OSCR

Disinhibitory signaling enables flexible coding of top-down information in cortical networks.

Code ↔ Paper

20 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 20 matches · 4 of them tie a paragraph to a whole file, not to given lines: weak matches, whose lines are not tinted
  1. [1] § Methods › SVM decoding and state space analyses ↔ analysis/decoding/svm_decoding_timeseries.py, lines 238–302 · score 0.79 · fold cross validated, support vector, firing rate, SVC, SVM, classifier
  2. [2] § Methods › Experimental data ↔ analysis/decoding/svm_decoding.m, lines 1–21 · score 0.70 · linear SVM decoding, random train, sliding, trained models, Classifiers, splits
  3. [3] § Results › Two-module RNN model with hierarchical organization ↔ analysis/two_layer_connection_timeconstant.m, lines 1–95 · score 0.68 · connection strengths, feedforward inhibitory, synaptic decay, sensory module, module RNN, excitatory
  4. [4] § Methods › Identifying task and attention cue selective units ↔ analysis/helpers/fnc_instr2_neus.m, the whole file · a weak match · score 0.68 · Wilcoxon rank sum, instruction window, firing rate, pro DMS, DMS task, selective
  5. [5] § Methods › CEBRA analysis ↔ analysis/cebra/cebra_for_matlab.ipynb, lines 102–125 · score 0.68 · offset10 model architecture, batch, cosine, offsets, CEBRA, distance
  6. [6] § Results › Dynamic encoding of top-down information by RNNs ↔ analysis/cebra/two_modality_attended_unattended.m, lines 1–139 · score 0.66 · modality attended, pro task, unattended modality, stimulus identities, teal, modality DMS
  7. [7] § Methods › CEBRA analysis ↔ analysis/two_layer_lesion/cebra_for_matlab.ipynb, lines 70–85 · score 0.66 · offset10 model architecture, batch, cosine, offsets, CEBRA, distance
  8. [8] § Methods › Cross-temporal discriminability analysis ↔ analysis/decoding/cross_temporal_across_models.m, lines 200–244 · score 0.65 · cross temporal discriminability, Fisher, Pearson, diagonal, correlation, transformation
  9. [9] § Results › Dynamic encoding of top-down information by RNNs ↔ analysis/cebra_euclidean_distance/euclidean_distance.py, lines 1–32 · score 0.65 · Euclidean distance, trajectory separation, unattended modality, quantify, modality DMS, RNN
  10. [10] § Methods › SVM decoding and state space analyses ↔ analysis/fixed_points/run_RNN_search.py, lines 82–133 · score 0.65 · outlier distance scale, instruction period, optimization
  11. [11] § Methods › Identifying task and attention cue selective units ↔ analysis/helpers/fnc_task_neus_two_layers.m, the whole file · a weak match · score 0.63 · Wilcoxon rank sum, instruction window, firing rate, pro DMS, selective, anti
  12. [12] § Methods › Continuous-rate RNN model ↔ analysis/helpers/fnc_eval_model.m, the whole file · a weak match · score 0.63 · connection weights, weight matrix, firing rates, channels, sigmoid, variable
  13. [13] § Methods › Continuous-rate RNN model ↔ analysis/helpers/fnc_eval_model2.m, lines 12–88 · score 0.63 · connection weights, weight matrix, firing rates, channels, sigmoid, variable
  14. [14] § Methods › Centrality measure ↔ 3eh_centrality_time.py, lines 103–143 · score 0.63 · closeness centrality, distance matrix, NetworkX
  15. [15] § Results › Two-module RNN model with hierarchical organization ↔ analysis/two_layer_connection_timeconstant.m, lines 1–95 · score 0.61 · feedforward inhibitory, sensory module, module RNN, inhibitory connections, excitatory, trained
  16. [16] § Methods › Experimental data ↔ analysis/decoding/svm_decoding_timeseries.py, lines 238–302 · score 0.59 · fold cross validated, SVM decoding, linear, cue
  17. [17] § Results › Dynamic encoding of top-down information by RNNs ↔ analysis/cebra/two_modality_attended_unattended_eucl_dist.m, lines 1–49 · score 0.58 · trajectory separation, stimulus identities, quantify, Euclidean, modality DMS, unattended
  18. [18] § Results › Dynamic encoding of top-down information by RNNs ↔ analysis/cebra/two_modality_pro_anti.m, the whole file · a weak match · score 0.57 · pro task, stimulus identities, modality DMS task, pro DMS, teal, task cue
  19. [19] § Results › Introducing retro-cue condition elicits shift in strategy ↔ analysis/decoding/plot_cross_temporal.m, lines 25–174 · score 0.56 · epoch boundaries, unattended modality, discriminability, temporal, onset, delay
  20. [20] § Results › Inhibitory units play a critical role in both encoding and maintaining top-down information ↔ 3eh_centrality_time.py, lines 103–143 · score 0.52 · closeness centrality, functional connectivity, activation, distance, neurons

