OSCR

Limitations of Variational Laplace-Based Dynamic Causal Modelling for Multistable Cortical Circuits.

Code ↔ Paper

8 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 8 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 › Statistical Analysis › Bayesian Model Selection ↔ Chaos/MATLAB_Codes/performDCMPEBAnalysis.m, lines 1–40 · score 0.60 · parametric empirical Bayes, winning model, PEB
  2. [2] § Methods › Statistical Analysis › Reconstructed Neural Activity from DCM Estimation ↔ Figures/FIG3/FIG3.m, lines 47–91 · score 0.58 · Lyapunov exponents, phase space, reconstructed, ground truth, period doubling
  3. [3] § Results › DCM Correctly Identifies Chaotic Model Architecture but not Connectivity Strengths ↔ Chaos/MATLAB_Codes/performDCMBMSBatch.m, the whole file · a weak match · score 0.57 · Bayesian model selection, forward backward, Free Energy, inference, BMS, lateral
  4. [4] § Results › DCM Correctly Identifies Chaotic Model Architecture but not Connectivity Strengths ↔ Single_point_steady_state/MATLAB_Codes/performDCMBMSBatch.m, the whole file · a weak match · score 0.57 · Bayesian model selection, forward backward, Free Energy, inference, BMS, lateral
  5. [5] § Methods › Cortical Columnar Neural Mass Models › Model with Bistable Fixed Points that Exhibits Decision-Making Behaviour ↔ Figures/FIG2/FIG2_G_I.m, lines 101–245 · score 0.55 · model exceedance probabilities, Corr, Err, lateral connection, ground truth, Acc
  6. [6] § Results › DCM Correctly Identifies Period-doubling Model Architecture but not Connectivity Strengths ↔ Figures/FIG3/FIG3.m, lines 47–91 · score 0.53 · Lyapunov exponent, phase space, reconstructed, ground truth, period doubling
  7. [7] § Methods › Cortical Columnar Neural Mass Models › Models with Period Doubling and Deterministic Chaos ↔ Chaos/MATLAB_Codes/NET_chaos.m, the whole file · a weak match · score 0.51 · feedback connections, membrane potential, chaos, inhibitory, excitatory
  8. [8] § Methods › DCM Estimation of Models ↔ Bistable_fixed_points/MATLAB_Codes/estimate_dcm_for_erp_unparallel.m, the whole file · a weak match · score 0.50 · forward backward connectivity, modulatory, lateral connectivity, SPM, ERP, bistable fixed point

