OSCR

Hypercapnia dissociates neuronal and hemodynamic responses impairing neurovascular coupling and functional brain connectivity.

Code ↔ Paper

1 match 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 1 match
  1. [1] § Methods › Dynamic functional connectivity analysis ↔ dfc_state_analysis.m, lines 864–924 · score 0.56 · cross correlation, state transition, lag, duration, dFC, window

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 · 1,383 lines · 67 KB · CC-BY-4.0 · 1 match

  1. close all; clc; clearvars;
  3. %% dFC state analysis: FL (10 Hz, 1 Hz), neural, and fMRI
  4. %
  5. % Required files in the working directory:
  6. % dFC_10Hz.mat
  7. % dFC_1Hz.mat
  8. % dFC_fMRI.mat
  9. % dFC_neural.mat
  11. %
  12. % Expected data dimensions: [num_regions x num_regions x num_windows x num_animals]
  14. %% Load data
  15. load('dFC_10Hz.mat');
  16. load('dFC_neural.mat');
  17. load('dFC_fMRI.mat');
  18. load('dFC_1Hz.mat');
  20. % Optional colormaps
  21. % cm_turbo_custom = load_colormap_safe('turbo_custom.mat', 'turbo_custom', parula(256));
  22. % cm_turbo_custom_endWhite = load_colormap_safe('turbo_custom_endWhite.mat', 'turbo_custom_endWhite', parula(256));
  23. % cm_custom_blueToRed = load_colormap_safe('custom_blueToRed.mat', 'custom_blueToRed', parula(256));
  24. % cm_custom_blue2red_2 = load_colormap_safe('custom_blue2red_2.mat', 'custom_blue2red_2', parula(256));
  26. rng('threefry');
  28. %% Parameters
  29. num_regions = size(dFC_10Hz, 1);
  30. num_animals = size(dFC_10Hz, 4);
  32. num_replicates = 20;
  33. max_clusters = 10;
  35. window_size_10Hz = 300;
  36. step_size_10Hz = 10;
  37. window_size_fMRI = 30;
  38. step_size_fMRI = 1;
  39. window_size_1Hz = 30;
  40. step_size_1Hz = 1;
  42. sampling_rate_10Hz = 10;
  43. sampling_rate_fMRI = 1;
  44. sampling_rate_1Hz = 1;
  46. num_windows_neural = size(dFC_neural, 3);
  47. num_windows_fMRI = size(dFC_fMRI, 3);
  48. num_windows_10Hz = size(dFC_10Hz, 3);
  49. num_windows_1Hz = size(dFC_1Hz, 3);
  51. optimal_num_clusters_combined = 6;
  52. optimal_num_clusters_fMRI = 5;
  54. % Region re-ordering for visualization
  55. new_order = [1:10, 11, 23, 12:15, 24, 16, 17, 18, 19, 20, 21, 22, 25, ...
  56. 26:35, 36, 48, 37:40, 49, 41, 42, 43, 44, 45, 46, 47, 50];
  58. %% CO2 timing
  59. % Start indices are subject-specific only for subject 3.
  60. co2_on_periods = [175, 535];
  61. co2_on_periods_sub3 = [235, 595];
  62. co2_duration = 165; % chosen analysis duration after accounting for windowing
  63. co2_on_duration_fMRI = co2_duration;
  65. co2_on_indices_10Hz = round(co2_on_periods * sampling_rate_10Hz / step_size_10Hz);
  66. co2_on_indices_10Hz_sub3 = round(co2_on_periods_sub3 * sampling_rate_10Hz / step_size_10Hz);
  67. co2_on_indices_fMRI = round(co2_on_periods * sampling_rate_fMRI / step_size_fMRI);
  68. co2_on_indices_fMRI_sub3 = round(co2_on_periods_sub3 * sampling_rate_fMRI / step_size_fMRI);
  69. co2_on_indices_1Hz = round(co2_on_periods * sampling_rate_1Hz / step_size_1Hz);
  70. co2_on_indices_1Hz_sub3 = round(co2_on_periods_sub3 * sampling_rate_1Hz / step_size_1Hz);
  72. %% Reshape data for clustering
  73. hemodynamic_10Hz_data = reshape(dFC_10Hz, num_regions * num_regions, [])';
  74. hemodynamic_1Hz_data = reshape(dFC_1Hz, num_regions * num_regions, [])';
  75. combined_hemodynamic_data = [hemodynamic_10Hz_data; hemodynamic_1Hz_data];
  77. fMRI_data = reshape(dFC_fMRI, num_regions * num_regions, [])';
  78. neural_data = reshape(dFC_neural, num_regions * num_regions, [])';
  80. %% K-means clustering
  81. rng('philox');
  82. [idx_combined, centroids_combined] = kmeans(combined_hemodynamic_data, ...
  83. optimal_num_clusters_combined, 'Replicates', num_replicates);
  85. idx_10Hz = reshape(idx_combined(1:size(hemodynamic_10Hz_data, 1)), num_windows_10Hz, num_animals);
  86. idx_1Hz = reshape(idx_combined(size(hemodynamic_10Hz_data, 1) + 1:end), num_windows_1Hz, num_animals);
  88. rng('philox');
  89. [idx_fMRI, centroids_fMRI] = kmeans(fMRI_data, optimal_num_clusters_fMRI, ...
  90. 'Replicates', num_replicates);
  91. idx_fMRI = reshape(idx_fMRI, num_windows_fMRI, num_animals);
  93. %% Remove states observed in only one subject, then fill missing labels by nearest centroid
  94. [idx_10Hz, removed_states_10Hz] = remove_single_subject_states(idx_10Hz);
  95. [idx_1Hz, removed_states_1Hz] = remove_single_subject_states(idx_1Hz);
  96. [idx_fMRI, removed_states_fMRI] = remove_single_subject_states(idx_fMRI);
  98. if ~isempty(removed_states_10Hz)
  99. fprintf('FL 10 Hz states removed for single-subject occurrence: %s\n', mat2str(removed_states_10Hz));
  100. end
  101. if ~isempty(removed_states_1Hz)
  102. fprintf('FL 1 Hz states removed for single-subject occurrence: %s\n', mat2str(removed_states_1Hz));
  103. end
  104. if ~isempty(removed_states_fMRI)
  105. fprintf('fMRI states removed for single-subject occurrence: %s\n', mat2str(removed_states_fMRI));
  106. end
  108. distance_matrix_combined = pdist2(centroids_combined, centroids_combined);
  109. distance_matrix_fMRI = pdist2(centroids_fMRI, centroids_fMRI);
  111. idx_10Hz = replace_nan_states_with_closest(idx_10Hz, distance_matrix_combined);
  112. idx_1Hz = replace_nan_states_with_closest(idx_1Hz, distance_matrix_combined);
  113. idx_fMRI = replace_nan_states_with_closest(idx_fMRI, distance_matrix_fMRI);
  115. %% Plot identified state centroids and mean neural dFC
  116. neural_plot = mean(dFC_neural, [3 4]);
  117. neural_2plot = neural_plot(new_order, new_order);
  119. figure('Position', [100 100 500 900]);
  120. for k = 1:optimal_num_clusters_combined
  121. FL_plot = reshape(centroids_combined(k, :), num_regions, num_regions);
  122. FL_2plot = FL_plot(new_order, new_order);
  124. subplot(ceil(optimal_num_clusters_combined / 2), 2, k);
  125. imagesc(FL_2plot);
  126. axis square;
  127. caxis([-1 1]);
  128. colormap(parula);
  129. title(sprintf('Identified State %d - FL', k));
  130. xlabel('Region'); ylabel('Region');