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Python · 318 lines · 13 KB · CC-BY-4.0 · 2 matches

  1. import numpy as np
  2. import scipy.io as si
  3. import os
  4. import h5py
  5. from sklearn.model_selection import KFold
  6. from imblearn.over_sampling import RandomOverSampler
  7. from imblearn.under_sampling import RandomUnderSampler
  8. from sklearn.svm import SVC
  9. import sklearn.pipeline as skp
  10. from sklearn.preprocessing import StandardScaler
  11. import matplotlib.pyplot as plt
  12. from functools import partial
  13. from joblib import Parallel, delayed
  14. import json
  15. from matplotlib.patches import Rectangle
  16. import pdb
  17. import scipy.stats as sst
  18. #rootpath = os.path.join(os.getcwd(), '..')
  19. #sys.path.append(rootpath)
  20. from custom_loadmat import custom_loadmat
  22. class NumpyEncoder(json.JSONEncoder):
  23. def default(self, obj):
  24. if isinstance(obj, np.ndarray):
  25. return obj.tolist()
  26. return json.JSONEncoder.default(self, obj)
  28. def trim_to_match_mean(fr_inh, fr_exc, atol=1e-12):
  29. fr_inh = np.asarray(fr_inh, dtype=float)
  30. fr_exc = np.asarray(fr_exc, dtype=float)
  32. # Optionally ignore NaNs
  33. mask_valid = ~np.isnan(fr_inh)
  34. x = fr_inh[mask_valid]
  36. target = np.nanmean(fr_exc)
  37. if x.size == 0:
  38. return x, np.array([], dtype=int), np.nan
  40. S = x.sum()
  41. N = x.size
  42. mu0 = S / N
  44. # If already <= target, nothing to remove
  45. if mu0 <= target + atol:
  46. kept_idx_global = np.nonzero(mask_valid)[0] # keep all valid
  47. return fr_inh[mask_valid], kept_idx_global, mu0
  49. # If even the minimum element is above target, not achievable by removing top values
  50. xmin = x.min()
  51. if target < xmin - atol:
  52. # Best you can do is keep only the smallest element(s)
  53. # Here we keep all elements equal to xmin
  54. keep_local = np.where(x == xmin)[0]
  55. kept = x[keep_local]
  56. kept_idx_global = np.nonzero(mask_valid)[0][keep_local]
  57. return kept, kept_idx_global, kept.mean()
  59. # Sort descending so cumulative removal is from the top
  60. order_desc = np.argsort(-x)
  61. xd = x[order_desc]
  62. csum = np.concatenate(([0.0], np.cumsum(xd))) # csum[k] = sum of top k removed
  63. k = np.arange(0, N + 1)
  64. denom = (N - k).astype(float)
  65. denom[denom == 0] = np.nan # avoid divide-by-zero for k=N
  66. rem_mean = (S - csum) / denom # remaining mean after removing top k
  68. # Find smallest k with remaining mean <= target
  69. feasible = np.where(rem_mean <= target + atol)[0]
  70. if feasible.size == 0:
  71. k_star = N # degenerate; shouldn't happen due to xmin check above
  72. else:
  73. k_star = feasible[0]
  75. # Build kept set and map back to original indices
  76. keep_local_desc = np.arange(k_star, N) # indices in the desc-sorted array
  77. keep_local_orig = order_desc[keep_local_desc]
  78. kept_idx_global = np.nonzero(mask_valid)[0][keep_local_orig]
  79. kept = fr_inh[kept_idx_global]
  80. return kept, kept_idx_global, kept.mean()
  82. # This function looks up the identities of attended modality stimuli, depending on which modality is cued in each trial
  83. def get_attended_id(id1,id2,cue):
  84. nTrials = id1.shape[1]
  85. attended_id = np.zeros((2,nTrials))
  86. unattended_id = np.zeros((2,nTrials))
  87. for tI in np.arange(nTrials):
  88. if cue[0,tI] == -1:
  89. attended_id[:,tI] = id1[:,tI]
  90. unattended_id[:,tI] = id2[:,tI]
  91. elif cue[0,tI] == 1:
  92. attended_id[:,tI] = id2[:,tI]
  93. unattended_id[:,tI] = id1[:,tI]
