OSCR

Global coincident bursts of high frequency oscillations across the human cortex coordinate large-scale memory processing.

Code ↔ Paper

3 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 3 matches
  1. [1] § Methods › Analysis of coincident HFO bursting ↔ sync_events.py, lines 70–70 · score 0.69 · 150–250 Hz, 250–500 Hz, 60–150 Hz, 60 Hz, events
  2. [2] § Methods › Electrophysiological recordings ↔ run_HFO_detect.py, lines 1–29 · score 0.62 · University Hospital, HFOs detected, Anne, Jan, Brno, electrophysiological
  3. [3] § Methods › Participants and electrode contact localization ↔ run_HFO_detect.py, lines 1–29 · score 0.58 · University Hospital, Anne, Jan, Czech, clinical, Brno

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 · 485 lines · 20 KB · MIT · 2 matches

  1. #!/usr/bin/env python3
  2. # -*- coding: utf-8 -*-
  3. """
  4. Created on Thu Nov 21 15:49:50 2019
  6. Script to run EPYCOM algorithms on memory encoding dataset
  8. Ing.,Mgr. (MSc.) Jan Cimbálník
  9. Biomedical engineering
  10. International Clinical Research Center
  11. St. Anne's University Hospital in Brno
  12. Czech Republic
  13. &
  14. Mayo systems electrophysiology lab
  15. Mayo Clinic
  16. 200 1st St SW
  17. Rochester, MN
  18. United States
  19. """
  21. import os, argparse, re
  22. from time import time
  24. import numpy as np
  25. import pandas as pd
  27. from bids import BIDSLayout, BIDSLayoutIndexer
  28. from epycom.event_detection import HilbertDetector
  29. from pymef import MefSession
  31. # from pathlib import Path
  32. # file_path = Path(__file__).parent.absolute()
  34. # %% Parse arguments
  36. parser = argparse.ArgumentParser(description='Memory encoding task detection of HFO')
  38. parser.add_argument('path_to_dataset', action="store",
  39. nargs='?', type=str, help='Specify BIDS dataset path')
  40. parser.add_argument('--subjects', default=None, action="store",
  41. nargs='+', type=int, help='Specify BIDS subject')
  42. parser.add_argument('--sessions', default=None, action="store",
  43. nargs='+',type=int, help='Specify BIDS session')
  44. parser.add_argument('--tasks', default=None, action="store",
  45. nargs='+',type=str, help='Specify BIDS task')
  46. parser.add_argument('--runs', default=None, action="store",
  47. nargs='+',type=int, help='Specify BIDS run')
  49. parser.add_argument('--channels', default=None, action="store",
  50. nargs='+',type=str, help='Specify iEEG channels run')
  52. parser.add_argument('--montage', default='unipolar', action="store",
  53. type=str, help='Specify montage for processing (unipolar, bipolar, wm)')
  55. parser.add_argument('--n_cores', default=None, action="store", type=int, help='Specify number of cores for multiprocessing')
  58. parsed_vals = parser.parse_args()
  60. print(parsed_vals)
  62. path_to_dataset = parsed_vals.path_to_dataset
  63. subjects = parsed_vals.subjects
  64. sessions = parsed_vals.sessions
  65. tasks = parsed_vals.tasks
  66. runs = parsed_vals.runs
  68. channels = parsed_vals.channels
  70. montage = parsed_vals.montage
  72. n_cores = parsed_vals.n_cores
  74. # %% Presets
  76. if 'fnusa' in path_to_dataset:
  77. fs = 5000 # We know this for iEEG
  78. elif 'mayo' in path_to_dataset:
  79. fs = 32000
  80. elif 'wroclaw' in path_to_dataset:
  81. fs = 4000
  83. if not path_to_dataset.endswith('/'):
  84. path_to_dataset += '/'
  85. path_to_source = path_to_dataset+'sourcedata/'
  86. path_to_derivatives = path_to_dataset+'derivatives/'
  87. os.makedirs(path_to_derivatives, exist_ok=True)
  89. detect_on = ['events', 'whole_rec', 'recall']
  91. mef_pwd = None
  93. pre_offset = 1.25 # in seconds
  94. post_offset = 1.25 # in seconds
  96. win_size = 10 # in seconds
  98. threshold = 2.5
  100. compute_instance = HilbertDetector(low_fc=60,