  131. end
  133. figure('Position', [100 100 500 900]);
  134. for k = 1:optimal_num_clusters_fMRI
  135. fMRI_plot = reshape(centroids_fMRI(k, :), num_regions, num_regions);
  136. fMRI_2plot = fMRI_plot(new_order, new_order);
  138. subplot(ceil(optimal_num_clusters_fMRI / 2), 2, k);
  139. imagesc(fMRI_2plot);
  140. axis square;
  141. caxis([-0.4 0.4]);
  142. colorbar;
  143. colormap(parula);
  144. title(sprintf('Identified State %d - fMRI', k));
  145. xlabel('Region'); ylabel('Region');
  146. end
  148. figure;
  149. imagesc(neural_2plot);
  150. axis square;
  151. colorbar;
  152. caxis([-1 1]);
  153. colormap(parula);
  154. title('Mean neural dFC');
  155. xlabel('Region'); ylabel('Region');
  157. %% Compare hemodynamic state centroids with neural dFC windows
  158. state_neural_similarity_combined = nan(optimal_num_clusters_combined, num_animals);
  159. state_neural_similarity_fMRI = nan(optimal_num_clusters_fMRI, num_animals);
  161. for k = 1:optimal_num_clusters_combined
  162. hemo_state = centroids_combined(k, :);
  163. for animal = 1:num_animals
  164. neural_dFC = neural_data((animal - 1) * num_windows_neural + (1:num_windows_neural), :);
  165. state_neural_similarity_combined(k, animal) = mean(corr(hemo_state', neural_dFC', 'Rows', 'pairwise'));
  166. end
  167. end
  169. for k = 1:optimal_num_clusters_fMRI
  170. hemo_state = centroids_fMRI(k, :);
  171. for animal = 1:num_animals
  172. neural_dFC = neural_data((animal - 1) * num_windows_neural + (1:num_windows_neural), :);
  173. state_neural_similarity_fMRI(k, animal) = mean(corr(hemo_state', neural_dFC', 'Rows', 'pairwise'));
  174. end
  175. end
  177. figure;
  178. subplot(2, 1, 1);
  179. imagesc(state_neural_similarity_combined);
  180. colorbar;
  181. colormap(cm_turbo_custom);
  182. clim([0.25 0.85]);
  183. title('Similarity between FL states and neural dFC windows');
  184. xlabel('Animal'); ylabel('FL state');
  186. subplot(2, 1, 2);
  187. imagesc(state_neural_similarity_fMRI);
  188. colorbar;
  189. colormap(cm_turbo_custom);
  190. clim([0.3 0.6]);
  191. title('Similarity between fMRI states and neural dFC windows');
  192. xlabel('Animal'); ylabel('fMRI state');
  194. add_jitter = @(values, scale) values + scale * (rand(size(values)) - 0.5);
  195. jitter_scale = 0.1;
  196. colors = lines(num_animals);
  198. figure('Position', [100, 100, 1000, 800]);
  199. subplot(2, 1, 1); hold on;
  200. for animal = 1:num_animals
  201. y_values = state_neural_similarity_combined(:, animal);
  202. x_values = 1:length(y_values);
  203. scatter(add_jitter(x_values, jitter_scale), y_values, 50, colors(animal, :), ...
  204. 'filled', 'DisplayName', sprintf('Animal %d', animal));
  205. end
  206. mean_values_combined = mean(state_neural_similarity_combined, 2, 'omitnan');
  207. for state = 1:length(mean_values_combined)
  208. plot([state - 0.4, state + 0.4], [mean_values_combined(state), mean_values_combined(state)], 'k-', 'LineWidth', 2);
  209. end
  210. title('FL state similarity to neural dFC windows');
  211. xlabel('FL state'); ylabel('Similarity'); legend('show', 'Location', 'eastoutside'); grid on; hold off;
  213. subplot(2, 1, 2); hold on;
  214. for animal = 1:num_animals
  215. y_values = state_neural_similarity_fMRI(:, animal);
  216. x_values = 1:length(y_values);
  217. scatter(add_jitter(x_values, jitter_scale), y_values, 50, colors(animal, :), ...
  218. 'filled', 'DisplayName', sprintf('Animal %d', animal));
  219. end
  220. mean_values_fMRI = mean(state_neural_similarity_fMRI, 2, 'omitnan');
  221. for state = 1:length(mean_values_fMRI)
  222. plot([state - 0.4, state + 0.4], [mean_values_fMRI(state), mean_values_fMRI(state)], 'k-', 'LineWidth', 2);
  223. end
  224. title('fMRI state similarity to neural dFC windows');
  225. xlabel('fMRI state'); ylabel('Similarity'); legend('show', 'Location', 'eastoutside'); grid on; hold off;
  227. %% Similarity between FL and fMRI state centroids
  228. similarity_fMRI_combined = corr(centroids_fMRI', centroids_combined', 'Rows', 'pairwise');
  229. similarity_fMRI_states = corr(centroids_fMRI', 'Rows', 'pairwise');
  230. similarity_FL_states = corr(centroids_combined', 'Rows', 'pairwise');
  232. figure;
  233. imagesc(similarity_fMRI_combined);
  234. colorbar;
  235. colormap(cm_turbo_custom);
  236. title('Similarity between fMRI states and FL states');
  237. xlabel('FL state'); ylabel('fMRI state');
  239. figure;
  240. imagesc(similarity_fMRI_states);
  241. colorbar;
  242. colormap(cm_turbo_custom_endWhite);
  243. caxis([0.65 0.9]);
  244. title('Similarity between fMRI states');
  245. xlabel('fMRI state'); ylabel('fMRI state');
  247. figure;
  248. imagesc(similarity_FL_states);
  249. colorbar;
  250. colormap(cm_turbo_custom_endWhite);
  251. caxis([0.6 0.96]);
  252. title('Similarity between FL states');
  253. xlabel('FL state'); ylabel('FL state');
  255. %% State distributions and transition matrices during CO2 on/off
  256. state_dist_10Hz_on = zeros(optimal_num_clusters_combined, num_animals);
  257. state_dist_10Hz_off = zeros(optimal_num_clusters_combined, num_animals);
  258. state_dist_1Hz_on = zeros(optimal_num_clusters_combined, num_animals);
  259. state_dist_1Hz_off = zeros(optimal_num_clusters_combined, num_animals);
  260. state_dist_fMRI_on = zeros(optimal_num_clusters_fMRI, num_animals);
  261. state_dist_fMRI_off = zeros(optimal_num_clusters_fMRI, num_animals);
  263. trans_mat_fMRI_on = zeros(optimal_num_clusters_fMRI, optimal_num_clusters_fMRI, num_animals);
  264. trans_mat_fMRI_off = zeros(optimal_num_clusters_fMRI, optimal_num_clusters_fMRI, num_animals);
  265. trans_mat_10Hz_on = zeros(optimal_num_clusters_combined, optimal_num_clusters_combined, num_animals);
  266. trans_mat_10Hz_off = zeros(optimal_num_clusters_combined, optimal_num_clusters_combined, num_animals);
  267. trans_mat_1Hz_on = zeros(optimal_num_clusters_combined, optimal_num_clusters_combined, num_animals);