  94. return attended_id, unattended_id
  96. # Model loader for each model type
  97. def load_model(mI,subfolders,model_type):
  98. print('Loading model ', mI)
  99. model_folder = subfolders[mI]
  100. trial_path = os.path.join(model_folder, 'trials.mat')
  101. trials = {}
  102. # Skip if trials have not yet been generated for this model
  103. if not os.path.isfile(trial_path):
  104. return {}
  106. # Getting excitatory and inhibitory neurons
  107. model_name = os.path.basename(os.path.normpath(model_folder))+'.mat'
  108. model_weights_folder = os.path.abspath(os.path.join(model_folder,os.pardir,os.pardir))
  109. model_weights_file = os.path.join(model_weights_folder,model_name)
  110. rnn_data = custom_loadmat(model_weights_file)
  111. exc = np.where(rnn_data['exc'])[0]
  112. inh = np.where(rnn_data['inh'])[0]
  114. # Get matrix of simulated trials
  115. with h5py.File(trial_path, "r") as f:
  116. simulated_trials = f['results']
  117. trials['instr_amp_trials'] = simulated_trials['instr_amp_trials'][()] # Instruction amplitude
  118. trials['instr_t_trials'] = simulated_trials['instr_t_trials'][()] # Instruction timing
  119. trials['oo_trials'] = simulated_trials['oo_trials'][()] # Outputs
  120. trials['rr_trials'] = simulated_trials['rr_trials'][()] # Firing rates
  121. trials['u_trials'] = simulated_trials['u_trials'][()] # Input to RNN
  122. if model_type == 'instr': # One-modality models
  123. trials['stim_id_trials'] = simulated_trials['stim_id_trials'][()]
  124. trials['stim_lab_trials'] = simulated_trials['stim_lab_trials'][()]
  125. elif model_type == 'instr2': # Two-modality models
  126. trials['stim_id1_trials'] = simulated_trials['stim_mod1_trials'][()]
  127. trials['stim_id2_trials'] = simulated_trials['stim_mod2_trials'][()]
  128. trials['stim_lab1_trials'] = simulated_trials['stim_lab1_trials'][()]
  129. trials['stim_lab2_trials'] = simulated_trials['stim_lab2_trials'][()]
  130. trials['cue_amp_trials'] = simulated_trials['cue_amp_trials'][()] # Modality cue
  131. # Get id of stim in attended/unattended modalities
  132. trials['attended_id_trials'],trials['unattended_id_trials'] = get_attended_id(trials['stim_id1_trials'],trials['stim_id2_trials'],trials['cue_amp_trials'])
  133. # Getting stims by order (stim1 / stim2)
  134. trials['attended_id_trials_1'], trials['attended_id_trials_2'] = trials['attended_id_trials'][0,:], trials['attended_id_trials'][1,:]
  135. trials['unattended_id_trials_1'], trials['unattended_id_trials_2'] = trials['unattended_id_trials'][0,:], trials['unattended_id_trials'][1,:]
  136. return trials, exc, inh
  138. def fit_svm(tI,X,y,train,test,ic):
  139. # Ensuring equal representation across classes with undersampling within fold
  140. undersample = RandomUnderSampler(sampling_strategy='all')
  141. X_train,X_test = np.squeeze(X[train,:]),np.squeeze(X[test,:])
  142. X_under, ic_under = undersample.fit_resample(X_train, ic[train])
  143. y_train,y_test = np.squeeze(y[train]),np.squeeze(y[test])
  144. y_under = y_train[undersample.sample_indices_]
  145. clf.fit(X_under, y_under)
  146. score = clf.score(X_test, y_test)
  147. return score
  149. def selectTrials(trials,sampleTrials_case):
  150. # Choosing which trials to decode from
  151. if sampleTrials_case == 'instruction_first':
  152. sampleTrials = np.squeeze(trials['instr_t_trials'] == -1) # Instruction first
  153. elif sampleTrials_case == 'instruction_second':
  154. sampleTrials = np.squeeze(trials['instr_t_trials'] == 1) # Instruction second
  155. elif sampleTrials_case == 'first_modality':
