◽

Network Activity ◽

Cell Type ◽

Splice Isoforms ◽

Mouse Hippocampus ◽

Specific Alternative ◽

Cell Type Specific ◽

Splice Isoform ◽

Functional Features

The unique anatomical and functional features of principal and interneuron populations are critical for the appropriate function of neuronal circuits. Cell type-specific properties are encoded by selective gene expression programs that shape molecular repertoires and synaptic protein complexes. However, the nature of such programs, particularly for post-transcriptional regulation at the level of alternative splicing is only beginning to emerge. We here demonstrate that transcripts encoding the synaptic adhesion molecules neurexin-1,2,3 are commonly expressed in principal cells and interneurons of the mouse hippocampus but undergo highly differential, cell type-specific alternative splicing. Principal cell-specific neurexin splice isoforms depend on the RNA-binding protein Slm2. By contrast, most parvalbumin-positive (PV+) interneurons lack Slm2, express a different neurexin splice isoform and co-express the corresponding splice isoform-specific neurexin ligand Cbln4. Conditional ablation of Nrxn alternative splice insertions selectively in PV+ cells results in elevated hippocampal network activity and impairment in a learning task. Thus, PV-cell-specific alternative splicing of neurexins is critical for neuronal circuit function

Download Full-text

The centrosome neither persistently leads migration nor determines the site of axonogenesis in migrating neurons in vivo

The Journal of Cell Biology ◽

10.1083/jcb.201004154 ◽

2010 ◽

Vol 191 (4) ◽

pp. 875-890 ◽

Cited By ~ 94

Author(s):

Martin Distel ◽

Jennifer C. Hocking ◽

Katrin Volkmann ◽

Reinhard W. Köster

Keyword(s):

Gene Expression ◽

Neuronal Migration ◽

Cell Type ◽

Vertebrate Brain ◽

Cell Type Specific ◽

Positive Correlations ◽

Experimental Findings ◽

Rhombic Lip

The position of the centrosome ahead of the nucleus has been considered crucial for coordinating neuronal migration in most developmental situations. The proximity of the centrosome has also been correlated with the site of axonogenesis in certain differentiating neurons. Despite these positive correlations, accumulating experimental findings appear to negate a universal role of the centrosome in determining where an axon forms, or in leading the migration of neurons. To further examine this controversy in an in vivo setting, we have generated cell type–specific multi-cistronic gene expression to monitor subcellular dynamics in the developing zebrafish cerebellum. We show that migration of rhombic lip–derived neurons is characterized by a centrosome that does not persistently lead the nucleus, but which is instead regularly overtaken by the nucleus. In addition, axonogenesis is initiated during the onset of neuronal migration and occurs independently of centrosome proximity. These in vivo data reveal a new temporal orchestration of organelle dynamics and provide important insights into the variation in intracellular processes during vertebrate brain differentiation.

Download Full-text