  101. high_fc=800,
  102. band_spacing='log',
  103. num_bands=100,
  104. cyc_th=1,
  105. threshold=threshold)
  107. # %% Processing - in small windows (2.5 seconds)
  108. """
  109. The middle point for the window is:
  110. COUNTDOWN - appearance of the countdown digit
  111. ENCODING - appearance of the word(s)
  112. DISTRACTOR - appearance of the equation
  113. RECALL - start of word vocalization
  114. """
  116. # Get data files
  117. mefd_pattern = re.compile(r'\.mefd$')
  118. indexer = BIDSLayoutIndexer(ignore=[mefd_pattern, 'sourcedata', 'derivatives', 'code'])
  119. l = BIDSLayout(path_to_dataset, indexer=indexer,
  120. database_path=path_to_dataset+'.sql', reset_database=True)
  121. filter_dict = {'suffix': 'ieeg',
  122. 'extension': 'json'}
  123. if subjects is not None:
  124. filter_dict['subject'] = [str(x).zfill(3) for x in subjects]
  125. if sessions is not None:
  126. filter_dict['session'] = [str(x).zfill(3) for x in sessions]
  127. if tasks is not None:
  128. filter_dict['task'] = tasks
  129. if runs is not None:
  130. filter_dict['run'] = runs
  131. json_files = l.get(**filter_dict)
  133. mef_sessions = []
  134. channel_dfs = []
  135. event_dfs = []
  136. for json_file in json_files:
  137. json_entities = json_file.entities
  139. # Get valid channels
  140. channel_file = l.get(suffix='channels', extension='tsv',
  141. task=json_entities['task'],
  142. subject=json_entities['subject'],
  143. session=json_entities['session'],
  144. run=json_entities['run'])[0]
  145. channel_df = channel_file.get_df()
  146. channel_dfs.append(channel_df[channel_df['type'].isin(['SEEG', 'ECOG'])])
  149. # Get events of interest
  150. events_file = l.get(suffix='events', extension='tsv',
  151. task=json_entities['task'],
  152. subject=json_entities['subject'],
  153. session=json_entities['session'],
  154. run=json_entities['run'])[0]
  155. events_df = events_file.get_df()
  156. # events_df = events_df.loc[~events_df['trial_type'].isna(), ['onset', 'sample', 'duration', 'trial_type']]
  157. event_dfs.append(events_df.loc[:, ['onset', 'sample', 'trial_type', 'value']])
  159. mef_session_path = os.path.splitext(json_file.path)[0]+'.mefd'
  160. mef_sessions.append(MefSession(mef_session_path, mef_pwd))
  162. if montage == 'unipolar':
  164. for ms, ch_df, ev_df, json_file in zip(mef_sessions, channel_dfs, event_dfs, json_files):
  166. print(f"Working on {json_file.filename[:-5]}")
  168. json_entities = json_file.entities
  170. ttl_df = ev_df.loc[ev_df['trial_type'].isna(), :]
  171. ev_df = ev_df.loc[~ev_df['trial_type'].isna(), :]
  173. # Get uUTC unique times (PAL task can have the times doubled)
  174. uq_times = ev_df['onset'].unique()
  175. uq_times *= 1e6
  176. uq_uutc_times = (uq_times + ms.session_md['session_specific_metadata']['earliest_start_time'][0]).astype(int)
  177. uq_samples = ev_df['sample'].unique()
  179. # Basic channel info
  180. bi = ms.read_ts_channel_basic_info()
  182. # Iterate over channels
  183. all_chan_seg_df_list = []
  184. all_chan_whole_df_list = []
  185. all_chan_recalls_df_list = []
  186. for ch in list(ch_df['name']):
  188. if channels is not None:
  189. if ch not in channels:
  190. continue
  192. if ch not in[x['name'] for x in ms.read_ts_channel_basic_info()]:
  193. continue
  195. print(f"\tDetecting channel {ch}")
  197. data = ms.read_ts_channels_sample(ch, [None, None])
  198. fs = int([x['fsamp'] for x in bi if x['name'] == ch][0][0])
  199. ch_start_time = [x['start_time'] for x in bi if x['name'] == ch][0][0]
  200. compute_instance.params['fs'] = fs
  202. # Samp offsets
  203. pre_samp_offset = int(pre_offset*fs)
  204. post_samp_offset = int(post_offset*fs)
  206. # Since our window is constant we can create "artifical" channel by gluing the segments together