  156. sampleTrials = np.squeeze(trials['cue_amp_trials'] == -1) # First modality
  157. elif sampleTrials_case == 'second_modality':
  158. sampleTrials = np.squeeze(trials['cue_amp_trials'] == 1) # Second modality
  159. elif sampleTrials_case == 'first_modality_instruction_first':
  160. sampleTrials = np.squeeze((trials['cue_amp_trials'] == -1) & (trials['instr_t_trials'] == -1)) # First modality and instruction first
  161. elif sampleTrials_case == 'second_modality_instruction_first':
  162. sampleTrials = np.squeeze((trials['cue_amp_trials'] == 1) & (trials['instr_t_trials'] == -1)) # Second modality and instruction first
  163. elif sampleTrials_case == 'all_trials':
  164. sampleTrials = np.squeeze(trials['instr_t_trials'] != 0) # All trials
  165. return sampleTrials
  168. # Load data generated for each model separately
  170. base_folder = '/home/shared/instruction_timing/models/instr/P_rec_0.2_Taus_4.0_25.0/abs/' # One modality
  171. #base_folder = '/home/shared/instruction_timing/models/instr2/P_rec_0.2_Taus_4.0_25.0/abs/' # Two modalities
  172. #base_folder = '/home/shared/instruction_timing/models/instr/P_rec_0.2_Taus_4.0_25.0/abs_two_layers/' # One modality / two layers
  174. #base_folder = '/home/shared/instruction_timing/models/instr/P_rec_0.2_Taus_4.0_25.0/abs_two_layers_wo_inh_feedforward/' # One modality / two layers
  177. os.walk(os.path.join(base_folder,'results'))
  178. #folders = [x[0] for x in os.walk(base_folder)]
  179. folders = [x[0] for x in os.walk(os.path.join(base_folder,'results'))]
  181. subfolders = folders[1:]
  182. # How many separate models were fit
  183. nModels = len(subfolders)
  184. print('# loaded models: ', nModels)
  185. # Which model type is being used (1 modality vs. 2 modality)
  186. if 'instr2' in base_folder:
  187. model_type = 'instr2'
  188. elif 'instr' in base_folder:
  189. model_type = 'instr'
  191. # Decoding each model separately
  192. nFolds = 5
  194. # Choosing which variable to decode
  195. decoded_label = 'instr_amp_trials'
  196. #decoded_label = 'cue_amp_trials'
  197. #decoded_label = 'attended_id_trials'
  198. #decoded_label = 'unattended_id_trials'
  200. # Which stimulus to decode and time window (0/1/-1=all)
  201. #stim_idx = 0
  202. #stim_idx = 1
  203. stim_idx = -1
  205. # Time axis
  206. if stim_idx == 0:
  207. time_axis = np.arange(0,130)
  208. elif stim_idx == 1:
  209. time_axis = np.arange(130,230)
  210. elif stim_idx == -1:
  211. time_axis = np.arange(350)
  212. nTimes = len(time_axis)
  214. # Choosing which trials subset to decode from (instruction_first, instruction_second, all_trials)
  215. sampleTrials_case = 'instruction_first'
  216. #sampleTrials_case = 'instruction_second'
  218. # Target neurons for decoding
  219. #target_neurons = 'layer1_inh'
  220. #target_neurons = 'layer1_exc'
  222. #target_neurons = 'layer2_inh'
  223. #target_neurons = 'layer2_exc'
  225. target_neurons = 'inh'
  226. #target_neurons = 'exc'
  228. #target_neurons = 'matched'
  231. if 'layer1' in target_neurons:
  232. layer_neurons = np.arange(200)
  233. elif 'layer2' in target_neurons:
  234. layer_neurons = np.arange(200,1000)
  235. else:
  236. layer_neurons = np.arange(1000)
  238. # Decoding models separately
  239. score = np.zeros((nModels,nFolds,nTimes))
  240. for mI in np.arange(nModels):
  241. trials, exc, inh = load_model(mI,subfolders,model_type)
  242. if not trials: # Skip if trials have not yet been generated for this model
  243. continue
  244. if 'exc' in target_neurons:
