CoChat
Sign Up
PreprintCited by 16Neural dynamics and brain functionNeural Networks and ApplicationsFunctional Brain Connectivity Studies

The DOI matches its Crossref record.

Prediction of neural activity in connectome-constrained recurrent networks

Open access
Beiran, Manuel, Litwin-Kumar, Ashok·2024·Nature Neuroscience
DOI 10.1101/2024.02.22.581667PMC PMC11160579

Preprint — published as 10.1038/s41593-025-02080-4

Open this paper in CoChat — free accountOpen PDF

Abstract

We develop a theory of connectome-constrained neural networks in which a student network is trained to reproduce the activity of a ground-truth teacher, representing a neural system for which a connectome is available. Unlike standard paradigms with unconstrained connectivity, here the two networks have the same connectivity but different biophysical parameters, reflecting uncertainty in neuronal and synaptic properties. We find that a connectome is often insufficient to constrain the dynamics of networks that perform a specific task, illustrating the difficulty of inferring function from connectivity alone. However, recordings from a small subset of neurons can remove this degeneracy, producing dynamics in the student that agree with the teacher. Our theory can also prioritize which neurons to record from to most efficiently predict unmeasured network activity. Our analysis shows that the solution spaces of connectome-constrained and unconstrained models are qualitatively different and provides a framework to determine when such models yield consistent dynamics.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/2024.02.22.581667