  268. trans_mat_1Hz_off = zeros(optimal_num_clusters_combined, optimal_num_clusters_combined, num_animals);
  270. for animal = 1:num_animals
  271. [starts_10Hz, starts_1Hz, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  272. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  273. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  274. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  276. ends_10Hz = starts_10Hz + co2_duration;
  277. ends_1Hz = starts_1Hz + co2_duration;
  278. ends_fMRI = starts_fMRI + co2_duration;
  280. off_starts_fMRI = [1, ends_fMRI + window_size_fMRI / 2];
  281. off_ends_fMRI = [starts_fMRI - window_size_fMRI / 2, num_windows_fMRI];
  283. off_starts_10Hz = [1, ends_10Hz + window_size_10Hz / sampling_rate_10Hz / 2];
  284. off_ends_10Hz = [starts_10Hz - window_size_10Hz / sampling_rate_10Hz / 2, num_windows_10Hz];
  286. off_starts_1Hz = [1, ends_1Hz + window_size_1Hz / 2];
  287. off_ends_1Hz = [starts_1Hz - window_size_1Hz / 2, num_windows_1Hz];
  289. % CO2-on state counts and transitions
  290. for cycle = 1:length(starts_10Hz)
  291. state_dist_10Hz_on(:, animal) = state_dist_10Hz_on(:, animal) + count_states_in_range(idx_10Hz(:, animal), starts_10Hz(cycle), min(ends_10Hz(cycle), num_windows_10Hz), optimal_num_clusters_combined);
  292. state_dist_1Hz_on(:, animal) = state_dist_1Hz_on(:, animal) + count_states_in_range(idx_1Hz(:, animal), starts_1Hz(cycle), min(ends_1Hz(cycle), num_windows_1Hz), optimal_num_clusters_combined);
  293. state_dist_fMRI_on(:, animal) = state_dist_fMRI_on(:, animal) + count_states_in_range(idx_fMRI(:, animal), starts_fMRI(cycle), min(ends_fMRI(cycle), num_windows_fMRI), optimal_num_clusters_fMRI);
  295. trans_mat_10Hz_on(:, :, animal) = trans_mat_10Hz_on(:, :, animal) + count_transitions_in_range(idx_10Hz(:, animal), starts_10Hz(cycle), min(ends_10Hz(cycle), num_windows_10Hz), optimal_num_clusters_combined);
  296. trans_mat_1Hz_on(:, :, animal) = trans_mat_1Hz_on(:, :, animal) + count_transitions_in_range(idx_1Hz(:, animal), starts_1Hz(cycle), min(ends_1Hz(cycle), num_windows_1Hz), optimal_num_clusters_combined);
  297. trans_mat_fMRI_on(:, :, animal) = trans_mat_fMRI_on(:, :, animal) + count_transitions_in_range(idx_fMRI(:, animal), starts_fMRI(cycle), min(ends_fMRI(cycle), num_windows_fMRI), optimal_num_clusters_fMRI);
  298. end
  300. % CO2-off state counts and transitions
  301. for cycle = 1:length(off_starts_10Hz)
  302. state_dist_10Hz_off(:, animal) = state_dist_10Hz_off(:, animal) + count_states_in_range(idx_10Hz(:, animal), max(1, round(off_starts_10Hz(cycle))), min(round(off_ends_10Hz(cycle)), num_windows_10Hz), optimal_num_clusters_combined);
  303. state_dist_1Hz_off(:, animal) = state_dist_1Hz_off(:, animal) + count_states_in_range(idx_1Hz(:, animal), max(1, round(off_starts_1Hz(cycle))), min(round(off_ends_1Hz(cycle)), num_windows_1Hz), optimal_num_clusters_combined);
  304. state_dist_fMRI_off(:, animal) = state_dist_fMRI_off(:, animal) + count_states_in_range(idx_fMRI(:, animal), max(1, round(off_starts_fMRI(cycle))), min(round(off_ends_fMRI(cycle)), num_windows_fMRI), optimal_num_clusters_fMRI);
  306. trans_mat_10Hz_off(:, :, animal) = trans_mat_10Hz_off(:, :, animal) + count_transitions_in_range(idx_10Hz(:, animal), max(1, round(off_starts_10Hz(cycle))), min(round(off_ends_10Hz(cycle)), num_windows_10Hz), optimal_num_clusters_combined);
  307. trans_mat_1Hz_off(:, :, animal) = trans_mat_1Hz_off(:, :, animal) + count_transitions_in_range(idx_1Hz(:, animal), max(1, round(off_starts_1Hz(cycle))), min(round(off_ends_1Hz(cycle)), num_windows_1Hz), optimal_num_clusters_combined);
  308. trans_mat_fMRI_off(:, :, animal) = trans_mat_fMRI_off(:, :, animal) + count_transitions_in_range(idx_fMRI(:, animal), max(1, round(off_starts_fMRI(cycle))), min(round(off_ends_fMRI(cycle)), num_windows_fMRI), optimal_num_clusters_fMRI);
  309. end
  310. end
  312. % Normalize per animal
  313. state_dist_10Hz_on = normalize_columns(state_dist_10Hz_on);
  314. state_dist_10Hz_off = normalize_columns(state_dist_10Hz_off);
  315. state_dist_1Hz_on = normalize_columns(state_dist_1Hz_on);
  316. state_dist_1Hz_off = normalize_columns(state_dist_1Hz_off);
  317. state_dist_fMRI_on = normalize_columns(state_dist_fMRI_on);
  318. state_dist_fMRI_off = normalize_columns(state_dist_fMRI_off);
  320. mean_state_dist_10Hz_on = mean(state_dist_10Hz_on, 2, 'omitnan');
  321. mean_state_dist_10Hz_off = mean(state_dist_10Hz_off, 2, 'omitnan');
  322. mean_state_dist_1Hz_on = mean(state_dist_1Hz_on, 2, 'omitnan');
  323. mean_state_dist_1Hz_off = mean(state_dist_1Hz_off, 2, 'omitnan');
  324. mean_state_dist_fMRI_on = mean(state_dist_fMRI_on, 2, 'omitnan');
  325. mean_state_dist_fMRI_off = mean(state_dist_fMRI_off, 2, 'omitnan');
  327. sem_state_dist_10Hz_on = std(state_dist_10Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  328. sem_state_dist_10Hz_off = std(state_dist_10Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  329. sem_state_dist_1Hz_on = std(state_dist_1Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  330. sem_state_dist_1Hz_off = std(state_dist_1Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  331. sem_state_dist_fMRI_on = std(state_dist_fMRI_on, 0, 2, 'omitnan') / sqrt(num_animals);
  332. sem_state_dist_fMRI_off = std(state_dist_fMRI_off, 0, 2, 'omitnan') / sqrt(num_animals);
  334. p_values_state_10Hz = nan(optimal_num_clusters_combined, 1);
  335. p_values_state_1Hz = nan(optimal_num_clusters_combined, 1);
  336. p_values_state_fMRI = nan(optimal_num_clusters_fMRI, 1);
  337. for i = 1:optimal_num_clusters_combined
  338. p_values_state_10Hz(i) = safe_paired_ttest(state_dist_10Hz_on(i, :), state_dist_10Hz_off(i, :));
  339. p_values_state_1Hz(i) = safe_paired_ttest(state_dist_1Hz_on(i, :), state_dist_1Hz_off(i, :));
  340. end
  341. for i = 1:optimal_num_clusters_fMRI
  342. p_values_state_fMRI(i) = safe_paired_ttest(state_dist_fMRI_on(i, :), state_dist_fMRI_off(i, :));
  343. end
  345. fprintf('P-values for state distributions - FL 10 Hz:\n'); disp(p_values_state_10Hz);