  207. starts = [x-pre_samp_offset for x in uq_samples if ((x-pre_samp_offset) > 0 and (x+post_samp_offset) < len(data))]
  208. stops = [x+post_samp_offset for x in uq_samples if ((x-pre_samp_offset) > 0 and (x+post_samp_offset) < len(data))]
  210. idx_arr = np.concatenate([np.arange(x, y) for x, y in zip(starts, stops)])
  212. seg_data = data[idx_arr]
  214. t = time()
  216. # ----- Segmented -----
  218. # Run windowed function over the segmented data
  219. res = compute_instance.run_windowed(seg_data,
  220. window_size=pre_samp_offset+post_samp_offset,
  221. n_cores=n_cores)
  222. res_df = pd.DataFrame(res)
  224. # Correct the window starts and stops based on sample starts/stops
  225. for win_idx in res_df.win_idx.unique():
  226. win_start = starts[win_idx]
  227. res_df.loc[res_df['win_idx'] == win_idx, ['event_start', 'event_stop']] += win_start
  229. res_df['event_start'] = (((res_df['event_start']/fs) * 1e6) + ch_start_time).astype(int)
  230. res_df['event_stop'] = (((res_df['event_stop']/fs) * 1e6) + ch_start_time).astype(int)
  232. # Assign channel
  233. res_df['channel'] = ch
  236. all_chan_seg_df_list.append(res_df)
  238. # ----- Whole recording
  240. # Run windowed function over the whole recroding
  241. whole_res = compute_instance.run_windowed(data,
  242. window_size=win_size*fs,
  243. n_cores=n_cores)
  244. whole_res_df = pd.DataFrame(whole_res)
  246. # Correct the window starts and stops based on sample starts/stops
  247. for win_idx in whole_res_df.win_idx.unique():
  248. whole_res_df.loc[whole_res_df['win_idx'] == win_idx, ['event_start', 'event_stop']] += win_idx*win_size*fs
  250. whole_res_df['event_start'] = (((whole_res_df['event_start']/fs) * 1e6) + ch_start_time).astype(int)
  251. whole_res_df['event_stop'] = (((whole_res_df['event_stop']/fs) * 1e6) + ch_start_time).astype(int)
  253. # Assign channel
  254. whole_res_df['channel'] = ch
  256. all_chan_whole_df_list.append(whole_res_df)
  258. # ----- Recall segments (only FR and PAL)-----
  259. if json_entities['task'] in ['FR', 'PAL']:
  260. # Find all recording starts
  261. rec_starts = ttl_df.loc[ttl_df['value'] == 232]
  262. rec_stops = ttl_df.loc[ttl_df['value'] == 233]
  264. # Determine the ones belonging to recall - every third recording start and stop
  265. recall_idxs = range(2, len(rec_starts), 3)
  266. recall_starts = rec_starts.iloc[recall_idxs, :]
  267. recall_stops = rec_stops.iloc[recall_idxs, :]
  269. starts = [x-pre_samp_offset for x in recall_starts['sample']]
  270. stops = [x+post_samp_offset-1 for x in recall_stops['sample']]
  272. recalls_res_df_list = []
  273. for start, stop in zip(starts, stops):
  274. if start < 0 or stop > len(data):
  275. continue
  276. idx_arr = np.arange(start, stop)
  277. recall_data = data[idx_arr]
  279. res = compute_instance.run_windowed(recall_data,
  280. window_size=pre_samp_offset+post_samp_offset,
  281. n_cores=n_cores)
  282. recall_res_df = pd.DataFrame(res)
  284. for win_idx in recall_res_df.win_idx.unique():
  285. recall_res_df.loc[recall_res_df['win_idx'] == win_idx, ['event_start', 'event_stop']] += win_idx*win_size*fs
  287. recall_res_df['event_start'] = (((recall_res_df['event_start']/fs) * 1e6) + ch_start_time).astype(int)
  288. recall_res_df['event_stop'] = (((recall_res_df['event_stop']/fs) * 1e6) + ch_start_time).astype(int)
  290. recalls_res_df_list.append(recall_res_df)
  292. recalls_res_df = pd.concat(recalls_res_df_list)
  293. recalls_res_df['channel'] = ch
  295. all_chan_recalls_df_list.append(recalls_res_df)
  297. print(f"\tProcessed in {time()-t} s")
  301. # Save the dataframe (on disk for now)
  302. all_chan_seg_df = pd.concat(all_chan_seg_df_list)
  303. all_chan_whole_df = pd.concat(all_chan_whole_df_list)
  304. sub = json_entities['subject']