Cites

Statistics

Figures

Figure 1: Task-constrained networks with the same connectivity. A A teacher RNN is trained to generate two different readout responses based on two different output pulses. B Properties of the teacher RNN. The teacher RNN has heterogeneous single neuron parameters (gains and biases of single neuron activation functions, left), sparse connectivity with connection probability p=0.5 (right), and neurons connect through either excitatory (red) or inhibitory (blue) synapses. C Student networks with the same connectivity as the teacher are trained to produce the teacher’s output. Error in the readout (training loss, MSE) as a function of training epochs. Each colored line corresponds to a different student network. D Error (mismatch in neural activity) between teacher and student RNNs. Gray line, for reference, corresponds to the average error in activity when the student reproduces the teacher’s activity but with shuffled neuron identities. E Error in gains and biases vs. training epochs. F Readout of teacher and student networks after training, for the two trial types (top and bottom). Teacher and student networks both solve the task. G Neural activity of an example excitatory (left) and inhibitory (right) neuron. Teacher and student neurons exhibit different single-neuron dynamics.
Figure 1: Task-constrained networks with the same connectivity. A A teacher RNN is trained to generate two different readout responses based on two different output pulses. B Properties of the teacher RNN. The teacher RNN has heterogeneous single neuron parameters (gains and biases of single neuron activation functions, left), sparse connectivity with connection probability p=0.5 (right), and neurons connect through either excitatory (red) or inhibitory (blue) synapses. C Student networks with the same connectivity as the teacher are trained to produce the teacher’s output. Error in the readout (training loss, MSE) as a function of training epochs. Each colored line corresponds to a different student network. D Error (mismatch in neural activity) between teacher and student RNNs. Gray line, for reference, corresponds to the average error in activity when the student reproduces the teacher’s activity but with shuffled neuron identities. E Error in gains and biases vs. training epochs. F Readout of teacher and student networks after training, for the two trial types (top and bottom). Teacher and student networks both solve the task. G Neural activity of an example excitatory (left) and inhibitory (right) neuron. Teacher and student neurons exhibit different single-neuron dynamics.CC BY-NC-ND · PubMed Central
Figure 2: Transition in the prediction of the activity of unrecorded neurons when training single-neuron parameters. A The student RNN is trained to mimic the activity of M recorded neurons in a teacher RNN. B Error in recorded activity (loss) vs. training epochs for students with trained single neuron parameters (left) and students with trained connectivity (right). Lines correspond to different numbers of recorded neurons M and show mean and SEM over 10 random seeds. All students successfully reproduce the recorded activity of the teacher after learning. C Left: Error between teacher and students in the activity of the N−M unrecorded neurons vs. training epochs. Right: Error in unrecorded neuronal activity after training, as a function of number of recorded neurons M. Error is substantially reduced when recording from M>30 neurons. D Analogous to Panel D, but training synaptic weights instead. The error in the activity of unrecorded neurons remains high across values of M. E Error in gains and biases vs. training epochs. Left: Parameters of recorded neurons. Right: Parameters of unrecorded neurons. F Analogous to panel E for connectivity weights between recorded neurons (left) and between unrecorded neurons (right).
Figure 2: Transition in the prediction of the activity of unrecorded neurons when training single-neuron parameters. A The student RNN is trained to mimic the activity of M recorded neurons in a teacher RNN. B Error in recorded activity (loss) vs. training epochs for students with trained single neuron parameters (left) and students with trained connectivity (right). Lines correspond to different numbers of recorded neurons M and show mean and SEM over 10 random seeds. All students successfully reproduce the recorded activity of the teacher after learning. C Left: Error between teacher and students in the activity of the N−M unrecorded neurons vs. training epochs. Right: Error in unrecorded neuronal activity after training, as a function of number of recorded neurons M. Error is substantially reduced when recording from M>30 neurons. D Analogous to Panel D, but training synaptic weights instead. The error in the activity of unrecorded neurons remains high across values of M. E Error in gains and biases vs. training epochs. Left: Parameters of recorded neurons. Right: Parameters of unrecorded neurons. F Analogous to panel E for connectivity weights between recorded neurons (left) and between unrecorded neurons (right).CC BY-NC-ND · PubMed Central