  346. fprintf('P-values for state distributions - FL 1 Hz:\n'); disp(p_values_state_1Hz);
  347. fprintf('P-values for state distributions - fMRI:\n'); disp(p_values_state_fMRI);
  349. figure('Position', [0 0 400 900]);
  350. plot_grouped_bars_with_paired_points(3, 1, 1, mean_state_dist_10Hz_on, mean_state_dist_10Hz_off, sem_state_dist_10Hz_on, sem_state_dist_10Hz_off, state_dist_10Hz_on, state_dist_10Hz_off, 'State Distribution - FL 10 Hz', 'Proportion', 1:optimal_num_clusters_combined, [0 0.68]);
  351. plot_grouped_bars_with_paired_points(3, 1, 2, mean_state_dist_1Hz_on, mean_state_dist_1Hz_off, sem_state_dist_1Hz_on, sem_state_dist_1Hz_off, state_dist_1Hz_on, state_dist_1Hz_off, 'State Distribution - FL 1 Hz', 'Proportion', 1:optimal_num_clusters_combined, [0 0.38]);
  352. plot_grouped_bars_with_paired_points(3, 1, 3, mean_state_dist_fMRI_on, mean_state_dist_fMRI_off, sem_state_dist_fMRI_on, sem_state_dist_fMRI_off, state_dist_fMRI_on, state_dist_fMRI_off, 'State Distribution - fMRI', 'Proportion', 1:optimal_num_clusters_fMRI, [0 0.83]);
  353. sgtitle('State distributions for FL 10 Hz, FL 1 Hz, and fMRI (CO2 on/off)');
  355. %% Transition probability matrices
  356. trans_matrix_10Hz = trans_mat_10Hz_on + trans_mat_10Hz_off;
  357. trans_matrix_1Hz = trans_mat_1Hz_on + trans_mat_1Hz_off;
  358. trans_matrix_fMRI = trans_mat_fMRI_on + trans_mat_fMRI_off;
  360. trans_prob_10Hz = normalize_transition_tensor(trans_matrix_10Hz);
  361. trans_prob_10Hz_on = normalize_transition_tensor(trans_mat_10Hz_on);
  362. trans_prob_10Hz_off = normalize_transition_tensor(trans_mat_10Hz_off);
  363. trans_prob_1Hz = normalize_transition_tensor(trans_matrix_1Hz);
  364. trans_prob_1Hz_on = normalize_transition_tensor(trans_mat_1Hz_on);
  365. trans_prob_1Hz_off = normalize_transition_tensor(trans_mat_1Hz_off);
  366. trans_prob_fMRI = normalize_transition_tensor(trans_matrix_fMRI);
  367. trans_prob_fMRI_on = normalize_transition_tensor(trans_mat_fMRI_on);
  368. trans_prob_fMRI_off = normalize_transition_tensor(trans_mat_fMRI_off);
  370. avg_trans_mat_10Hz_on = mean(trans_prob_10Hz_on, 3, 'omitnan');
  371. avg_trans_mat_10Hz_off = mean(trans_prob_10Hz_off, 3, 'omitnan');
  372. avg_trans_mat_1Hz_on = mean(trans_prob_1Hz_on, 3, 'omitnan');
  373. avg_trans_mat_1Hz_off = mean(trans_prob_1Hz_off, 3, 'omitnan');
  374. avg_trans_mat_fMRI_on = mean(trans_prob_fMRI_on, 3, 'omitnan');
  375. avg_trans_mat_fMRI_off = mean(trans_prob_fMRI_off, 3, 'omitnan');
  377. p_values_trans_10Hz = nan(optimal_num_clusters_combined, optimal_num_clusters_combined);
  378. p_values_trans_1Hz = nan(optimal_num_clusters_combined, optimal_num_clusters_combined);
  379. p_values_trans_fMRI = nan(optimal_num_clusters_fMRI, optimal_num_clusters_fMRI);
  380. for i = 1:optimal_num_clusters_combined
  381. for j = 1:optimal_num_clusters_combined
  382. p_values_trans_10Hz(i, j) = safe_paired_ttest(squeeze(trans_prob_10Hz_on(i, j, :)), squeeze(trans_prob_10Hz_off(i, j, :)));
  383. p_values_trans_1Hz(i, j) = safe_paired_ttest(squeeze(trans_prob_1Hz_on(i, j, :)), squeeze(trans_prob_1Hz_off(i, j, :)));
  384. end
  385. end
  386. for i = 1:optimal_num_clusters_fMRI
  387. for j = 1:optimal_num_clusters_fMRI
  388. p_values_trans_fMRI(i, j) = safe_paired_ttest(squeeze(trans_prob_fMRI_on(i, j, :)), squeeze(trans_prob_fMRI_off(i, j, :)));
  389. end
  390. end
  392. fprintf('P-values for transition probabilities - FL 10 Hz:\n'); disp(p_values_trans_10Hz);
  393. fprintf('P-values for transition probabilities - FL 1 Hz:\n'); disp(p_values_trans_1Hz);
  394. fprintf('P-values for transition probabilities - fMRI:\n'); disp(p_values_trans_fMRI);
  396. figure;
  397. threshold_transition = 0.05;
  398. plot_transition_matrix_with_sig(3, 2, 1, avg_trans_mat_10Hz_on, p_values_trans_10Hz, threshold_transition, 'Average Transition Matrix - FL 10 Hz (CO2 On)', 'Next State', 'Current State', [-0.01 0.9], cm_custom_blueToRed);
  399. plot_transition_matrix_with_sig(3, 2, 2, avg_trans_mat_10Hz_off, p_values_trans_10Hz, threshold_transition, 'Average Transition Matrix - FL 10 Hz (CO2 Off)', 'Next State', 'Current State', [-0.01 0.9], cm_custom_blueToRed);
  400. plot_transition_matrix_with_sig(3, 2, 3, avg_trans_mat_1Hz_on, p_values_trans_1Hz, threshold_transition, 'Average Transition Matrix - FL 1 Hz (CO2 On)', 'Next State', 'Current State', [-0.01 0.9], cm_custom_blueToRed);
  401. plot_transition_matrix_with_sig(3, 2, 4, avg_trans_mat_1Hz_off, p_values_trans_1Hz, threshold_transition, 'Average Transition Matrix - FL 1 Hz (CO2 Off)', 'Next State', 'Current State', [-0.01 0.9], cm_custom_blueToRed);
  402. plot_transition_matrix_with_sig(3, 2, 5, avg_trans_mat_fMRI_on, p_values_trans_fMRI, threshold_transition, 'Average Transition Matrix - fMRI (CO2 On)', 'Next State', 'Current State', [-0.02 0.9], cm_custom_blueToRed);
  403. plot_transition_matrix_with_sig(3, 2, 6, avg_trans_mat_fMRI_off, p_values_trans_fMRI, threshold_transition, 'Average Transition Matrix - fMRI (CO2 Off)', 'Next State', 'Current State', [-0.02 0.9], cm_custom_blueToRed);
  404. sgtitle('Average transition matrices for FL 10 Hz, FL 1 Hz, and fMRI (CO2 on/off)');
  406. %% Joint FL-fMRI state distributions (overall)
  407. count_10Hz_fMRI = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  408. count_1Hz_fMRI = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  409. count_fMRI_10Hz = zeros(optimal_num_clusters_fMRI, optimal_num_clusters_combined);
  410. count_fMRI_1Hz = zeros(optimal_num_clusters_fMRI, optimal_num_clusters_combined);
  412. for animal = 1:num_animals
  413. for t = 1:num_windows_fMRI
  414. f_state = idx_fMRI(t, animal);
  415. if t <= num_windows_10Hz
  416. fl10_state = idx_10Hz(t, animal);
  417. count_10Hz_fMRI(fl10_state, f_state) = count_10Hz_fMRI(fl10_state, f_state) + 1;
  418. end
  419. if t <= num_windows_1Hz
  420. fl1_state = idx_1Hz(t, animal);
  421. count_1Hz_fMRI(fl1_state, f_state) = count_1Hz_fMRI(fl1_state, f_state) + 1;
  422. end