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

MATLAB · 459 lines · 15 KB · GPL-3.0 · 2 matches

  1. figure('Units', 'normalized', 'Position', [0, 0, 1, 1]);
  3. % Colors for the bars
  4. colors_lateral = [0.5, 0.5, 0.5; % Gray for ground truth
  5. 0, 0, 1; % Blue for correct
  6. 1, 0, 0; % Red for error
  7. 0.5, 0, 0.5]; % Purple for all
  9. %%%%% Figure 2D %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  11. load("3D&E.mat")
  12. load("3g pd.mat")
  13. load('fig3_PD.mat');
  15. w=20;
  16. subplot(2,3,1)
  18. plot(ts,no1,'LineWidth',2,'color', colors_lateral(1, :))
  19. hold on
  20. plot(ts, noo1, 'LineWidth', 1, 'color', colors_lateral(2, :))
  21. xlim([0 0.75])
  23. ax = gca;
  24. ax.LineWidth = 2;
  25. set(gca, 'TickDir', 'out');
  26. % Add legend, labels, and formatting
  27. %legend('Column 1', 'Column 2','FontSize', 24)
  28. %xlabel('Time (s)', 'FontSize', 26)
  29. %ylabel('NMP of e_1 (a.u.)', 'FontSize', 26)
  30. set(gca, 'FontSize', w);
  31. set(gca, 'box', 'off')
  32. xticks([0,0.25,0.5,0.75])
  33. xticklabels({});
  34. h = ylabel('Normalised EPSP of {e}^1_1 (a.u.)', 'FontSize', w, 'FontWeight','bold');
  36. % Adjust the position of the x-label
  37. % currentPos = get(h, 'Position'); % Get current position [x, y, z]
  38. % newYPos = currentPos(2)-0.8 ; % Adjust the y-position (move up by increasing y-value)
  39. % set(h, 'Position', [currentPos(1), newYPos, currentPos(3)]); % Apply the new position
  41. %title('Period-doubling - \lambda = 0.03')
  42. % axis square
  43. legend('Ground-truth','Estimated','FontSize', w-2,'location','southeast')
  44. legend boxoff
  45. % t = title('Period-doubling');
  46. % set(t, 'Position', get(t, 'Position') + [0.6 0 0]); % Move left by 1 unit
  47. %%%%% Figure 2E %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  48. load("3D&E.mat")
  50. subplot(2,3,2)
  51. plot3(no2(fs:end)*nan,no3(fs:end)*nan,no1(fs:end)*nan,'LineWidth', 2,'LineStyle', '-','color', colors_lateral(1, :))
  52. hold on
  53. plot3(noo2(fs:end)*nan,noo3(fs:end)*nan,noo1(fs:end)*nan,'LineWidth', 2,'LineStyle', '-','color', colors_lateral(2, :))
  54. hold on
  55. plot3(no2(fs:end),no3(fs:end),no1(fs:end),'.','MarkerSize',6,'color', colors_lateral(1, :))
  56. hold on
  57. plot3(noo2(fs:end),noo3(fs:end),noo1(fs:end),'.','MarkerSize',6,'color', colors_lateral(2, :))
  59. ax = gca;
  60. ax.LineWidth = 2;
  61. set(gca, 'TickDir', 'out');
  63. % Add legend, labels, and formatting
  64. %legend('Column 1', 'Column 2','FontSize', 24)
  65. %xlabel('NMP of e_2 (a.u.)', 'FontSize', w)
  66. %ylabel('NMP of i (a.u.)', 'FontSize', w)
  67. %zlabel('NMP of e_1 (a.u.)', 'FontSize', w)
  68. set(gca, 'FontSize', w);
  69. set(gca, 'box', 'off')
  70. % axis square
  71. % legend('Ground-truth','Est.','FontSize', w-2,'location','southeast')
  72. % legend boxoff
  74. [~,lag] = phaseSpaceReconstruction(ch(1:end),[],3);
  75. eRange = [10 300];
  76. ly=lyapunovExponent(ch(fs:end),fs,lag,3,'ExpansionRange',eRange);
  77. %title(['Period-doubling - \lambda = ' num2str(ly)])
  78. xlim([0.0 1.0])
  79. ylim([0.0 1.0])
  80. zlim([0.0 1.0])
  82. text(0.25, 0.9,0.32, ['\lambda = ' num2str(ly, '%.2f')], 'BackgroundColor', 'none', 'EdgeColor', 'none', 'Margin', 5, 'FontSize', w-2,'color', colors_lateral(1, :));
  83. text(0.25, 0.9,0.15, ['\lambda = ' num2str(0, '%.2f')], 'BackgroundColor', 'none', 'EdgeColor', 'none', 'Margin', 5, 'FontSize', w-2,'color', colors_lateral(2, :));
  85. xticks([0,0.25,0.5,0.75,1])
  86. yticks([0,0.25,0.5,0.75,1])
  87. zticks([0,0.25,0.5,0.75,1])
  88. xticklabels({});
  89. zlabel('Normalised EPSP of {e}^1_1 (a.u.)', 'FontSize', w, 'FontWeight','bold');
  90. yticklabels({});
  91. %zticklabels({});
  93. %%%%% Figure 3F %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  94. h = subplot(2,3,3);
  95. pos = get(h, 'Position');
  97. % Delete the axes, since we'll replace it with two subplots
  98. delete(h);
  100. % Directories containing the BMS.mat files for different conditions
  101. BMSDir = 'C:\Users\aaa210\OneDrive - University of Sussex\Research projects\DCM validation\Different models test\Period Doubling';