Figure 3: Prediction of unrecorded neurons’ activity depends on the geometry of dynamics, not network size. A Set of teacher RNNs with variable network size N but fixed rank of the connectivity matrix (see Methods). B Teachers of different size produce the same low-dimensional dynamics. Left: Dynamics projected on the top two principal components (PCs). All RNNs generate a limit cycle largely constrained to a 2D linear subspace. Right: Variance screeplot. C Error in the activity of unrecorded neurons, after training. We measured the correlation distance between activity in the teacher and student. Empirical average and SEM for each network size (10 networks per condition). Gray line sets represents the baseline error corresponding to shuffled neuronal identities. D Set of teacher RNNs with variable network size N with random (full-rank) connectivity. E Left: Networks generate high-dimensional chaotic dynamics. Sample activity of four units for networks of different sizes. Right: Variance scree plot. Larger networks generate higher dimensional dynamics. F Error in the activity of unrecorded neurons after training. Larger networks require recording from a larger number of neurons M to predict unrecorded activity. Inset: Number of neurons M^(*) needed to predict unrecorded activity above a certain threshold (set to 0.2; dotted line), as a function of network size.
Figure 3: Prediction of unrecorded neurons’ activity depends on the geometry of dynamics, not network size. A Set of teacher RNNs with variable network size N but fixed rank of the connectivity matrix (see Methods). B Teachers of different size produce the same low-dimensional dynamics. Left: Dynamics projected on the top two principal components (PCs). All RNNs generate a limit cycle largely constrained to a 2D linear subspace. Right: Variance screeplot. C Error in the activity of unrecorded neurons, after training. We measured the correlation distance between activity in the teacher and student. Empirical average and SEM for each network size (10 networks per condition). Gray line sets represents the baseline error corresponding to shuffled neuronal identities. D Set of teacher RNNs with variable network size N with random (full-rank) connectivity. E Left: Networks generate high-dimensional chaotic dynamics. Sample activity of four units for networks of different sizes. Right: Variance scree plot. Larger networks generate higher dimensional dynamics. F Error in the activity of unrecorded neurons after training. Larger networks require recording from a larger number of neurons M to predict unrecorded activity. Inset: Number of neurons M^(*) needed to predict unrecorded activity above a certain threshold (set to 0.2; dotted line), as a function of network size.CC BY-NC-ND · PubMed Central
Figure 4: Model mismatch between teacher and student. A Mismatch in activation functions of teacher and student neurons. B The activation function is a smooth rectification but with different degrees of smoothness, parameterized by a parameter β. Teacher RNN from Fig. 2. C Errors in the activity of recorded (left) and unrecorded (right) neurons for different values of model mismatch between teacher and student. Within a large range of mismatch we observe a decrease in the error in unrecorded neurons when M>30. D Mismatch in the connectivity between teacher and student, mimicking errors in connectome reconstruction. E Eigenvalues of the teacher and student connectivity matrices, for different levels of connectivity mismatch. F Errors in the activity of recorded (left) and unrecorded (right) neurons for different levels of mismatch in the connectivity.
Figure 4: Model mismatch between teacher and student. A Mismatch in activation functions of teacher and student neurons. B The activation function is a smooth rectification but with different degrees of smoothness, parameterized by a parameter β. Teacher RNN from Fig. 2. C Errors in the activity of recorded (left) and unrecorded (right) neurons for different values of model mismatch between teacher and student. Within a large range of mismatch we observe a decrease in the error in unrecorded neurons when M>30. D Mismatch in the connectivity between teacher and student, mimicking errors in connectome reconstruction. E Eigenvalues of the teacher and student connectivity matrices, for different levels of connectivity mismatch. F Errors in the activity of recorded (left) and unrecorded (right) neurons for different levels of mismatch in the connectivity.CC BY-NC-ND · PubMed Central
Figure 5: Linear teacher-student model A Left: The activity of a neuron is a linear mapping A(J), which depends on the connectivity matrix J, of the single neuron parameters b. Right: Neuronal activation functions are linear with heterogeneous biases. B Singular values of the connectivity matrix, which is random and has rank D=60. C Errors in activity and biases as a function of the number of recorded neurons. D Single neuron biases evolve over training through gradient descent. Parameter modes are described as stiff or sloppy based on the effect of changes along each mode near the optimal solution. E Singular value decomposition of the mapping A determines stiff and sloppy parameter modes. Stiffer modes are learned more quickly. F Effective singular value decomposition when recording from a subset of M neurons. Inset shows the maximum angle between the M stiffest modes and the M sub-sampled parameter modes. G-H Evolution of errors in activity and biases for D>M=10 and D<M=160, for 10 different initializations of parameters. Error in biases is projected along one stiff (1st) and one sloppy (50th) parameter mode.