  245. selected_neurons = np.intersect1d(layer_neurons,exc)
  246. elif 'inh' in target_neurons:
  247. selected_neurons = np.intersect1d(layer_neurons,inh)
  248. elif 'matched' in target_neurons:
  249. exc_layer = np.intersect1d(layer_neurons,exc)
  250. inh_layer = np.intersect1d(layer_neurons,inh)
  251. sampleTrials = selectTrials(trials,sampleTrials_case)
  252. X = np.transpose(trials['rr_trials'][:,:,sampleTrials],(2,1,0))
  253. fr_all = np.mean(np.mean(X,axis=0),axis=0)
  254. fr_inh = fr_all[inh_layer]
  255. fr_exc = fr_all[exc_layer]
  256. kept_values, kept_indices, final_mean = trim_to_match_mean(fr_inh, fr_exc)
  257. selected_neurons = np.sort(inh_layer[kept_indices])
  259. else:
  260. selected_neurons = layer_neurons
  262. sampleTrials = selectTrials(trials,sampleTrials_case)
  263. # Loading firing rate data (nTrials x nTimes x nNeurons)
  264. X = np.transpose(trials['rr_trials'][:,:,sampleTrials],(2,1,0))
  265. X = X[:,:,selected_neurons]
  267. # Loading labels to be decoded
  268. # These labels only allow stim_idx = 0
  269. if decoded_label == 'instr_amp_trials' or decoded_label == 'cue_amp_trials' or stim_idx == -1:
  270. decoded_idx = 0
  271. else:
  272. decoded_idx = stim_idx
  275. y = trials[decoded_label][decoded_idx,sampleTrials]
  277. # Ensuring all classes have equal representation
  278. instr_amp = np.expand_dims(trials['instr_amp_trials'][decoded_idx,sampleTrials],axis=1)
  279. if model_type == 'instr2':
  280. cue_amp = np.expand_dims(trials['cue_amp_trials'][decoded_idx,sampleTrials],axis=1)
  281. attended_id = np.expand_dims(trials['attended_id_trials'][decoded_idx,sampleTrials],axis=1)
  282. unattended_id = np.expand_dims(trials['unattended_id_trials'][decoded_idx,sampleTrials],axis=1)
  283. all_labels = np.concatenate((instr_amp,cue_amp,attended_id,unattended_id,np.expand_dims(y,axis=1)),axis=1)
  284. elif model_type == 'instr':
  285. all_labels = np.concatenate((instr_amp,np.expand_dims(y,axis=1)),axis=1)
  286. uniqueConditions,ic = np.unique(all_labels,axis=0,return_inverse=True)
  288. # Standard pipeline to z-score and run support vector classifcation
  289. clf = skp.make_pipeline(SVC(kernel='linear'))
  291. # Performing kFold cross-validated SVM decoding
  292. model_score = np.zeros((nFolds, nTimes))
  293. kf = KFold(n_splits=nFolds)
  294. for tI,time_val in enumerate(time_axis):
  295. timeX = np.squeeze(X[:,time_val,:])
  296. if tI % 50 == 0:
  297. print('Model: ', mI, ' / Time: ', time_val)
  298. # Parallelize results for each fold
  299. score[mI,:,tI] = Parallel(n_jobs=24)(
  300. delayed(fit_svm)(time_val,timeX,y,train_index,test_index,ic)
  301. for fI, (train_index, test_index) in enumerate(kf.split(timeX))
  302. )
  304. # Save score to file
  305. valid_models = (np.sum(np.sum(score,axis=1),axis=1)) != 0
  306. valid_score = score[valid_models,:,:]
  307. # Save score to file
  308. save_data = {'score':valid_score, 'time_axis':time_axis}
  309. base_save_folder = os.path.abspath(os.path.join(base_folder))
  310. if model_type == 'instr': # One-modality models
  311. save_folder = os.path.join(base_save_folder,'decoding',sampleTrials_case,decoded_label,target_neurons,str(stim_idx))
  312. elif model_type == 'instr2': # Two-modality models
  313. save_folder = os.path.join(base_save_folder,'decoding',sampleTrials_case,decoded_label,'modality'+str(stim_idx),target_neurons)
  315. if not os.path.exists(save_folder):
  316. os.makedirs(save_folder)
  317. with open(os.path.join(save_folder, 'score.json'), 'w') as f:
  318. json.dump(save_data, f, cls=NumpyEncoder)