  423. end
  425. for t = 1:num_windows_10Hz
  426. fl10_state = idx_10Hz(t, animal);
  427. if t <= num_windows_fMRI
  428. f_state = idx_fMRI(t, animal);
  429. count_fMRI_10Hz(f_state, fl10_state) = count_fMRI_10Hz(f_state, fl10_state) + 1;
  430. end
  431. end
  433. for t = 1:num_windows_1Hz
  434. fl1_state = idx_1Hz(t, animal);
  435. if t <= num_windows_fMRI
  436. f_state = idx_fMRI(t, animal);
  437. count_fMRI_1Hz(f_state, fl1_state) = count_fMRI_1Hz(f_state, fl1_state) + 1;
  438. end
  439. end
  440. end
  442. percent_10Hz_fMRI = safe_percent_by_column(count_10Hz_fMRI);
  443. percent_1Hz_fMRI = safe_percent_by_column(count_1Hz_fMRI);
  444. percent_fMRI_10Hz = safe_percent_by_column(count_fMRI_10Hz);
  445. percent_fMRI_1Hz = safe_percent_by_column(count_fMRI_1Hz);
  447. ulim = 40;
  448. figure;
  449. subplot(1, 2, 1); imagesc(percent_10Hz_fMRI); colorbar; axis square; colormap(cm_turbo_custom); caxis([0 ulim]); title('Percentage distribution of 10 Hz FL states by fMRI state'); xlabel('fMRI State'); ylabel('10 Hz FL State');
  450. subplot(1, 2, 2); imagesc(percent_1Hz_fMRI); colorbar; axis square; colormap(cm_turbo_custom); caxis([0 ulim]); title('Percentage distribution of 1 Hz FL states by fMRI state'); xlabel('fMRI State'); ylabel('1 Hz FL State');
  452. figure;
  453. subplot(1, 2, 1); imagesc(percent_fMRI_10Hz); colorbar; axis square; colormap(cm_turbo_custom); caxis([0 ulim]); title('Percentage distribution of fMRI states by 10 Hz FL state'); xlabel('10 Hz FL State'); ylabel('fMRI State');
  454. subplot(1, 2, 2); imagesc(percent_fMRI_1Hz); colorbar; axis square; colormap(cm_turbo_custom); caxis([0 ulim]); title('Percentage distribution of fMRI states by 1 Hz FL state'); xlabel('1 Hz FL State'); ylabel('fMRI State');
  456. %% Joint FL-fMRI state distributions by CO2 condition (10 Hz FL, fMRI)
  457. count_10Hz_fMRI_on = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI, num_animals);
  458. count_10Hz_fMRI_off = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI, num_animals);
  459. count_fMRI_10Hz_on = zeros(optimal_num_clusters_fMRI, optimal_num_clusters_combined, num_animals);
  460. count_fMRI_10Hz_off = zeros(optimal_num_clusters_fMRI, optimal_num_clusters_combined, num_animals);
  462. for animal = 1:num_animals
  463. [starts_10Hz, ~, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  464. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  465. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  466. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  468. timing_info_fMRI = build_timing_mask(num_windows_fMRI, starts_fMRI, starts_fMRI + co2_duration);
  469. timing_info_10Hz = build_timing_mask(num_windows_10Hz, starts_10Hz, starts_10Hz + co2_duration);
  471. for t = 1:num_windows_fMRI
  472. f_state = idx_fMRI(t, animal);
  473. if t <= num_windows_10Hz
  474. fl_state = idx_10Hz(t, animal);
  475. if timing_info_10Hz(t)
  476. count_10Hz_fMRI_on(fl_state, f_state, animal) = count_10Hz_fMRI_on(fl_state, f_state, animal) + 1;
  477. else
  478. count_10Hz_fMRI_off(fl_state, f_state, animal) = count_10Hz_fMRI_off(fl_state, f_state, animal) + 1;
  479. end
  480. end
  481. end
  483. for t = 1:num_windows_10Hz
  484. fl_state = idx_10Hz(t, animal);
  485. if t <= num_windows_fMRI
  486. f_state = idx_fMRI(t, animal);
  487. if timing_info_fMRI(t)
  488. count_fMRI_10Hz_on(f_state, fl_state, animal) = count_fMRI_10Hz_on(f_state, fl_state, animal) + 1;
  489. else
  490. count_fMRI_10Hz_off(f_state, fl_state, animal) = count_fMRI_10Hz_off(f_state, fl_state, animal) + 1;
  491. end
  492. end
  493. end
  494. end
  496. total_count_10Hz_fMRI_on = sum(count_10Hz_fMRI_on, 3);
  497. total_count_10Hz_fMRI_off = sum(count_10Hz_fMRI_off, 3);
  498. total_count_fMRI_10Hz_on = sum(count_fMRI_10Hz_on, 3);
  499. total_count_fMRI_10Hz_off = sum(count_fMRI_10Hz_off, 3);
  501. percent_10Hz_fMRI_on = safe_percent_by_column(total_count_10Hz_fMRI_on);
  502. percent_10Hz_fMRI_off = safe_percent_by_column(total_count_10Hz_fMRI_off);
  503. percent_fMRI_10Hz_on = safe_percent_by_column(total_count_fMRI_10Hz_on);
  504. percent_fMRI_10Hz_off = safe_percent_by_column(total_count_fMRI_10Hz_off);
  506. p_values_10Hz_fMRI = nan(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  507. p_values_fMRI_10Hz = nan(optimal_num_clusters_fMRI, optimal_num_clusters_combined);
  508. for i = 1:optimal_num_clusters_combined
  509. for j = 1:optimal_num_clusters_fMRI
  510. p_values_10Hz_fMRI(i, j) = safe_paired_ttest(squeeze(count_10Hz_fMRI_on(i, j, :)), squeeze(count_10Hz_fMRI_off(i, j, :)));
  511. end
  512. end
  513. for i = 1:optimal_num_clusters_fMRI
  514. for j = 1:optimal_num_clusters_combined
  515. p_values_fMRI_10Hz(i, j) = safe_paired_ttest(squeeze(count_fMRI_10Hz_on(i, j, :)), squeeze(count_fMRI_10Hz_off(i, j, :)));
  516. end
  517. end
  519. fprintf('P-values for joint state distributions - 10 Hz FL to fMRI:\n'); disp(p_values_10Hz_fMRI);
  520. fprintf('P-values for joint state distributions - fMRI to 10 Hz FL:\n'); disp(p_values_fMRI_10Hz);
  522. figure;
  523. threshold_joint = 0.05;
  524. ulim = 50;
  525. plot_matrix_with_sig(2, 2, 1, percent_10Hz_fMRI_on, p_values_10Hz_fMRI, threshold_joint, 'Percentage Distribution - 10 Hz FL to fMRI (CO2 On)', 'fMRI State', '10 Hz FL State', [0 ulim], cm_custom_blue2red_2);
  526. plot_matrix_with_sig(2, 2, 2, percent_10Hz_fMRI_off, p_values_10Hz_fMRI, threshold_joint, 'Percentage Distribution - 10 Hz FL to fMRI (CO2 Off)', 'fMRI State', '10 Hz FL State', [0 ulim], cm_custom_blue2red_2);
  527. plot_matrix_with_sig(2, 2, 3, percent_fMRI_10Hz_on, p_values_fMRI_10Hz, threshold_joint, 'Percentage Distribution - fMRI to 10 Hz FL (CO2 On)', '10 Hz FL State', 'fMRI State', [0 ulim], cm_custom_blue2red_2);
  528. plot_matrix_with_sig(2, 2, 4, percent_fMRI_10Hz_off, p_values_fMRI_10Hz, threshold_joint, 'Percentage Distribution - fMRI to 10 Hz FL (CO2 Off)', '10 Hz FL State', 'fMRI State', [0 ulim], cm_custom_blue2red_2);