  103. BMS_folder = fullfile(BMSDir, 'BMS');
  104. correct_trials_dir = BMS_folder;
  106. % Load BMS.mat files
  107. PD_trials_file = fullfile(correct_trials_dir,'BMS.mat');
  109. PD_trials_data = load(PD_trials_file);
  111. % Extract model names and exceedance probabilities
  112. model_names = PD_trials_data.BMS.DCM.rfx.family.names;
  113. PD_xp = PD_trials_data.BMS.DCM.rfx.model.xp;
  115. % Number of models
  116. num_models = length(model_names);
  118. % Combine data for plotting
  119. data_BMS = [PD_xp]';
  121. % Labels for x-axis
  122. x_labels_BMS = {'lat.', 'FB', 'FC'}; %model_names;
  124. % Colors for the bars
  125. colors = [0, 0, 0; % Red for error
  126. 0, 0, 1; % Blue for correct
  127. 0.5, 0, 0.5]; % Purple for all
  129. % Ground truth values
  130. ground_truth = [74.802466469706870, 74.802466469706870];
  132. % DCM Bayesian Averaging results for correct trials
  133. PD = [PD_trials_data.BMS.DCM.rfx.bma.mEp.A{3}(2,1), PD_trials_data.BMS.DCM.rfx.bma.mEp.A{3}(1,2)];
  136. % Take the absolute values of the connections
  137. PD = abs(PD);
  139. % Combine data for plotting
  140. data_lateral = [ground_truth; PD]';
  142. % Labels for x-axis
  143. x_labels_lateral = {'Column 1 to 2', 'Column 2 to 1'};
  147. % Define the gap between subplots
  148. gap_fraction = 0.25; % Fraction of the tile's width to use as gap (adjust as desired)
  149. gap = gap_fraction * pos(3); % Calculate the gap width
  151. % Adjust the widths of the subplots to accommodate the gap
  152. subplot_width = (pos(3) - gap) / 2;
  154. % Position for the first subplot (left column)
  155. pos1 = pos;
  156. pos1(3) = subplot_width; % Set the width of the first subplot
  157. % pos1(1) remains the same (left edge)
  159. % Position for the second subplot (right column)
  160. pos2 = pos;
  161. pos2(3) = subplot_width; % Set the width of the second subplot
  162. pos2(1) = pos(1) + subplot_width + gap; % Move it to the right by one subplot width plus the gap
  164. % Create the first subplot (left column)
  165. ax3a = axes('Position', pos1);
  166. axes(ax3a); % Explicitly make ax3a the active axis
  167. hold on; % Ensure "hold on" applies to ax3a
  169. b1 = bar(data_BMS, 'grouped');
  171. % Set colors for each bar group
  172. for k = 1:length(b1)
  173. b1(k).FaceColor = colors(k, :);
  174. end
  176. % Set x-axis labels
  177. set(gca, 'XTickLabel', x_labels_BMS, 'XTick', 1:numel(x_labels_BMS), 'XTickLabel', '', 'XTickLabelMode', 'manual');
  179. % Remove x-ticks but keep x-tick labels
  180. ax1 = gca;
  181. ax1.XAxis.TickLength = [0 0];
  183. % Set y-axis limits
  184. ylim([0, 1]);
  186. set(gca, 'TickDir', 'out');
  187. % Add legend, labels, and formatting
  188. set(gca, 'FontSize', w);
  189. set(gca, 'box', 'off');
  190. ax = gca;
  191. ax.LineWidth = 2;
  192. hold off;
  193. set(ax3a, 'FontSize', w);
  194. ylabel('Model Exceedance Probability', 'FontSize', w, 'FontWeight','bold');
  195. ylim([0 0.65])
  197. ax = ax3a;
  198. ax.LineWidth = 2;
  199. set(ax3a, 'TickDir', 'out');
  200. set(ax3a, 'box', 'off');
  202. % Create the second subplot (right column)
  203. ax3b = axes('Position', pos2);
  204. axes(ax3b); % Explicitly make ax3b the active axis
  205. hold on; % Ensure "hold on" applies to ax3b
  207. b2 = bar(data_lateral);
  209. % Set colors for each bar group
  210. for k = 1:length(b2)
  211. b2(k).FaceColor = colors_lateral(k, :);
  212. end
  214. % Set x-axis labels
  215. set(gca, 'XTickLabel', x_labels_lateral, 'XTick', 1:numel(x_labels_lateral), 'YScale', 'log', 'XTickLabel', '', 'XTickLabelMode', 'manual');
  217. % Remove x-ticks but keep x-tick labels
  218. ax2 = gca;
  219. ax2.XAxis.TickLength = [0 0];
  220. % Set y-axis limits
  221. % ylim([-0.001, ground_truth(1) * 1.05]);
  222. set(ax3b, 'FontSize', w);
  223. ylabel(ax3b, 'Lateral Connection (log scale)', 'FontSize', w, 'FontWeight','bold'); % Increased y-label font size
  224. ax = ax3b;
  225. ax.LineWidth = 2;
  226. set(ax3b, 'TickDir', 'out');
  228. set(ax3b, 'box', 'off');