svm_decoding_timeseries.py, under CC-BY-4.0 · at the source

Overview

Authors: Tomas G. Aquino1, Robert Kim2, Nuttida Rungratsameetaweemana1
ORCID iDs: Robert Kim, Nuttida Rungratsameetaweemana
  1. Department of Biomedical Engineering, Columbia University, New York, New York, United States of America
  2. Department of Neurology, Cedars-Sinai Medical Center, Los Angeles, California, United States of America
Institutions: Columbia University (United States); Cedars-Sinai Medical Center (United States)
Journal: PLoS biology, volume 24, issue 7, article e3003831
Dates: received 6 July 2025; accepted 18 May 2026; published online 2 July 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1371/journal.pbio.3003831 · PMID 42391203 · PMCID PMC13327264 · OpenAlex W7167020486
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: mouse (organism), systems (subfield)
Methods: Statistics, Smoothing, state filtering, decompositions, Machine learning, Preprocessing, Connectivity, Single-unit activity, calcium imaging
MeSH: Nerve Net*, Visual Cortex*, Animals, Interneurons, Mice, Models, Neurological, Recurrent Neural Networks, Signal Transduction (* major topic)
Journal subjects: Biology and Life Sciences, Neuroscience, Cognitive Science, Cognitive Psychology, Perception, Sensory Perception, Psychology, Social Sciences, Sensory Cues, Attention, Cell Biology, Cellular Types, Animal Cells, Neurons, Cellular Neuroscience, Signal Transduction, Cell Signaling, Signal Inhibition, Computer and Information Sciences, Network Analysis, Centrality, Vision, Engineering and Technology, Signal Processing
Topic: Neural dynamics and brain function (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Funding: DEVCOM Army Research Laboratory (W911NF-23-2-0067, W911NF-22-2-0148)
Citations: not cited yet (Europe PMC); 103 references in the paper