  529. sgtitle('Percentage distributions for FL and fMRI states during CO2 on/off');
  531. %% Joint transition probabilities: FL to fMRI (1 Hz and 10 Hz), overall and by CO2
  532. joint_transitions_10Hz_fMRI = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  533. joint_transitions_1Hz_fMRI = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  535. for animal = 1:num_animals
  536. for t = 1:(min(num_windows_10Hz, num_windows_fMRI) - 1)
  537. joint_transitions_10Hz_fMRI(idx_10Hz(t, animal), idx_fMRI(t, animal)) = joint_transitions_10Hz_fMRI(idx_10Hz(t, animal), idx_fMRI(t, animal)) + 1;
  538. end
  539. for t = 1:(min(num_windows_1Hz, num_windows_fMRI) - 1)
  540. joint_transitions_1Hz_fMRI(idx_1Hz(t, animal), idx_fMRI(t, animal)) = joint_transitions_1Hz_fMRI(idx_1Hz(t, animal), idx_fMRI(t, animal)) + 1;
  541. end
  542. end
  544. joint_trans_prob_10Hz_fMRI = normalize_rows_2d(joint_transitions_10Hz_fMRI);
  545. joint_trans_prob_1Hz_fMRI = normalize_rows_2d(joint_transitions_1Hz_fMRI);
  547. figure;
  548. subplot(1, 2, 1); imagesc(joint_trans_prob_10Hz_fMRI); colorbar; colormap(cm_turbo_custom); caxis([0 0.5]); title('Joint transition probabilities - 10 Hz FL to fMRI'); xlabel('fMRI State'); ylabel('10 Hz FL State');
  549. subplot(1, 2, 2); imagesc(joint_trans_prob_1Hz_fMRI); colorbar; colormap(cm_turbo_custom); caxis([0 0.5]); title('Joint transition probabilities - 1 Hz FL to fMRI'); xlabel('fMRI State'); ylabel('1 Hz FL State');
  551. joint_trans_prob_10Hz_fMRI_on = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI, num_animals);
  552. joint_trans_prob_10Hz_fMRI_off = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI, num_animals);
  553. joint_trans_prob_1Hz_fMRI_on = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI, num_animals);
  554. joint_trans_prob_1Hz_fMRI_off = zeros(optimal_num_clusters_combined, optimal_num_clusters_fMRI, num_animals);
  556. for animal = 1:num_animals
  557. [~, ~, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  558. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  559. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  560. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  561. ends_fMRI = starts_fMRI + co2_on_duration_fMRI;
  563. for t = 1:(min(num_windows_10Hz, num_windows_fMRI) - 1)
  564. current_state_10Hz = idx_10Hz(t, animal);
  565. next_state_fMRI = idx_fMRI(t + 1, animal);
  566. if current_state_10Hz > 0 && next_state_fMRI > 0
  567. if any(t >= starts_fMRI & t <= ends_fMRI)
  568. joint_trans_prob_10Hz_fMRI_on(current_state_10Hz, next_state_fMRI, animal) = joint_trans_prob_10Hz_fMRI_on(current_state_10Hz, next_state_fMRI, animal) + 1;
  569. else
  570. joint_trans_prob_10Hz_fMRI_off(current_state_10Hz, next_state_fMRI, animal) = joint_trans_prob_10Hz_fMRI_off(current_state_10Hz, next_state_fMRI, animal) + 1;
  571. end
  572. end
  573. end
  575. for t = 1:(min(num_windows_1Hz, num_windows_fMRI) - 1)
  576. current_state_1Hz = idx_1Hz(t, animal);
  577. next_state_fMRI = idx_fMRI(t + 1, animal);
  578. if current_state_1Hz > 0 && next_state_fMRI > 0
  579. if any(t >= starts_fMRI & t <= ends_fMRI)
  580. joint_trans_prob_1Hz_fMRI_on(current_state_1Hz, next_state_fMRI, animal) = joint_trans_prob_1Hz_fMRI_on(current_state_1Hz, next_state_fMRI, animal) + 1;
  581. else
  582. joint_trans_prob_1Hz_fMRI_off(current_state_1Hz, next_state_fMRI, animal) = joint_trans_prob_1Hz_fMRI_off(current_state_1Hz, next_state_fMRI, animal) + 1;
  583. end
  584. end
  585. end
  586. end
  588. joint_trans_prob_10Hz_fMRI_on = normalize_transition_tensor(joint_trans_prob_10Hz_fMRI_on);
  589. joint_trans_prob_10Hz_fMRI_off = normalize_transition_tensor(joint_trans_prob_10Hz_fMRI_off);
  590. joint_trans_prob_1Hz_fMRI_on = normalize_transition_tensor(joint_trans_prob_1Hz_fMRI_on);
  591. joint_trans_prob_1Hz_fMRI_off = normalize_transition_tensor(joint_trans_prob_1Hz_fMRI_off);
  593. p_values_joint_10Hz_fMRI = nan(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  594. p_values_joint_1Hz_fMRI = nan(optimal_num_clusters_combined, optimal_num_clusters_fMRI);
  595. for i = 1:optimal_num_clusters_combined
  596. for j = 1:optimal_num_clusters_fMRI
  597. p_values_joint_10Hz_fMRI(i, j) = safe_paired_ttest(squeeze(joint_trans_prob_10Hz_fMRI_on(i, j, :)), squeeze(joint_trans_prob_10Hz_fMRI_off(i, j, :)));
  598. p_values_joint_1Hz_fMRI(i, j) = safe_paired_ttest(squeeze(joint_trans_prob_1Hz_fMRI_on(i, j, :)), squeeze(joint_trans_prob_1Hz_fMRI_off(i, j, :)));
  599. end
  600. end
  602. fprintf('P-values for joint transition probabilities - 10 Hz FL to fMRI:\n'); disp(p_values_joint_10Hz_fMRI);
  603. fprintf('P-values for joint transition probabilities - 1 Hz FL to fMRI:\n'); disp(p_values_joint_1Hz_fMRI);
  605. figure;
  606. threshold_joint_trans = 0.1;
  607. plot_matrix_with_sig(2, 2, 1, mean(joint_trans_prob_10Hz_fMRI_on, 3, 'omitnan'), p_values_joint_10Hz_fMRI, threshold_joint_trans, 'Joint Transition Probabilities - 10 Hz FL to fMRI (CO2 On)', 'fMRI State', '10 Hz FL State', [0 0.5], cm_turbo_custom);
  608. plot_matrix_with_sig(2, 2, 2, mean(joint_trans_prob_10Hz_fMRI_off, 3, 'omitnan'), p_values_joint_10Hz_fMRI, threshold_joint_trans, 'Joint Transition Probabilities - 10 Hz FL to fMRI (CO2 Off)', 'fMRI State', '10 Hz FL State', [0 0.5], cm_turbo_custom);
  609. plot_matrix_with_sig(2, 2, 3, mean(joint_trans_prob_1Hz_fMRI_on, 3, 'omitnan'), p_values_joint_1Hz_fMRI, threshold_joint_trans, 'Joint Transition Probabilities - 1 Hz FL to fMRI (CO2 On)', 'fMRI State', '1 Hz FL State', [0 0.5], cm_turbo_custom);
  610. plot_matrix_with_sig(2, 2, 4, mean(joint_trans_prob_1Hz_fMRI_off, 3, 'omitnan'), p_values_joint_1Hz_fMRI, threshold_joint_trans, 'Joint Transition Probabilities - 1 Hz FL to fMRI (CO2 Off)', 'fMRI State', '1 Hz FL State', [0 0.5], cm_turbo_custom);
  611. sgtitle('Joint transition probabilities for FL to fMRI (CO2 on/off)');
  613. %% Similarity between FL and fMRI dFC matrices over time
  614. similarity_10Hz_fMRI = nan(num_windows_fMRI, num_animals);