Figure 5: Linear teacher-student model A Left: The activity of a neuron is a linear mapping A(J), which depends on the connectivity matrix J, of the single neuron parameters b. Right: Neuronal activation functions are linear with heterogeneous biases. B Singular values of the connectivity matrix, which is random and has rank D=60. C Errors in activity and biases as a function of the number of recorded neurons. D Single neuron biases evolve over training through gradient descent. Parameter modes are described as stiff or sloppy based on the effect of changes along each mode near the optimal solution. E Singular value decomposition of the mapping A determines stiff and sloppy parameter modes. Stiffer modes are learned more quickly. F Effective singular value decomposition when recording from a subset of M neurons. Inset shows the maximum angle between the M stiffest modes and the M sub-sampled parameter modes. G-H Evolution of errors in activity and biases for D>M=10 and D<M=160, for 10 different initializations of parameters. Error in biases is projected along one stiff (1st) and one sloppy (50th) parameter mode.CC BY-NC-ND · PubMed Central
Figure 6: Loss landscape in nonlinear networks. A We study a rank-two RNN with two populations. Neurons in each population share the same gains and network statistics. B Dynamics of the target network. Left: Phase-space in the two-dimensional latent space. Right: Activity as a function of time for 20 sampled neurons. For illustrative purposes, neurons 1 and 2 are selected based on their alignment with the two latent variables. C The loss landscape of the full network depends only on the gains of each population, g1 and g2. White dot indicates the parameters of the teacher RNN. D Loss landscape when recording the activity of neuron 1 (left) or neuron 2 (right). Blue and red squares correspond respectively to solutions where the training loss is close to zero. E Target trajectory (black) and dynamics of the teacher RNN. Blue and red trajectories correspond to the solutions found in D. F Predicted activity for neurons 1 and 2 for the solutions found in D. Left: Error in the activity of the recorded neuron (neuron 1) is small, while error for the unrecorded neuron (neuron 2) is large. Right: Similar to left, but when neuron 2 is recorded and neuron 1 is unrecorded. G Full-rank non-linear RNN, same as in Fig. 1. H Average squared error in parameters projected on the different stiff and sloppy parameter modes. The stiff and sloppy dimensions are determined by approximating the full-sampled loss function around the teacher’s values (see Methods). Average over 10 realizations.
Figure 6: Loss landscape in nonlinear networks. A We study a rank-two RNN with two populations. Neurons in each population share the same gains and network statistics. B Dynamics of the target network. Left: Phase-space in the two-dimensional latent space. Right: Activity as a function of time for 20 sampled neurons. For illustrative purposes, neurons 1 and 2 are selected based on their alignment with the two latent variables. C The loss landscape of the full network depends only on the gains of each population, g1 and g2. White dot indicates the parameters of the teacher RNN. D Loss landscape when recording the activity of neuron 1 (left) or neuron 2 (right). Blue and red squares correspond respectively to solutions where the training loss is close to zero. E Target trajectory (black) and dynamics of the teacher RNN. Blue and red trajectories correspond to the solutions found in D. F Predicted activity for neurons 1 and 2 for the solutions found in D. Left: Error in the activity of the recorded neuron (neuron 1) is small, while error for the unrecorded neuron (neuron 2) is large. Right: Similar to left, but when neuron 2 is recorded and neuron 1 is unrecorded. G Full-rank non-linear RNN, same as in Fig. 1. H Average squared error in parameters projected on the different stiff and sloppy parameter modes. The stiff and sloppy dimensions are determined by approximating the full-sampled loss function around the teacher’s values (see Methods). Average over 10 realizations.CC BY-NC-ND · PubMed Central
Figure 7: Optimal selection of recorded neurons. A Recording from specific subsets of neurons (right) in the teacher RNN leads to different performance. B We linearized the mapping from changes in single neuron parameters to changes in neural activity. C Teacher RNN with linear single-neuron activation functions, unknown biases, and connectivity with rank D=60 (as in Fig. 5). D Error in activity of unrecorded neurons as a function of number of recorded neurons M. Lines correspond to theoretical prediction, dots to numerical simulation (mean ± SEM). We selected neurons following the estimated best ranking (red), 5 different random rankings (black), and the worst ranking (blue). E Error in recorded neurons for the same networks. F - H Analogous to C-E but for a nonlinear network. The teacher is the RNN from Fig. 2 with sparse E-I connectivity. Single-neuron parameters are both gains and biases. The linearization in the mapping from parameters to activity assumes homogeneous single-neuron parameters (see Methods).
Figure 7: Optimal selection of recorded neurons. A Recording from specific subsets of neurons (right) in the teacher RNN leads to different performance. B We linearized the mapping from changes in single neuron parameters to changes in neural activity. C Teacher RNN with linear single-neuron activation functions, unknown biases, and connectivity with rank D=60 (as in Fig. 5). D Error in activity of unrecorded neurons as a function of number of recorded neurons M. Lines correspond to theoretical prediction, dots to numerical simulation (mean ± SEM). We selected neurons following the estimated best ranking (red), 5 different random rankings (black), and the worst ranking (blue). E Error in recorded neurons for the same networks. F - H Analogous to C-E but for a nonlinear network. The teacher is the RNN from Fig. 2 with sparse E-I connectivity. Single-neuron parameters are both gains and biases. The linearization in the mapping from parameters to activity assumes homogeneous single-neuron parameters (see Methods).CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