Abstract

Flexible behavior requires the ability to modulate sensory processing based on task context, yet the circuit-level mechanisms supporting this capacity remain poorly understood. Here, we combine recurrent neural network modeling and neural recordings from mouse visual cortex to investigate how task context shapes sensory coding. Networks trained on an instruction-based discrimination task develop a disinhibitory interneuron-to-interneuron motif that dynamically gates task-relevant sensory information. Perturbation and lesion analyses show that this motif is necessary for task performance and for maintaining distinct sensory representations across contexts. We validate key predictions in mouse visual cortex, where interneuron activity patterns exhibit comparable task-dependent modulation. These results identify a biologically plausible circuit motif that supports flexible sensory processing and link recurrent connectivity structure to adaptive context integration in both artificial and biological systems.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above, with 20 matches between paragraphs and lines of code.

Zenodo 20148662

License: CC-BY-4.0
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Languages: MATLAB (40), Python (14), Shell (4), Jupyter (2)
Size: 64 files, 60 scripts
Software Heritage: not checked
Found in: “Data Availability”
Holds: README, 2 notebooks
Not found: license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: Statistics and Machine Learning Toolbox (41 files), NumPy (31 files), SciPy (29 files), Matplotlib (25 files), h5py (17 files), scikit-learn (16 files), pandas (8 files), TensorFlow (6 files), imbalanced-learn (4 files), boundedline (2 files), Image Processing Toolbox (2 files), NetworkX (2 files), PyTorch (2 files), scikit-posthocs (2 files)
Availability: 1 check, the latest on 27 September 2026: the link answers (HTTP 200)
  • 27 September 2026: the link answers (HTTP 200)
123 files
At the source: doi.org/10.5281/zenodo.20148662

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 122 scripts, each with its path and the digest of its content;
  • 20 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

Datasets cited

Data Availability

The code for the analyses performed in this study is available in Zenodo at https://doi.org/10.5281/zenodo.20148662 (Shared code for Disinhibitory signaling enables flexible coding of top-down information in cortical networks). The trained RNN models used in the study, as well as the analysis data, are available in Zenodo at https://doi.org/10.5281/zenodo.20148306 (Shared data for Disinhibitory signaling enables flexible coding of top-down information in cortical networks).

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 3 authors, 8 MeSH terms, 1 funder, 96 references.

Cite

This paper

Aquino, T. G., Kim, R., & Rungratsameetaweemana, N. (2026). Disinhibitory signaling enables flexible coding of top-down information in cortical networks. PLoS biology, 24(7), e3003831. https://doi.org/10.1371/journal.pbio.3003831

BibTeX

@article{aquino2026disinhibitory,
author = {Aquino, Tomas G. and Kim, Robert and Rungratsameetaweemana, Nuttida},
title = {{Disinhibitory signaling enables flexible coding of top-down information in cortical networks}},
journal = {PLoS biology},
year = {2026},
month = jul,
volume = {24},
number = {7},
pages = {e3003831},
publisher = {PLOS},
issn = {1544-9173},
doi = {10.1371/journal.pbio.3003831},
url = {https://doi.org/10.1371/journal.pbio.3003831},
pmid = {42391203},
pmcid = {PMC13327264}
}