  305. ses = json_entities['session']
  306. run = json_entities['run']
  307. task = json_entities['task']
  308. all_chan_seg_df.to_pickle(f"{path_to_derivatives}/{montage}/sub-{sub}_ses-{ses}_run-{run}_task-{task}_th-{threshold}_mont-{montage}.pkl")
  309. all_chan_whole_df.to_pickle(f"{path_to_derivatives}/{montage}/sub-{sub}_ses-{ses}_run-{run}_task-{task}_th-{threshold}_mont-{montage}_whole_channel.pkl")
  310. if json_entities['task'] in ['FR', 'PAL']:
  311. all_chan_recalls_df = pd.concat(all_chan_recalls_df_list)
  312. all_chan_recalls_df.to_pickle(f"{path_to_derivatives}/{montage}/sub-{sub}_ses-{ses}_run-{run}_task-{task}_th-{threshold}_mont-{montage}_whole_recalls.pkl")
  314. elif montage == 'bipolar':
  316. for ms, ch_df, ev_df, json_file in zip(mef_sessions, channel_dfs, event_dfs, json_files):
  318. ch_df = ch_df[~(ch_df['name'].str.replace('[^0-9]', '', regex=True) == '')]
  320. print(f"Working on {json_file.filename[:-5]}")
  322. json_entities = json_file.entities
  324. ttl_df = ev_df.loc[ev_df['trial_type'].isna(), :]
  325. ev_df = ev_df.loc[~ev_df['trial_type'].isna(), :]
  327. # Get uUTC unique times (PAL task can have the times doubled)
  328. uq_times = ev_df['onset'].unique()
  329. uq_times *= 1e6
  330. uq_uutc_times = (uq_times + ms.session_md['session_specific_metadata']['earliest_start_time'][0]).astype(int)
  331. uq_samples = ev_df['sample'].unique()
  333. # Basic channel info
  334. bi = ms.read_ts_channel_basic_info()
  336. # Create channel pairs
  337. ch_df = ch_df.sort_values('name').reset_index(drop=True)
  339. # Create channel pairs
  340. ch_df['group'] = ch_df['name'].str.replace(r'\d+', '', regex=True)
  341. ch_df['group'] = ch_df['group'].str.replace('_', '')
  342. ch_df['number'] = ch_df['name'].str.replace('[^0-9]', '', regex=True).astype(int)
  344. chan_pairs = []
  345. for i, r in ch_df.iterrows():
  346. second_chan = ch_df[(ch_df['group'] == r['group']) & (ch_df['number'] == r['number']+1)]
  347. if len(second_chan):
  348. chan_pairs.append([r['name'], second_chan['name'].values[0]])
  351. # Iterate over channels
  352. all_chan_seg_df_list = []
  353. all_chan_whole_df_list = []
  354. all_chan_recalls_df_list = []
  355. for ch_pair in chan_pairs:
  357. if ch_pair[0] not in ch_df['name'].unique() or ch_pair[1] not in ch_df['name'].unique():
  358. continue
  360. if ch_pair[0] not in[x['name'] for x in ms.read_ts_channel_basic_info()] or ch_pair[1] not in[x['name'] for x in ms.read_ts_channel_basic_info()]:
  361. continue
  363. print(f"\tDetecting channel {ch_pair[0]}-{ch_pair[1]}")
  365. data = ms.read_ts_channels_sample(ch_pair, [None, None])
  366. fs = int([x['fsamp'] for x in bi if x['name'] == ch_pair[0]][0][0])
  367. ch_start_time = [x['start_time'] for x in bi if x['name'] == ch_pair[0]][0][0]
  368. compute_instance.params['fs'] = fs
  370. if len(data[0]) != len(data[1]):
  371. continue
  373. data = data[0]-data[1]
  375. # Samp offsets
  376. pre_samp_offset = int(pre_offset*fs)
  377. post_samp_offset = int(post_offset*fs)
  379. # Since our window is constant we can create "artifical" channel by gluing the segments together
  380. starts = [x-pre_samp_offset for x in uq_samples if ((x-pre_samp_offset) > 0 and (x+post_samp_offset) < len(data))]
  381. stops = [x+post_samp_offset for x in uq_samples if ((x-pre_samp_offset) > 0 and (x+post_samp_offset) < len(data))]
  383. idx_arr = np.concatenate([np.arange(x, y) for x, y in zip(starts, stops)])
  385. seg_data = data[idx_arr]
  387. t = time()
  389. # ----- Segmented -----
  390. # Run windowed function over the segmented data
  391. res = compute_instance.run_windowed(seg_data,
  392. window_size=pre_samp_offset+post_samp_offset,
  393. n_cores=n_cores)
  394. res_df = pd.DataFrame(res)