No references could be resolved yet.

Research works printed but not linked: 51

  1. [1]Das A., Fiete I. R., Systematic errors in connectivity inferred from activity in strongly recurrent networks, Nature Neuroscience 2020 23:10 23 (10) (2020) 1286– 1296. 32895567
  2. [4]Liang T., Brinkman B. A. W., Statistically inferred neuronal connections in subsampled neural networks strongly correlate with spike train covariance, bioRxiv (2023) 2023.02.01.526673. 38755896
  3. [5]Dinc F., Shai A., Schnitzer M., Tanaka H., CORNN: Convex optimization of recurrent neural networks for rapid inference of neural dynamics (nov 2023). arXiv:2311.10200. · arXiv
  4. [6]White J. G., Southgate E., Thomson J. N., Brenner S., The Structure of the Nervous System of the Nematode Caenorhabditis elegans, Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 314 (1165) (1986) 1– 340. 22462104
  5. [7]Ohyama T., Schneider-Mizell C. M., Fetter R. D., Aleman J. V., Franconville R., Rivera-Alba M., Mensh B. D., Branson K. M., Simpson J. H., Truman J. W., Cardona A., Zlatic M., A multilevel multimodal circuit enhances action selection in Drosophila, Nature 520 (7549) (2015) 633– 639. 25896325
  6. [8]Zheng Z., Lauritzen J. S., Perlman E., Saalfeld S., Fetter R. D., Bock Correspondence D. D., A Complete Electron Microscopy Volume of the Brain of Adult Drosophila melanogaster, Cell 174 (2018) 730– 743.e22. 30033368 PMC6063995
  7. [9]Scheffer L. K., Xu C. S., Januszewski M., Lu Z., Takemura S. Y., Hayworth K. J., Huang G. B., Shinomiya K., Maitin-Shepard J., Berg S., Clements J., Hubbard P. M., Katz W. T., Umayam L., Zhao T., Ackerman D., Blakely T., Bogovic J., Dolafi T., Kainmueller D., Kawase T., Khairy K. A., Leavitt L., Li P. H., Lindsey L., Neubarth N., Olbris D. J., Otsuna H., Trautman E. T., Ito M., Bates A. S., Goldammer J., Wolff T., Svirskas R., Schlegel P., Neace E. R., Knecht C. J., Alvarado C. X., Bailey D. A., Ballinger S., Borycz J. A., Canino B. S., Cheatham N., Cook M., Dreher M., Duclos O., Eubanks B., Fairbanks K., Finley S., Forknall N., Francis A., Hopkins G. P., Joyce E. M., Kim S., Kirk N. A., Kovalyak J., Lauchie S. A., Lohff A., Maldonado C., Manley E. A., McLin S., Mooney C., Ndama M., Ogundeyi O., Okeoma N., Ordish C., Padilla N., Patrick C., Paterson T., Phillips E. E., Phillips E. M., Rampally N., Ribeiro C., Robertson M. K., Rymer J. T., Ryan S. M., Sammons M., Scott A. K., Scott A. L., Shinomiya A., Smith C., Smith K., Smith N. L., Sobeski M. A., Suleiman A., Swift J., Takemura S., Talebi I., Tarnogorska D., Tenshaw E., Tokhi T., Walsh J. J., Yang T., Horne J. A., Li F., Parekh R., Rivlin P. K., Jayaraman V., Costa M., Jefferis G. S., Ito K., Saalfeld S., George R., Meinertzhagen I. A., Rubin G. M., Hess H. F., Jain V., Plaza S. M., A connectome and analysis of the adult drosophila central brain, eLife 9 (2020) 1– 74. PMC7546738 32880371
  8. [10]Dorkenwald S., Matsliah A., Sterling A. R., Schlegel P., Yu S.-c., McKellar C. E., Lin A., Costa M., Eichler K., Yin Y., Silversmith W., Schneider-Mizell C., Jordan C. S., Brittain D., Halageri A., Kuehner K., Ogedengbe O., Morey R., Gager J., Kruk K., Perlman E., Yang R., Deutsch D., Bland D., Sorek M., Lu R., Macrina T., Lee K., Bae J. A., Mu S., Nehoran B., Mitchell E., Popovych S., Wu J., Jia Z., Castro M., Kemnitz N., Ih D., Bates A. S., Eckstein N., Funke J., Collman F., Bock D. D., Jefferis G. S., Seung H. S., Murthy M., Consortium F., Neuronal wiring diagram of an adult brain, bioRxiv (jul 2023). PMC11446842 39358518
  9. [11]Hildebrand D. G. C., Cicconet M., Torres R. M., Choi W., Quan T. M., Moon J., Wetzel A. W., Scott Champion A., Graham B. J., Randlett O., Plummer G. S., Portugues R., Bianco I. H., Saalfeld, Baden A. D., Lillaney K., Burns R., Vogelstein J. T., Schier A. F., Lee W. C. A., Jeong W. K., Lichtman J. W., Engert F., Whole-brain serial-section electron microscopy in larval zebrafish, Nature 545 (7654) (2017) 345– 349. 28489821 PMC5594570
  10. [12]Turner M. H., Mann K., Clandinin T. R., The connectome predicts resting-state functional connectivity across the Drosophila brain, Current biology 31 (11) (2021) 2386– 2394.e3. 33770490 PMC8519013
  11. [13]Randi F., Sharma A. K., Dvali S., Leifer A. M., Neural signal propagation atlas of Caenorhabditis elegans, Nature 2023 623:7986 623 (7986) (2023) 406– 414. PMC10632145 37914938