  230. %%%%% Figure 2G %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  231. load("3G&H.mat")
  232. load("3g ch.mat")
  233. load("fig3_chaos.mat")
  235. subplot(2,3,4)
  237. plot(ts,no1,'LineWidth',2,'color', colors_lateral(1, :))
  238. hold on
  239. plot(ts, noo1, 'LineWidth', 1, 'color', colors_lateral(2, :))
  241. ax = gca;
  242. ax.LineWidth = 2;
  243. set(gca, 'TickDir', 'out');
  244. % Add legend, labels, and formatting
  245. %legend('Column 1', 'Column 2','FontSize', 24)
  246. set(gca, 'FontSize', w);
  247. xlabel('Time (s)', 'FontSize', w, 'FontWeight','bold')
  248. h = ylabel('Normalised EPSP of {e}^1_1 (a.u.)', 'FontSize', w, 'FontWeight','bold');
  250. set(gca, 'box', 'off')
  252. % axis square
  253. %legend('O1','FontSize', 16)
  254. %legend boxoff
  255. xlim([0 0.75])
  256. xticks([0,0.25,0.5,0.75])
  257. yticks([0,0.2,0.4,0.6,0.8,1])
  258. % t = title('Chaos');
  259. % set(t, 'Position', get(t, 'Position') + [0.6 0 0]); % Move left by 1 unit
  260. %%%%% Figure 2H %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  261. load("3G&H.mat")
  263. subplot(2,3,5)
  264. plot3(no2(fs:end),no3(fs:end),no1(fs:end),'.','MarkerSize',6,'color', colors_lateral(1, :))
  265. hold on
  266. plot3(noo2(fs:end),noo3(fs:end),noo1(fs:end),'.','MarkerSize',6,'color', colors_lateral(2, :))
  267. ax = gca;
  268. ax.LineWidth = 2;
  269. set(gca, 'TickDir', 'out');
  271. % Add legend, labels, and formatting
  272. %legend('Column 1', 'Column 2','FontSize', 24)
  274. %zlabel('NMP of e_1 (a.u.)', 'FontSize', w)
  275. set(gca, 'FontSize', w);
  276. set(gca, 'box', 'off')
  277. % axis square
  279. %lyapExp = lyapunovExponent(ch(1:end),fs)
  281. [~,lag] = phaseSpaceReconstruction(ch(1:end),[],3);
  282. eRange = [50 400];
  283. ly=lyapunovExponent(ch(fs:end),fs,lag,3,'ExpansionRange',eRange);
  284. %title(['Chaos - \lambda = ' num2str(ly)])
  285. xlim([0.0 1.0])
  286. ylim([0.0 1.0])
  287. zlim([0.0 1.0])
  288. xticks([0,0.25,0.5,0.75,1])
  289. yticks([0,0.25,0.5,0.75,1])
  290. zticks([0,0.25,0.5,0.75,1])
  291. zlabel('Normalised EPSP of {e}^1_1 (a.u.)', 'FontSize', w, 'FontWeight','bold');
  292. hy = ylabel('Norm. IPSP of {i}^1 (a.u.)', 'FontSize', 0.9*w, 'FontWeight','bold');
  293. set(hy, 'Rotation', -30); % Rotate Y-label by -30 degrees
  294. posy = get(hy, 'Position'); % Get current position
  297. hx = xlabel('Norm. EPSP of e^1_2 (a.u.)', 'FontSize', 0.9*w, 'FontWeight','bold');
  298. set(hx, 'Rotation', 19); % Rotate Z-label by 30 degrees
  299. posx = get(hx, 'Position'); % Get current position
  300. text(0.25, 0.9,0.30, ['\lambda = ' num2str(ly, '%.2f')], 'BackgroundColor', 'none', 'EdgeColor', 'none', 'Margin', 5, 'FontSize', w-2,'color', colors_lateral(1, :));
  302. text(0.25, 0.9,0.15, ['\lambda = ' num2str(0, '%.2f')], 'BackgroundColor', 'none', 'EdgeColor', 'none', 'Margin', 5, 'FontSize', w-2,'color', colors_lateral(2, :));
  304. % Adjust Y-label position
  305. xpos = posy(1)+ 0.38; % Move left (decrease X value)
  306. ypos = posy(2) + 0.59; % Move up (increase Y value)
  307. set(hy, 'Position', [xpos ypos posy(3)]); % Apply new position
  309. % Adjust X-label position
  310. xposx = posx(1)+ 0.24; % Move left (decrease X value)
  311. yposx = posx(2) + 0.17; % Move up (increase Y value)
  312. set(hx, 'Position', [xposx yposx posx(3)]); % Apply new position
  314. %%%%% Figure 3F %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  315. h = subplot(2,3,6);
  316. pos = get(h, 'Position');
  318. % Delete the axes, since we'll replace it with two subplots
  319. delete(h);
  321. % Directories containing the BMS.mat files for different conditions
  322. BMSDir = 'C:\Users\aaa210\OneDrive - University of Sussex\Research projects\DCM validation\Different models test\Chaos';
  324. BMS_folder = fullfile(BMSDir, 'BMS');
  325. correct_trials_dir = BMS_folder;
  327. % Load BMS.mat files
  328. PD_trials_file = fullfile(correct_trials_dir,'BMS.mat');
  330. PD_trials_data = load(PD_trials_file);
  332. % Extract model names and exceedance probabilities
  333. model_names = PD_trials_data.BMS.DCM.rfx.family.names;