  396. # Correct the window starts and stops based on sample starts/stops
  397. for win_idx in res_df.win_idx.unique():
  398. win_start = starts[win_idx]
  399. res_df.loc[res_df['win_idx'] == win_idx, ['event_start', 'event_stop']] += win_start
  401. res_df['event_start'] = (((res_df['event_start']/fs) * 1e6) + ch_start_time).astype(int)
  402. res_df['event_stop'] = (((res_df['event_stop']/fs) * 1e6) + ch_start_time).astype(int)
  404. # Assign channel
  405. res_df['channel'] = '-'.join(ch_pair)
  407. all_chan_seg_df_list.append(res_df)
  409. # ----- Whole recording
  411. # Run windowed function over the whole recroding
  412. whole_res = compute_instance.run_windowed(data,
  413. window_size=win_size*fs,
  414. n_cores=n_cores)
  415. whole_res_df = pd.DataFrame(whole_res)
  417. # Correct the window starts and stops based on sample starts/stops
  418. for win_idx in whole_res_df.win_idx.unique():
  419. whole_res_df.loc[whole_res_df['win_idx'] == win_idx, ['event_start', 'event_stop']] += win_idx*win_size*fs
  421. whole_res_df['event_start'] = (((whole_res_df['event_start']/fs) * 1e6) + ch_start_time).astype(int)
  422. whole_res_df['event_stop'] = (((whole_res_df['event_stop']/fs) * 1e6) + ch_start_time).astype(int)
  424. # Assign channel
  425. whole_res_df['channel'] = '-'.join(ch_pair)
  427. all_chan_whole_df_list.append(whole_res_df)
  429. # ----- Recall segments (only FR and PAL)-----
  430. if json_entities['task'] in ['FR', 'PAL']:
  431. # Find all recording starts
  432. rec_starts = ttl_df.loc[ttl_df['value'] == 232]
  433. rec_stops = ttl_df.loc[ttl_df['value'] == 233]
  435. # Determine the ones belonging to recall - every third recording start and stop
  436. # !!! This should be veryfied for the older recordings!!!!
  437. recall_idxs = range(2, len(rec_starts), 3)
  438. recall_starts = rec_starts.iloc[recall_idxs, :]
  439. recall_stops = rec_stops.iloc[recall_idxs, :]
  441. starts = [x-pre_samp_offset for x in recall_starts['sample']]
  442. stops = [x+post_samp_offset-1 for x in recall_stops['sample']]
  444. recalls_res_df_list = []
  445. for start, stop in zip(starts, stops):
  446. if start < 0 or stop > len(data):
  447. continue
  448. idx_arr = np.arange(start, stop)
  449. recall_data = data[idx_arr]
  451. res = compute_instance.run_windowed(recall_data,
  452. window_size=pre_samp_offset+post_samp_offset,
  453. n_cores=n_cores)
  454. recall_res_df = pd.DataFrame(res)
  456. for win_idx in recall_res_df.win_idx.unique():
  457. recall_res_df.loc[recall_res_df['win_idx'] == win_idx, ['event_start', 'event_stop']] += win_idx*win_size*fs
  459. recall_res_df['event_start'] = (((recall_res_df['event_start']/fs) * 1e6) + ch_start_time).astype(int)
  460. recall_res_df['event_stop'] = (((recall_res_df['event_stop']/fs) * 1e6) + ch_start_time).astype(int)
  462. recalls_res_df_list.append(recall_res_df)
  464. recalls_res_df = pd.concat(recalls_res_df_list)
  465. recalls_res_df['channel'] = '-'.join(ch_pair)
  467. all_chan_recalls_df_list.append(recalls_res_df)
  469. print(f"\tProcessed in {time()-t} s")
  471. # Save the dataframe (on disk for now)
  472. all_chan_seg_df = pd.concat(all_chan_seg_df_list)
  474. all_chan_whole_df = pd.concat(all_chan_whole_df_list)
  475. sub = json_entities['subject']
  476. ses = json_entities['session']
  477. run = json_entities['run']
  478. task = json_entities['task']
  479. all_chan_whole_df.to_pickle(f"{path_to_derivatives}/{montage}/sub-{sub}_ses-{ses}_run-{run}_task-{task}_th-{threshold}_mont-{montage}_whole_channel.pkl")
  480. if json_entities['task'] in ['FR', 'PAL']:
  481. all_chan_recalls_df = pd.concat(all_chan_recalls_df_list)
  482. all_chan_recalls_df.to_pickle(f"{path_to_derivatives}/{montage}/sub-{sub}_ses-{ses}_run-{run}_task-{task}_th-{threshold}_mont-{montage}_recalls.pkl")