  12. [15]Pospisil D. A., Aragon M. J., Dorkenwald S., Matsliah A., Sterling A. R., Schlegel P., chieh Yu S., McKellar C. E., Costa M., Eichler K., Jefferis G. S., Murthy M., Pillow J. W., From connectome to effectome: learning the causal interaction map of the fly brain, bioRxiv (2024). arXiv: https://www.biorxiv.org/content/early/2024/02/09/2023.10.31.564922.full.pdf, doi: 10.1101/2023.10.31.564922. URL https://www.biorxiv.org/content/early/2024/02/09/2023.10.31.564922 PMC11446844 39358526 · DOI
  13. [16]Zarin A. A., Mark B., Cardona A., Litwin-Kumar A., Doe C. Q., A multilayer circuit architecture for the generation of distinct locomotor behaviors in Drosophila, eLife 8 (dec 2019). PMC6994239 31868582
  14. [17]Keller A. J., Dipoppa M., Roth M. M., Caudill M. S., Ingrosso A., Miller K. D., Scanziani M., A Disinhibitory Circuit for Contextual Modulation in Primary Visual Cortex, Neuron 108 (6) (2020) 1181– 1193.e8. 33301712 PMC7850578
  15. [18]Kohn J. R., Portes J. P., Christenson M. P., Abbott L. F., Behnia R., Flexible filtering by neural inputs supports motion computation across states and stimuli, Current biology 31 (23) (2021) 5249– 5260.e5. 34670114 PMC8725177
  16. [19]Lappalainen J. K., Tschopp F. D., Prakhya S., McGill M., Nern A., Shinomiya K., Takemura S.-y., Gruntman E., Macke J. H., Turaga S. C., Connectome-constrained deep mechanistic networks predict neural responses across the fly visual system at single-neuron resolution, bioRxiv (2023) 2023.03.11.532232.
  17. [20]Eckstein N., Bates A. S., Champion A., Du M., Yin Y., Schlegel P., Lu A. K.-Y., Rymer T., Finley-May S., Paterson T., Parekh R., Dorkenwald S., Matsliah A., Yu S.-C., McKellar C., Sterling A., Eichler K., Costa M., Seung S., Murthy M., Hartenstein V., Jefferis G. S., Funke J., Neurotransmitter Classification from Electron Microscopy Images at Synaptic Sites in Drosophila Melanogaster, bioRxiv (2023) 2020.06.12.148775. PMC11106717 38729112
  18. [21]Barnes C. L., Bonnéry D., Cardona A., Synaptic counts approximate synaptic contact area in Drosophila, PLOS ONE 17 (4) (2022) e0266064. 35377898 PMC8979427
  19. [22]Kasai H., Fukuda M., Watanabe S., Hayashi-Takagi A., Noguchi J., Structural dynamics of dendritic spines in memory and cognition, Trends in neurosciences 33 (3) (2010) 121– 129. 20138375
  20. [23]Bargmann C. I., Marder E., From the connectome to brain function, Nature methods 10 (6) (2013) 483– 490. 23866325
  21. [24]Marder E., Neuromodulation of neuronal circuits: back to the future, Neuron 76 (1) (2012) 1– 11. 23040802 PMC3482119
  22. [25]Gutierrez G. J., O’Leary T., Marder E., Multiple mechanisms switch an electrically coupled, synaptically inhibited neuron between competing rhythmic oscillators, Neuron 77 (5) (2013) 845– 858. 23473315 PMC3664401
  23. [26]Stroud J. P., Porter M. A., Hennequin G., Vogels T. P., Motor primitives in space and time via targeted gain modulation in cortical networks, Nature Neuroscience 2018 21:12 21 (12) (2018) 1774– 1783. PMC6276991 30482949
  24. [27]Ferguson K. A., Cardin J. A., Mechanisms underlying gain modulation in the cortex, Nature Reviews Neuroscience 2020 21:2 21 (2) (2020) 80– 92. PMC7408409 31911627
  25. [28]Connors B. W., Gutnick M. J., Intrinsic firing patterns of diverse neocortical neurons, Trends in neurosciences 13 (3) (1990) 99– 104. 1691879
  26. [29]Kubota Y., Hatada S., Kondo S., Karube F., Kawaguchi Y., Neocortical inhibitory terminals innervate dendritic spines targeted by thalamocortical afferents, The Journal of neuroscience: the official journal of the Society for Neuroscience 27 (5) (2007) 1139– 1150. 17267569 PMC6673192
  27. [30]Perich M. G., Arlt C., Soares S., Young M. E., Mosher C. P., Minxha J., Carter E., Rutishauser U., Rudebeck P. H., Harvey C. D., Rajan K., Inferring brain-wide interactions using data-constrained recurrent neural network models, bioRxiv (2021) 2020.12.18.423348.
  28. [31]Seung H. S., Sompolinsky H., Tishby N., Statistical mechanics of learning from examples, Physical Review A 45 (8) (1992) 6056. 9907706
  29. [32]Saad D., Solla S. A., Exact Solution for On-Line Learning in Multilayer Neural Networks, Physical Review Letters 74 (21) (1995) 4337. 10058475