  334. PD_xp = PD_trials_data.BMS.DCM.rfx.model.xp;
  336. % Number of models
  337. num_models = length(model_names);
  339. % Combine data for plotting
  340. data_BMS = [PD_xp]';
  342. % Labels for x-axis
  343. x_labels_BMS = {'lat.', 'FB', 'FC'}; %model_names;
  345. % Colors for the bars
  346. colors = [0, 0, 0; % Red for error
  347. 0, 0, 1; % Blue for correct
  348. 0.5, 0, 0.5]; % Purple for all
  350. % Ground truth values
  351. ground_truth = [74.802466469706870, 74.802466469706870];
  353. % DCM Bayesian Averaging results for correct trials
  354. PD = [PD_trials_data.BMS.DCM.rfx.bma.mEp.A{3}(2,1), PD_trials_data.BMS.DCM.rfx.bma.mEp.A{3}(1,2)];
  357. % Take the absolute values of the connections
  358. PD = abs(PD);
  360. % Combine data for plotting
  361. data_lateral = [ground_truth; PD]';
  363. % Labels for x-axis
  364. x_labels_lateral = {'Column 1 to 2', 'Column 2 to 1'};
  366. % Colors for the bars
  367. colors_lateral = [0.5, 0.5, 0.5; % Gray for ground truth
  368. 0, 0, 1; % Blue for correct
  369. 1, 0, 0; % Red for error
  370. 0.5, 0, 0.5]; % Purple for all
  372. % Define the gap between subplots
  373. % gap_fraction = 0.1; % Fraction of the tile's width to use as gap (adjust as desired)
  374. gap = gap_fraction * pos(3); % Calculate the gap width
  376. % Adjust the widths of the subplots to accommodate the gap
  377. subplot_width = (pos(3) - gap) / 2;
  379. % Position for the first subplot (left column)
  380. pos1 = pos;
  381. pos1(3) = subplot_width; % Set the width of the first subplot
  382. % pos1(1) remains the same (left edge)
  384. % Position for the second subplot (right column)
  385. pos2 = pos;
  386. pos2(3) = subplot_width; % Set the width of the second subplot
  387. pos2(1) = pos(1) + subplot_width + gap; % Move it to the right by one subplot width plus the gap
  389. % Create the first subplot (left column)
  390. ax3a = axes('Position', pos1);
  391. axes(ax3a); % Explicitly make ax3a the active axis
  392. hold on; % Ensure "hold on" applies to ax3a
  394. b1 = bar(data_BMS, 'grouped');
  396. % Set colors for each bar group
  397. for k = 1:length(b1)
  398. b1(k).FaceColor = colors(k, :);
  399. end
  401. % Set x-axis labels
  402. set(gca, 'XTickLabel', x_labels_BMS, 'XTick', 1:numel(x_labels_BMS));
  404. % Remove x-ticks but keep x-tick labels
  405. ax1 = gca;
  406. ax1.XAxis.TickLength = [0 0];
  408. % Set y-axis limits
  409. ylim([0, 1]);
  411. set(gca, 'TickDir', 'out');
  412. % Add legend, labels, and formatting
  413. set(gca, 'FontSize', w);
  414. set(gca, 'box', 'off');
  415. ax = gca;
  416. ax.LineWidth = 2;
  417. hold off;
  419. set(ax3a, 'FontSize', w);
  420. ylabel('Model Exceedance Probability', 'FontSize', w, 'FontWeight','bold');
  421. ylim([0 0.65])
  423. ax = ax3a;
  424. ax.XAxis.FontWeight = 'bold'; % Set x-axis tick labels to bold
  425. ax.XAxis.FontSize = w; % Set x-axis tick labels font size to w
  426. ax.LineWidth = 2;
  427. set(ax3a, 'TickDir', 'out');
  428. set(ax3a, 'box', 'off');
  430. % Create the second subplot (right column)
  431. ax3b = axes('Position', pos2);
  432. axes(ax3b); % Explicitly make ax3b the active axis
  433. hold on; % Ensure "hold on" applies to ax3b
  435. b2 = bar(data_lateral);
  437. % Set colors for each bar group
  438. for k = 1:length(b2)
  439. b2(k).FaceColor = colors_lateral(k, :);
  440. end
  442. % Set x-axis labels
  443. set(gca, 'XTickLabel', x_labels_lateral, 'XTick', 1:numel(x_labels_lateral), 'YScale', 'log');
  445. % Remove x-ticks but keep x-tick labels
  446. ax2 = gca;
  447. ax2.XAxis.TickLength = [0 0];
  448. % Set y-axis limits
  449. % ylim([log10(0.001), log10(ground_truth(1) * 1.05)]);
  451. set(ax3b, 'FontSize', w);
  452. ylabel(ax3b, 'Lateral Connection (log scale)', 'FontSize', w, 'FontWeight','bold'); % Increased y-label font size
  453. ax = ax3b;
  454. ax.LineWidth = 2;
  455. ax.XAxis.FontWeight = 'bold'; % Set x-axis tick labels to bold
  456. ax.XAxis.FontSize = w; % Set x-axis tick labels font size to w
  457. set(ax3b, 'TickDir', 'out');
  459. set(ax3b, 'box', 'off');