run_HFO_detect.py at commit 12fe14a, under MIT · at the source

Overview

Authors: Sathwik Prathapagiri1, Jan Cimbalnik2,3, Jesús S García-Salinas1, Marina Galanina1, Lenka Jurkovicova2,3,4, Pavel Daniel3,4, Martin Kojan3,4, Robert Roman3,4, Martin Pail4, Wojciech Fortuna5, Monika Sluzewska-Niedzwiedz6, Pawel Tabakow5, Andrzej Czyzewski1, Milan Brazdil3,4, Michal T Kucewicz1,7
ORCID iDs: Sathwik Prathapagiri, Jesús S García-Salinas, Lenka Jurkovicova, Pavel Daniel, Martin Kojan, Robert Roman, Andrzej Czyzewski, Michal T Kucewicz
  1. Brain & Mind Electrophysiology Laboratory, BioTechMed Center, Department of Multimedia Systems, Faculty of Electronics, Telecommunications and Informatics, Gdańsk University of Technology, 80-233 Gdańsk, Poland
  2. Department of Biomedical Engineering & Neurology, St. Anne’s University Hospital in Brno, Brno, 60200 Czech Republic
  3. Central European Institute of Technology Masaryk University, Brno, Czech Republic
  4. Brno Epilepsy Center, Department of Neurology, St. Anne’s University Hospital, member of ERN-EpiCARE, Faculty of Medicine, Masaryk University, Pekarska 53, 602 00 Brno, Czech Republic
  5. Clinical Department of Neurosurgery, Faculty of Medicine, Wroclaw Medical University, Borowska 213, 50-556 Wroclaw, Poland
  6. Clinical Department of Neurology, Faculty of Medicine, Wroclaw Medical University, Borowska 213, 50-556 Wroclaw, Poland
  7. Department of Neurology, Mayo Clinic, 200 1st Street SW, Rochester, MN 55905 USA
Institutions: Gdańsk University of Technology (Poland); Central European Institute of Technology – Masaryk University (Czechia); University Hospital Brno (Czechia); St. Anne's University Hospital Brno (Czechia); Masaryk University (Czechia); Wroclaw Medical University (Poland); Mayo Clinic (United States)
Journal: Nature communications, volume 17, issue 1, article 3996
Dates: received 15 April 2025; accepted 2 March 2026; published online 15 March 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1038/s41467-026-70633-7 · PMID 41832174 · PMCID PMC13136343 · OpenAlex W7135423466
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: human (organism), epilepsy (population), cognitive (subfield)
Methods: Connectivity, Statistics, Machine learning, Spectral & time-frequency, Preprocessing, Evoked potentials, Smoothing, state filtering, decompositions, Physiology & signal measures
Keywords: Cognitive neuroscience, Learning and memory, Computational neuroscience
MeSH: Cerebral Cortex*, Memory*, Mental Recall*, Adult, Epilepsy, Female, Hippocampus, Humans, Male, Neurons (* major topic)
Topic: Neural dynamics and brain function (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Funding: Grantová Agentura České Republiky (Grant Agency of the Czech Republic) (21-44843L); Narodowe Centrum Nauki (National Science Centre) (2020/39/I/NZ4/02070)
Citations: not cited yet (Europe PMC); 84 references in the paper