  30. [33]Gao P., Trautmann E., Yu B., Santhanam G., Ryu S., Shenoy K., Ganguli S., A theory of multineuronal dimensionality, dynamics and measurement, bioRxiv (2017) 214262.
  31. [34]Kim C. M., Finkelstein A., Chow C. C., Svoboda K., Darshan R., Distributing task-related neural activity across a cortical network through task-independent connections, Nature Communications 2023 14:1 14 (1) (2023) 1– 21. PMC10195842 37202424
  32. [35]Russo A. A., Bittner S. R., Perkins S. M., Seely J. S., London B. M., Lara A. H., Miri A., Marshall N. J., Kohn A., Jessell T. M., Abbott L. F., Cunningham J. P., Churchland M. M., Motor cortex embeds muscle-like commands in an untangled population response, Neuron 97 (4) (2018) 953. 29398358 PMC5823788
  33. [36]Beiran M., Dubreuil A., Valente A., Mastrogiuseppe F., Ostojic S., Shaping Dynamics With Multiple Populations in Low-Rank Recurrent Networks, Neural Computation 33 (6) (2021) 1572– 1615. 34496384
  34. [37]Sompolinsky H., Crisanti A., Sommers H. J., Chaos in Random Neural Networks, Physical Review Letters 61 (3) (1988) 259. 10039285
  35. [38]Clark D. G., Abbott L. F., Litwin-Kumar A., Dimension of Activity in Random Neural Networks, Physical Review Letters 131 (11) (2023) 118401. 37774280
  36. [39]Hamood A. W., Marder E., Animal-to-Animal Variability in Neuromodulation and Circuit Function, Cold Spring Harbor symposia on quantitative biology 79 (2014) 21– 28. 25876630 PMC4610821
  37. [40]Prinz A. A., Bucher D., Marder E., Similar network activity from disparate circuit parameters, Nature Neuroscience 7 (12) (2004) 1345– 1352. 15558066
  38. [41]Mastrogiuseppe F., Ostojic S., Linking Connectivity, Dynamics, and Computations in Low-Rank Recurrent Neural Networks, Neuron 99 (3) (2018) 609– 623.e29. 30057201
  39. [42]Dubreuil A., Valente A., Beiran M., Mastrogiuseppe F., Ostojic S., The role of population structure in computations through neural dynamics, Nature neuroscience 25 (6) (2022) 783– 794. 35668174 PMC9284159
  40. [43]Synchronization of cortical activity and its putative role in information processing and learning, Annual review of physiology 55 (1) (1993) 349– 374. 8466179
  41. [44]Koch C., Segev I., The role of single neurons in information processing, Nature neuroscience 3 (11) (2000) 1171– 1177. 11127834
  42. [45]Goaillard J.-M., Marder E., Ion channel degeneracy, variability, and covariation in neuron and circuit resilience, Annual review of neuroscience 44 (2021) 335– 357. 33770451
  43. [46]Li H., Xu Z., Taylor G., Studer C., Goldstein T., Visualizing the Loss Landscape of Neural Nets, Advances in Neural Information Processing Systems 31 (2018).
  44. [47]Fort S., Jastrzebski S., Large Scale Structure of Neural Network Loss Landscapes, Advances in Neural Information Processing Systems 32 (2019).
  45. [49]Gutenkunst R. N., Waterfall J. J., Casey F. P., Brown K. S., Myers C. R., Sethna J. P., Universally Sloppy Parameter Sensitivities in Systems Biology Models, PLOS Computational Biology 3 (10) (2007) e189. 17922568 PMC2000971
  46. [50]Daniels B. C., Chen Y. J., Sethna J. P., Gutenkunst R. N., Myers C. R., Sloppiness, robustness, and evolvability in systems biology, Current opinion in biotechnology 19 (4) (2008) 389– 395. 18620054
  47. [51]Otopalik A. G., Goeritz M. L., Sutton A. C., Brookings T., Guerini C., Marder E., Sloppy morphological tuning in identified neurons of the crustacean stomatogastric ganglion, eLife 6 (feb 2017). PMC5323045 28177286
  48. [52]O’Leary T., Sutton A. C., Marder E., Computational models in the age of large datasets, Current opinion in neurobiology 32 (2015) 87– 94. 25637959 PMC4517826
  49. [53]Werbos P. J., Backpropagation Through Time: What It Does and How to Do It, Proceedings of the IEEE 78 (10) (1990) 1550– 1560.
  50. [54]Kingma D. P., Ba J. L., Adam: A method for stochastic optimization, in: 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings, International Conference on Learning Representations, ICLR, 2015. arXiv:1412.6980. · arXiv
  51. [55]Paszke A., Gross S., Chintala S., Chanan G., Yang E., Facebook Z. D., Research A. I., Lin Z., Desmaison A., Antiga L., Srl O., Lerer A., Automatic differentiation in PyTorch, in: Advances in Neural Information Processing Systems, 2017, pp. 8024– 8035.