RIS

TY - JOUR
AU - Aquino, Tomas G.
AU - Kim, Robert
AU - Rungratsameetaweemana, Nuttida
TI - Disinhibitory signaling enables flexible coding of top-down information in cortical networks
T2 - PLoS biology
J2 - PLoS Biol
PY - 2026
DA - 2026/07/02
VL - 24
IS - 7
SP - e3003831
SN - 1544-9173
PB - PLOS
DO - 10.1371/journal.pbio.3003831
UR - https://doi.org/10.1371/journal.pbio.3003831
LA - en
ER -

CSL-JSON

{
"id": "10.1371/journal.pbio.3003831",
"type": "article-journal",
"title": "Disinhibitory signaling enables flexible coding of top-down information in cortical networks",
"container-title": "PLoS biology",
"author": [
{
"family": "Aquino",
"given": "Tomas G."
},
{
"family": "Kim",
"given": "Robert"
},
{
"family": "Rungratsameetaweemana",
"given": "Nuttida"
}
],
"container-title-short": "PLoS Biol",
"volume": "24",
"issue": "7",
"page": "e3003831",
"DOI": "10.1371/journal.pbio.3003831",
"PMID": "42391203",
"PMCID": "PMC13327264",
"ISSN": "1544-9173",
"publisher": "PLOS",
"URL": "https://doi.org/10.1371/journal.pbio.3003831",
"language": "en",
"issued": {
"date-parts": [
[
2026,
7,
2
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1523/jneurosci.0987-25.2026 [code]
Cell-Type-Specific Synaptic Scaling Mechanisms Differentially Contribute to Associative Learning.
Journal: The Journal of neuroscience : the official journal of the Society for Neuroscience
In common: pandas, SciPy, Matplotlib, 1 other tool, 10 references
[2] doi:10.1016/j.isci.2026.116776 [code]
Selective perturbation of mirror and non-mirror neurons in an in silico model of the action observation network.
Journal: iScience
In common: TensorFlow, SciPy, NumPy, 9 references
[3] doi:10.7554/elife.109717 [code]
Retrosplenial cortex enables context-dependent goal-directed sensorimotor transformation.
Journal: eLife
In common: h5py, Statistics and Machine Learning Toolbox, scikit-learn, 4 other tools, systems, mouse, 5 references
[4] doi:10.1038/s41467-026-76104-3 [code]
Sensorimotor remapping drives task specialization in prefrontal cortex.
Journal: Nature communications
In common: NetworkX, scikit-learn, pandas, 3 other tools, systems, 6 references
[5] doi:10.1038/s41586-026-10448-0 [code]
Plasticity and language in the anaesthetized human hippocampus.
Journal: Nature
In common: TensorFlow, Image Processing Toolbox, PyTorch, 4 other tools, 1 reference, author Robert Kim
[6] doi:10.1371/journal.pcbi.1014162 [code]
Exploring neural manifolds across a wide range of intrinsic dimensions.
Journal: PLoS computational biology
In common: TensorFlow, h5py, scikit-learn, 4 other tools, 5 references
[7] doi:10.7554/elife.110588 [code]
Opening the black box toward a modular approach to spike sorting.
Journal: eLife
In common: imbalanced-learn, TensorFlow, NetworkX, 7 other tools, mouse, 1 reference
[8] doi:10.1016/j.neuron.2026.03.034 [code]
Dentate gyrus interneurons modulate winner-take-all network dynamics in freely behaving mice.
Journal: Neuron
In common: boundedline, NetworkX, h5py, 8 other tools, systems, mouse
[9] doi:10.1016/j.neuron.2026.07.016 [code]
Inferring brain-wide interactions using data-constrained recurrent neural network models.
Journal: Neuron
In common: Image Processing Toolbox, Statistics and Machine Learning Toolbox, Matplotlib, 1 other tool, systems, mouse, 6 references
[10] doi:10.1016/j.isci.2026.117375 [code]
Motor priming is associated with widespread recruitment into neural ensembles and more rapid ensemble transitions.
Journal: iScience
In common: scikit-posthocs, NetworkX, h5py, 8 other tools, systems

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.