Abstract

Oscillations in the high gamma and ripple frequency ranges are known to coordinate local hippocampal and neocortical neuronal assemblies during memory encoding and recall. Here, we explored spatiotemporal dynamics and the role of global coordination of these fast oscillatory discharges across the sensory and associational cortical areas in distinct phases of memory processing. Individual bursts of high frequency oscillations were detected in intracranial recordings from epilepsy patients remembering word lists for immediate free recall. We found constant coincident bursting across visual and higher order processing areas, peaking before recall and elevated during encoding of words. This global co-bursting was modulated by memory processing, engaged approximately half of the recorded electrode contact sites, and clustered into a sequence of multiple consecutive bursting events. Our results suggest a general role of global coincident high frequency oscillations in organizing large-scale information processing across the brain necessary especially, but not exclusively, for memory functions.

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 3 matches between paragraphs and lines of code.

brainandmindlab/coincident-hfo

License: MIT
State: the link answers, verified on 30 September 2026
Evidence: files inventoried
Commit: 12fe14abe4a45efa8c14fe30e5771f71b2d86669, 10 September 2025
Languages: Python (5)
Size: 7 files, 5 scripts
Software Heritage: not archived
Found in: “Code availability”
Holds: README
Not found: license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: NumPy (5 files), pandas (5 files), Matplotlib (3 files), Nilearn (3 files), SciPy (2 files), seaborn (2 files), statsmodels (2 files), PyBIDS (1 file), scikit-learn (1 file)
Availability: 1 check, the latest on 30 September 2026: the link answers
  • 30 September 2026: the link answers
6 files
At the source: gitlab.com/brainandmindlab/coincident-hfo

Code availability

The code for the analysis is available in the following GitLab repository: https://gitlab.com/brainandmindlab/coincident-hfo.

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

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;
  • 5 scripts, each with its path and the digest of its content;
  • 3 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 Human brain local field potential recordings during a battery of multilingual cognitive and eye-tracking tasks (v1) data used in this study are available in the EBRAINS database at 10.25493/4FZH-ZCG. All analyses were conducted on data from this openly accessible resource. The raw LFP data are protected and are not available in the public repository due to data privacy laws regarding human participants. The processed data used in this study are available at EBRAINS. Source data for all plots and figures generated in this study are provided as a Source Data file with this paper. Source data are provided with this paper.

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, 30 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 15 authors, 3 keywords, 10 MeSH terms, 2 funders, 78 references.

Cite

This paper

Prathapagiri, S., Cimbalnik, J., García-Salinas, J. S., Galanina, M., Jurkovicova, L., Daniel, P., Kojan, M., Roman, R., Pail, M., Fortuna, W., Sluzewska-Niedzwiedz, M., Tabakow, P., Czyzewski, A., Brazdil, M., & Kucewicz, M. T. (2026). Global coincident bursts of high frequency oscillations across the human cortex coordinate large-scale memory processing. Nature communications, 17(1), 3996. https://doi.org/10.1038/s41467-026-70633-7

BibTeX

@article{prathapagiri2026global,
author = {Prathapagiri, Sathwik and Cimbalnik, Jan and García-Salinas, Jesús S and Galanina, Marina and Jurkovicova, Lenka and Daniel, Pavel and Kojan, Martin and Roman, Robert and Pail, Martin and Fortuna, Wojciech and Sluzewska-Niedzwiedz, Monika and Tabakow, Pawel and Czyzewski, Andrzej and Brazdil, Milan and Kucewicz, Michal T},
title = {{Global coincident bursts of high frequency oscillations across the human cortex coordinate large-scale memory processing}},
journal = {Nature communications},
year = {2026},
month = mar,
volume = {17},
number = {1},
pages = {3996},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/s41467-026-70633-7},
url = {https://doi.org/10.1038/s41467-026-70633-7},
pmid = {41832174},
pmcid = {PMC13136343}
}