FIG3.m at commit c94e7dd, under GPL-3.0 · at the source

Overview

Authors: Abdoreza Asadpour1,2, Amin Azimi1,3, KongFatt Wong-Lin1
ORCID iDs: Abdoreza Asadpour, KongFatt Wong-Lin
  1. Intelligent Systems Research Centre, School of Computing, Engineering and Intelligent Systems, Ulster University, Magee campus, Derry~Londonderry, Northern Ireland UK
  2. Sussex Neuroscience, School of Life Sciences, University of Sussex, Brighton, UK
  3. School of Psychology, Manchester Metropolitan University, Manchester, UK
Institutions: University of Ulster (United Kingdom); University of Sussex (United Kingdom); Ulster University, Derry-Londonderry Campus (United Kingdom); Manchester Metropolitan University (United Kingdom)
Journal: Neuroinformatics, volume 24, issue 2, article 17
Dates: received 29 September 2025; accepted 18 November 2025; published online 1 April 2026; in print 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1007/s12021-025-09759-w · PMID 41920413 · PMCID PMC13043521 · OpenAlex W7147233438
Open access: hybrid, a free copy (OpenAlex)
Status: code verified
Categories: human (organism), computational (subfield)
Methods: Connectivity, Evoked potentials, Single-unit activity, calcium imaging, Machine learning
Keywords: Dynamic causal modelling DCM, Bayesian model averaging, Cortical column neural mass model, Multistable dynamical states, Oscillations, Chaos, Decision making
MeSH: Cerebral Cortex*, Models, Neurological*, Nerve Net*, Neurons*, Nonlinear Dynamics*, Animals, Bayes Theorem, Computer Simulation, Humans, Neural Pathways (* major topic)
Topic: Functional Brain Connectivity Studies (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Funding: HSC R&D (STL/5540/19); Medical Research Council (MC_PC_20020); HSC R&D (STL/5540/19)
Citations: not cited yet (Europe PMC); 65 references in the paper