  615. similarity_1Hz_fMRI = nan(num_windows_fMRI, num_animals);
  617. dFC_10Hz_reshaped = reshape(dFC_10Hz, num_regions * num_regions, num_windows_10Hz, num_animals);
  618. dFC_1Hz_reshaped = reshape(dFC_1Hz, num_regions * num_regions, num_windows_1Hz, num_animals);
  619. dFC_fMRI_reshaped = reshape(dFC_fMRI, num_regions * num_regions, num_windows_fMRI, num_animals);
  621. for animal = 1:num_animals
  622. for t = 1:num_windows_fMRI
  623. start_time_fMRI = (t - 1) * step_size_fMRI + 1;
  625. start_time_10Hz = round(start_time_fMRI / (step_size_10Hz / 10)) - 4;
  626. end_time_10Hz = start_time_10Hz + 5;
  627. if start_time_10Hz > 0 && end_time_10Hz <= num_windows_10Hz
  628. avg_dFC_10Hz = mean(dFC_10Hz_reshaped(:, start_time_10Hz:end_time_10Hz, animal), 2);
  629. similarity_10Hz_fMRI(t, animal) = corr(avg_dFC_10Hz, dFC_fMRI_reshaped(:, t, animal), 'Rows', 'pairwise');
  630. end
  632. t_1Hz = round(((t - 1) * step_size_fMRI) / step_size_1Hz) + 1;
  633. if t_1Hz > 0 && t_1Hz <= num_windows_1Hz
  634. similarity_1Hz_fMRI(t, animal) = corr(dFC_1Hz_reshaped(:, t_1Hz, animal), dFC_fMRI_reshaped(:, t, animal), 'Rows', 'pairwise');
  635. end
  636. end
  637. end
  639. similarity_10Hz_fMRI_on = [];
  640. similarity_10Hz_fMRI_off = [];
  641. similarity_1Hz_fMRI_on = [];
  642. similarity_1Hz_fMRI_off = [];
  643. cc = 30;
  645. for animal = 1:num_animals
  646. [starts_10Hz, starts_1Hz, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  647. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  648. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  649. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  651. ends_10Hz = starts_10Hz + co2_duration;
  652. ends_1Hz = starts_1Hz + co2_duration;
  653. ends_fMRI = starts_fMRI + co2_duration;
  655. off_starts_fMRI = [1, ends_fMRI + window_size_fMRI / 2];
  656. off_ends_fMRI = [starts_fMRI - window_size_fMRI / 2, num_windows_fMRI];
  657. off_starts_10Hz = [1, ends_10Hz + window_size_10Hz / sampling_rate_10Hz / 2];
  658. off_ends_10Hz = [starts_10Hz - window_size_10Hz / sampling_rate_10Hz / 2, num_windows_10Hz];
  660. for cycle = 1:length(starts_10Hz)
  661. for t = (starts_10Hz(cycle) + cc):min(ends_10Hz(cycle), num_windows_fMRI) - cc
  662. if t > 0 && t <= num_windows_10Hz
  663. similarity_10Hz_fMRI_on = [similarity_10Hz_fMRI_on; similarity_10Hz_fMRI(t, animal)]; %#ok<AGROW>
  664. end
  665. if t > 0 && t <= num_windows_1Hz
  666. similarity_1Hz_fMRI_on = [similarity_1Hz_fMRI_on; similarity_1Hz_fMRI(t, animal)]; %#ok<AGROW>
  667. end
  668. end
  669. end
  671. for cycle = 1:length(off_starts_10Hz)
  672. for t = max(1, round(off_starts_10Hz(cycle))) + cc : round(off_ends_10Hz(cycle)) - cc
  673. if t > 0 && t <= num_windows_10Hz
  674. similarity_10Hz_fMRI_off = [similarity_10Hz_fMRI_off; similarity_10Hz_fMRI(t, animal)]; %#ok<AGROW>
  675. end
  676. if t > 0 && t <= num_windows_1Hz
  677. similarity_1Hz_fMRI_off = [similarity_1Hz_fMRI_off; similarity_1Hz_fMRI(t, animal)]; %#ok<AGROW>
  678. end
  679. end
  680. end
  681. end
  683. llim = 10;
  684. ulim = 90;
  685. similarity_1Hz_fMRI_on_filtered = percentile_filter(similarity_1Hz_fMRI_on, llim, ulim);
  686. similarity_1Hz_fMRI_off_filtered = percentile_filter(similarity_1Hz_fMRI_off, llim, ulim);
  687. similarity_10Hz_fMRI_on_filtered = percentile_filter(similarity_10Hz_fMRI_on, llim, ulim);
  688. similarity_10Hz_fMRI_off_filtered = percentile_filter(similarity_10Hz_fMRI_off, llim, ulim);
  690. figure;
  691. subplot(1, 2, 1);
  692. plot_distribution_comparison(similarity_1Hz_fMRI_on_filtered, similarity_1Hz_fMRI_off_filtered, '1 Hz similarity');
  693. subplot(1, 2, 2);
  694. plot_distribution_comparison(similarity_10Hz_fMRI_on_filtered, similarity_10Hz_fMRI_off_filtered, '10 Hz similarity');
  696. [~, p_10Hz] = ttest2(similarity_10Hz_fMRI_on_filtered, similarity_10Hz_fMRI_off_filtered);
  697. [~, p_1Hz] = ttest2(similarity_1Hz_fMRI_on_filtered, similarity_1Hz_fMRI_off_filtered);
  699. effect_1Hz = cohens_d_two_sample(similarity_1Hz_fMRI_on_filtered, similarity_1Hz_fMRI_off_filtered);
  700. effect_10Hz = cohens_d_two_sample(similarity_10Hz_fMRI_on_filtered, similarity_10Hz_fMRI_off_filtered);
  702. fprintf('P-value for similarity comparison (10 Hz FL): %.6f\n', p_10Hz);
  703. fprintf('P-value for similarity comparison (1 Hz FL): %.6f\n', p_1Hz);
  704. fprintf('Cohen''s d for similarity comparison (10 Hz FL): %.6f\n', effect_10Hz);
  705. fprintf('Cohen''s d for similarity comparison (1 Hz FL): %.6f\n', effect_1Hz);
  707. %% Dwell times during CO2 on/off
  708. [dwell_times_10Hz_on, dwell_times_10Hz_off] = deal(nan(optimal_num_clusters_combined, num_animals));
  709. [dwell_times_1Hz_on, dwell_times_1Hz_off] = deal(nan(optimal_num_clusters_combined, num_animals));
  710. [dwell_times_fMRI_on, dwell_times_fMRI_off] = deal(nan(optimal_num_clusters_fMRI, num_animals));
  712. for animal = 1:num_animals
  713. [starts_10Hz, starts_1Hz, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  714. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  715. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  716. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  718. timing_info_10Hz = build_timing_mask(num_windows_10Hz, starts_10Hz, starts_10Hz + co2_duration);
  719. timing_info_1Hz = build_timing_mask(num_windows_1Hz, starts_1Hz, starts_1Hz + co2_duration);
  720. timing_info_fMRI = build_timing_mask(num_windows_fMRI, starts_fMRI, starts_fMRI + co2_duration);
  722. dwell_times_10Hz_on(:, animal) = calculate_dwell_times(idx_10Hz(:, animal), timing_info_10Hz, optimal_num_clusters_combined);
  723. dwell_times_10Hz_off(:, animal) = calculate_dwell_times(idx_10Hz(:, animal), ~timing_info_10Hz, optimal_num_clusters_combined);
  724. dwell_times_1Hz_on(:, animal) = calculate_dwell_times(idx_1Hz(:, animal), timing_info_1Hz, optimal_num_clusters_combined);
  725. dwell_times_1Hz_off(:, animal) = calculate_dwell_times(idx_1Hz(:, animal), ~timing_info_1Hz, optimal_num_clusters_combined);