RIS

TY - JOUR
AU - Prathapagiri, Sathwik
AU - Cimbalnik, Jan
AU - García-Salinas, Jesús S
AU - Galanina, Marina
AU - Jurkovicova, Lenka
AU - Daniel, Pavel
AU - Kojan, Martin
AU - Roman, Robert
AU - Pail, Martin
AU - Fortuna, Wojciech
AU - Sluzewska-Niedzwiedz, Monika
AU - Tabakow, Pawel
AU - Czyzewski, Andrzej
AU - Brazdil, Milan
AU - Kucewicz, Michal T
TI - Global coincident bursts of high frequency oscillations across the human cortex coordinate large-scale memory processing
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/03/15
VL - 17
IS - 1
SP - 3996
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/s41467-026-70633-7
UR - https://doi.org/10.1038/s41467-026-70633-7
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41467-026-70633-7",
"type": "article-journal",
"title": "Global coincident bursts of high frequency oscillations across the human cortex coordinate large-scale memory processing",
"container-title": "Nature communications",
"author": [
{
"family": "Prathapagiri",
"given": "Sathwik"
},
{
"family": "Cimbalnik",
"given": "Jan"
},
{
"family": "García-Salinas",
"given": "Jesús S"
},
{
"family": "Galanina",
"given": "Marina"
},
{
"family": "Jurkovicova",
"given": "Lenka"
},
{
"family": "Daniel",
"given": "Pavel"
},
{
"family": "Kojan",
"given": "Martin"
},
{
"family": "Roman",
"given": "Robert"
},
{
"family": "Pail",
"given": "Martin"
},
{
"family": "Fortuna",
"given": "Wojciech"
},
{
"family": "Sluzewska-Niedzwiedz",
"given": "Monika"
},
{
"family": "Tabakow",
"given": "Pawel"
},
{
"family": "Czyzewski",
"given": "Andrzej"
},
{
"family": "Brazdil",
"given": "Milan"
},
{
"family": "Kucewicz",
"given": "Michal T"
}
],
"container-title-short": "Nat Commun",
"volume": "17",
"issue": "1",
"page": "3996",
"DOI": "10.1038/s41467-026-70633-7",
"PMID": "41832174",
"PMCID": "PMC13136343",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41467-026-70633-7",
"language": "en",
"issued": {
"date-parts": [
[
2026,
3,
15
]
]
}
}

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:10.1038/s41467-026-75345-6 [code]
Hippocampal ripples initiate cortical dimensionality expansion for memory retrieval.
Journal: Nature communications
In common: epilepsy, cognitive, 10 references
[2] doi:10.7554/elife.108023 [code]
Challenges in replay detection by TDLM in post-encoding resting state.
Journal: eLife
In common: statsmodels, seaborn, scikit-learn, 4 other tools, 6 references
[3] doi:10.1038/s42003-025-08618-3 [code]
Physical activity simultaneously improves working memory and ripple-spindle coupling
Journal: n/a
In common: cognitive, 8 references
[4] doi:10.1093/braincomms/fcag041
Exercise enhances hippocampal-cortical ripple interactions in the human brain.
Journal: Brain communications
In common: 8 references
[5] doi:10.1162/netn.a.544 [code]
Brain network reconfiguration during reward prediction error processing.
Journal: Network neuroscience (Cambridge, Mass.)
In common: PyBIDS, Nilearn, statsmodels, 6 other tools, cognitive, 1 reference
[6] doi:10.1038/s41593-026-02345-6 [code]
Human hippocampal ripples tune cortical responses based on predicted uncertainty.
Journal: Nature neuroscience
In common: cognitive, 7 references
[7] doi:10.64898/2026.08.18.26360725 [code]
Temporal pole blurring in hippocampal sclerosis reflects seizure-disrupted myelination
Journal: medRxiv (preprint)
In common: PyBIDS, Nilearn, statsmodels, 6 other tools, epilepsy
[8] doi:10.1007/s00234-026-04103-8 [code]
Enhanced detection of subtle cortical abnormalities in focal epilepsy using 7 T MRI surface-based models and graph neural networks.
Journal: Neuroradiology
In common: PyBIDS, Nilearn, statsmodels, 6 other tools, epilepsy
[9] doi:10.1016/j.isci.2026.117180 [code]
Developmental changes in similarity between neural representations of mental arithmetic and artificial neural networks.
Journal: iScience
In common: PyBIDS, Nilearn, statsmodels, 6 other tools
[10] doi:10.1073/pnas.2505464123 [code]
Opioid-specific brain connectivity dynamics distinguish analgesia from secondary effects: Studies in male mice.
Journal: Proceedings of the National Academy of Sciences of the United States of America
In common: PyBIDS, Nilearn, statsmodels, 6 other tools

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.