Cited by

Sorted by citations

Metadata

Authors
Beiran, M., & Litwin-Kumar, A.
Affiliations
Columbia University Irving Medical Center (US)
Corresponding author
Litwin-Kumar, Ashok
ORCID
Venue
Nature Neuroscience
Publisher
openRxiv
Published
2024-02-27
Volume
28
Issue
12
Pages
2561-2574
ISSN
1097-6256 (ISSN-L)
DOI
10.1101/2024.02.22.581667
PMC
PMC11160579
OpenAlex
W4415648663
Cited by
16
References
0
Citations by year
2026: 13 · 2025: 2 · 2024: 1
FWCI
8.55
Type
journal article
Language
en
Open access
Licence
cc-by-nc-nd
Locations
  • Nature Neuroscience (publishedVersion)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • HAL (submittedVersion, other-oa)
  • PubMed Central (submittedVersion, cc-by-nc-nd)
Funders
  • U.S. Department of Health and Human Services (R01EB029858, RF1DA060772)
  • Burroughs Wellcome Fund
  • Kavli Foundation
  • McKnight Foundation
  • Gatsby Charitable Foundation (GAT3708)
  • National Institutes of Health (1r01eb029858-01, 9rf1da060772-02)
  • National Institute on Drug Abuse (RF1 DA060772)
  • National Institute of Biomedical Imaging and Bioengineering (R01 EB029858)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
  • Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Functional Brain Connectivity Studies
Keywords
connectome, recurrent neural networks, neural dynamics, synaptic connectivity, teacher-student models, neural activity prediction

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review