Abstract

Dynamic causal modelling (DCM) is widely used to infer effective connectivity from neuroimaging data. However, its applicability to neural systems with complex, multistable dynamics remains uncertain—particularly when using the standard estimation approach based on variational Bayesian inference under the Laplace approximation. To investigate this limitation, we constructed biologically grounded cortical columnar neural mass models exhibiting three distinct multistable regimes: bistable fixed points associated with decision-making, coexisting oscillatory states through period-doubling bifurcations, and deterministic chaotic dynamics. These models were used to simulate local field potentials, which served as inputs to DCM. Bayesian model selection successfully identified the correct model architecture in all cases. However, Bayesian model averaging of the winning models failed to accurately estimate extrinsic connectivity parameters, leading to substantial discrepancies between the dynamics of the reconstructed and ground-truth systems. These results suggest that even when model selection is accurate, parameter estimation can break down under complex dynamics. Compared to previous applications of DCM to simpler neural systems, our study highlights significant limitations in its ability to capture the structure of multistable and globally nonlinear dynamics. We conclude that caution is warranted when applying variational Laplace-based DCM procedures to experimental paradigms involving bifurcations, chaotic trajectories, or other forms of dynamical complexity.

Supplementary Information: The online version contains supplementary material available at 10.1007/s12021-025-09759-w.

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 8 matches between paragraphs and lines of code.

asadpouretal/DCM_Multistabilityor

License: GPL-3.0
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Commit: c94e7dd1a84c0e720e1403b24a76fd6661ab5ec8, 14 April 2026
Languages: MATLAB (120)
Size: 146 files, 120 scripts
Software Heritage: not archived
Found in: “Data Availability”
Holds: README, license file
Not found: CITATION.cff, environment file, tests, continuous integration, documentation
Tools: SPM (54 files), Parallel Computing Toolbox (22 files), Statistics and Machine Learning Toolbox (6 files), Signal Processing Toolbox (3 files)
Availability: 1 check, the latest on 28 September 2026: the link answers
  • 28 September 2026: the link answers
122 files
At the source: github.com/asadpouretal/DCM_Multistabilityor

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;
  • 120 scripts, each with its path and the digest of its content;
  • 8 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

No dataset and no data link were found in the paper.

Data Availability Statement

In numerical simulations for designing ground truth models before implementing DCM, the deterministic nonlinear differential equations for period-doubling and chaotic dynamics were solved using the Euler forward method, while stochastic differential equations for noisy bistable decision-making were solved using the Euler-Maruyama numerical method (Higham, 2001). All simulations were performed in MATLAB R2023b software with a time step of 1 millisecond. Sufficient accuracy for these differential equations is achieved with Euler-based methods as long as the simulation time step used is sufficiently small, and it was verified that smaller time steps did not affect the results. For simulations using stochastic differential equations, the same set of pseudo-random seed was used when comparing between the generated data of the ground-truth models and corresponding DCM-estimated models, akin to Lenfesty et al. (2025). For DCM estimation, BMS, and BMA, computations were performed using MATLAB R2022a software via the Northern Ireland High-Performance Computing (NI-HPC) facility (https://www.ni-hpc.ac.uk).

The source code for our implementation and generated data are available and can be accessed at the following GitHub repository: https://github.com/asadpouretal/DCM_Multistabilityor. This repository contains all the necessary scripts, models, and instructions to reproduce the results presented in this study, ensuring transparency and facilitating further research.

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 2, 28 September 2026

  • Publisher: n/a → Springer Science+Business Media

Version 1, 28 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 3 authors, 7 keywords, 10 MeSH terms, 3 funders, 63 references.

Cite

This paper

Asadpour, A., Azimi, A., & Wong-Lin, K. (2026). Limitations of Variational Laplace-Based Dynamic Causal Modelling for Multistable Cortical Circuits. Neuroinformatics, 24(2), 17. https://doi.org/10.1007/s12021-025-09759-w

BibTeX

@article{asadpour2026limitations,
author = {Asadpour, Abdoreza and Azimi, Amin and Wong-Lin, KongFatt},
title = {{Limitations of Variational Laplace-Based Dynamic Causal Modelling for Multistable Cortical Circuits}},
journal = {Neuroinformatics},
year = {2026},
month = apr,
volume = {24},
number = {2},
pages = {17},
publisher = {Springer Science+Business Media},
issn = {1539-2791},
doi = {10.1007/s12021-025-09759-w},
url = {https://doi.org/10.1007/s12021-025-09759-w},
pmid = {41920413},
pmcid = {PMC13043521}
}

RIS

TY - JOUR
AU - Asadpour, Abdoreza
AU - Azimi, Amin
AU - Wong-Lin, KongFatt
TI - Limitations of Variational Laplace-Based Dynamic Causal Modelling for Multistable Cortical Circuits
T2 - Neuroinformatics
J2 - Neuroinformatics
PY - 2026
DA - 2026/04/01
VL - 24
IS - 2
SP - 17
SN - 1539-2791
PB - Springer Science+Business Media
DO - 10.1007/s12021-025-09759-w
UR - https://doi.org/10.1007/s12021-025-09759-w
LA - en
ER -

CSL-JSON

{
"id": "10.1007/s12021-025-09759-w",
"type": "article-journal",
"title": "Limitations of Variational Laplace-Based Dynamic Causal Modelling for Multistable Cortical Circuits",
"container-title": "Neuroinformatics",
"author": [
{
"family": "Asadpour",
"given": "Abdoreza"
},
{
"family": "Azimi",
"given": "Amin"
},
{
"family": "Wong-Lin",
"given": "KongFatt"
}
],
"container-title-short": "Neuroinformatics",
"volume": "24",
"issue": "2",
"page": "17",
"DOI": "10.1007/s12021-025-09759-w",
"PMID": "41920413",
"PMCID": "PMC13043521",
"ISSN": "1539-2791",
"publisher": "Springer Science+Business Media",
"URL": "https://doi.org/10.1007/s12021-025-09759-w",
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
1
]
]
}
}

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:10.1162/imag.a.1234 [code]
Multiscale parcellation of dynamic causal models of the brain.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: SPM, Statistics and Machine Learning Toolbox, computational, 6 references
[2] doi:10.1038/s41467-026-71414-y [code]
Bed nucleus of the stria terminalis connectivity during food cue and taste processing under stress.
Journal: Nature communications
In common: Parallel Computing Toolbox, SPM, Signal Processing Toolbox, 1 other tool, 3 references
[3] doi:10.1038/s41593-026-02205-3 [code]
Competitive interactions shape mammalian brain network dynamics and computation.
Journal: Nature neuroscience
In common: Parallel Computing Toolbox, SPM, Signal Processing Toolbox, 1 other tool, 2 references
[4] doi:10.1371/journal.pcbi.1014222 [code]
Neural population models for EEG: From Canonical models to alternative model structures.
Journal: PLoS computational biology
In common: computational, 5 references
[5] doi:10.1162/imag.a.1250 [code]
Dynamics-informed priors (DIP) for neural mass modelling.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: SPM, Statistics and Machine Learning Toolbox, 4 references
[6] doi:10.1038/s42003-026-10276-y [code]
The cellular correlates and adolescent reorganisation of cortical myelination networks in the common marmoset.
Journal: Communications biology
In common: Parallel Computing Toolbox, SPM, Signal Processing Toolbox, 1 other tool, 1 reference
[7] doi:10.1093/braincomms/fcag334 [code]
Functional and effective connectivity alterations relate to cognitive performance in aquaporin-4 antibody-positive neuromyelitis optica spectrum disorder.
Journal: Brain communications
In common: SPM, Statistics and Machine Learning Toolbox, 3 references
[8] doi:10.1162/imag.a.1130 [code]
Computational parametric mapping of functional neuroimaging data.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: SPM, Statistics and Machine Learning Toolbox, 3 references
[9] doi:10.1038/s41586-026-10910-z [code]
Psychedelics align brain activity with context.
Journal: Nature
In common: SPM, Statistics and Machine Learning Toolbox, 3 references
[10] doi:10.1162/netn_a_00453 [code]
Joint estimation of source dynamics and interactions from MEG data
Journal: n/a
In common: Parallel Computing Toolbox, Signal Processing Toolbox, Statistics and Machine Learning Toolbox, 2 references

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.