  726. dwell_times_fMRI_on(:, animal) = calculate_dwell_times(idx_fMRI(:, animal), timing_info_fMRI, optimal_num_clusters_fMRI);
  727. dwell_times_fMRI_off(:, animal) = calculate_dwell_times(idx_fMRI(:, animal), ~timing_info_fMRI, optimal_num_clusters_fMRI);
  728. end
  730. p_values_dwell_10Hz = nan(optimal_num_clusters_combined, 1);
  731. p_values_dwell_1Hz = nan(optimal_num_clusters_combined, 1);
  732. p_values_dwell_fMRI = nan(optimal_num_clusters_fMRI, 1);
  733. for i = 1:optimal_num_clusters_combined
  734. p_values_dwell_10Hz(i) = safe_paired_ttest(dwell_times_10Hz_on(i, :), dwell_times_10Hz_off(i, :));
  735. p_values_dwell_1Hz(i) = safe_paired_ttest(dwell_times_1Hz_on(i, :), dwell_times_1Hz_off(i, :));
  736. end
  737. for i = 1:optimal_num_clusters_fMRI
  738. p_values_dwell_fMRI(i) = safe_paired_ttest(dwell_times_fMRI_on(i, :), dwell_times_fMRI_off(i, :));
  739. end
  741. fprintf('P-values for dwell times - FL 10 Hz:\n'); disp(p_values_dwell_10Hz);
  742. fprintf('P-values for dwell times - FL 1 Hz:\n'); disp(p_values_dwell_1Hz);
  743. fprintf('P-values for dwell times - fMRI:\n'); disp(p_values_dwell_fMRI);
  745. mean_dwell_times_10Hz_on = mean(dwell_times_10Hz_on, 2, 'omitnan');
  746. mean_dwell_times_10Hz_off = mean(dwell_times_10Hz_off, 2, 'omitnan');
  747. mean_dwell_times_1Hz_on = mean(dwell_times_1Hz_on, 2, 'omitnan');
  748. mean_dwell_times_1Hz_off = mean(dwell_times_1Hz_off, 2, 'omitnan');
  749. mean_dwell_times_fMRI_on = mean(dwell_times_fMRI_on, 2, 'omitnan');
  750. mean_dwell_times_fMRI_off = mean(dwell_times_fMRI_off, 2, 'omitnan');
  752. sem_dwell_times_10Hz_on = std(dwell_times_10Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  753. sem_dwell_times_10Hz_off = std(dwell_times_10Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  754. sem_dwell_times_1Hz_on = std(dwell_times_1Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  755. sem_dwell_times_1Hz_off = std(dwell_times_1Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  756. sem_dwell_times_fMRI_on = std(dwell_times_fMRI_on, 0, 2, 'omitnan') / sqrt(num_animals);
  757. sem_dwell_times_fMRI_off = std(dwell_times_fMRI_off, 0, 2, 'omitnan') / sqrt(num_animals);
  759. figure('Position', [0 0 400 900]);
  760. plot_grouped_bars_with_paired_points(3, 1, 1, mean_dwell_times_10Hz_on, mean_dwell_times_10Hz_off, sem_dwell_times_10Hz_on, sem_dwell_times_10Hz_off, dwell_times_10Hz_on, dwell_times_10Hz_off, 'Dwell Times - FL 10 Hz', 'Time (windows)', 1:optimal_num_clusters_combined, []);
  761. plot_grouped_bars_with_paired_points(3, 1, 2, mean_dwell_times_1Hz_on, mean_dwell_times_1Hz_off, sem_dwell_times_1Hz_on, sem_dwell_times_1Hz_off, dwell_times_1Hz_on, dwell_times_1Hz_off, 'Dwell Times - FL 1 Hz', 'Time (windows)', 1:optimal_num_clusters_combined, []);
  762. plot_grouped_bars_with_paired_points(3, 1, 3, mean_dwell_times_fMRI_on, mean_dwell_times_fMRI_off, sem_dwell_times_fMRI_on, sem_dwell_times_fMRI_off, dwell_times_fMRI_on, dwell_times_fMRI_off, 'Dwell Times - fMRI', 'Time (windows)', 1:optimal_num_clusters_fMRI, []);
  763. sgtitle('Dwell times for FL 10 Hz, FL 1 Hz, and fMRI (CO2 on/off)');
  765. %% Entry and exit rates during CO2 on/off
  766. [entry_rate_10Hz_on, entry_rate_10Hz_off] = deal(nan(optimal_num_clusters_combined, num_animals));
  767. [entry_rate_1Hz_on, entry_rate_1Hz_off] = deal(nan(optimal_num_clusters_combined, num_animals));
  768. [entry_rate_fMRI_on, entry_rate_fMRI_off] = deal(nan(optimal_num_clusters_fMRI, num_animals));
  770. [exit_rate_10Hz_on, exit_rate_10Hz_off] = deal(nan(optimal_num_clusters_combined, num_animals));
  771. [exit_rate_1Hz_on, exit_rate_1Hz_off] = deal(nan(optimal_num_clusters_combined, num_animals));
  772. [exit_rate_fMRI_on, exit_rate_fMRI_off] = deal(nan(optimal_num_clusters_fMRI, num_animals));
  774. for animal = 1:num_animals
  775. [starts_10Hz, starts_1Hz, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  776. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  777. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  778. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  780. timing_info_10Hz = build_timing_mask(num_windows_10Hz, starts_10Hz, starts_10Hz + co2_duration);
  781. timing_info_1Hz = build_timing_mask(num_windows_1Hz, starts_1Hz, starts_1Hz + co2_duration);
  782. timing_info_fMRI = build_timing_mask(num_windows_fMRI, starts_fMRI, starts_fMRI + co2_duration);
  784. entry_rate_10Hz_on(:, animal) = calculate_entry_rates(idx_10Hz(:, animal), timing_info_10Hz, optimal_num_clusters_combined);
  785. entry_rate_10Hz_off(:, animal) = calculate_entry_rates(idx_10Hz(:, animal), ~timing_info_10Hz, optimal_num_clusters_combined);
  786. exit_rate_10Hz_on(:, animal) = calculate_exit_rates(idx_10Hz(:, animal), timing_info_10Hz, optimal_num_clusters_combined);
  787. exit_rate_10Hz_off(:, animal) = calculate_exit_rates(idx_10Hz(:, animal), ~timing_info_10Hz, optimal_num_clusters_combined);
  789. entry_rate_1Hz_on(:, animal) = calculate_entry_rates(idx_1Hz(:, animal), timing_info_1Hz, optimal_num_clusters_combined);
  790. entry_rate_1Hz_off(:, animal) = calculate_entry_rates(idx_1Hz(:, animal), ~timing_info_1Hz, optimal_num_clusters_combined);
  791. exit_rate_1Hz_on(:, animal) = calculate_exit_rates(idx_1Hz(:, animal), timing_info_1Hz, optimal_num_clusters_combined);
  792. exit_rate_1Hz_off(:, animal) = calculate_exit_rates(idx_1Hz(:, animal), ~timing_info_1Hz, optimal_num_clusters_combined);
  794. entry_rate_fMRI_on(:, animal) = calculate_entry_rates(idx_fMRI(:, animal), timing_info_fMRI, optimal_num_clusters_fMRI);
  795. entry_rate_fMRI_off(:, animal) = calculate_entry_rates(idx_fMRI(:, animal), ~timing_info_fMRI, optimal_num_clusters_fMRI);
  796. exit_rate_fMRI_on(:, animal) = calculate_exit_rates(idx_fMRI(:, animal), timing_info_fMRI, optimal_num_clusters_fMRI);
  797. exit_rate_fMRI_off(:, animal) = calculate_exit_rates(idx_fMRI(:, animal), ~timing_info_fMRI, optimal_num_clusters_fMRI);
  798. end
  800. p_values_entry_10Hz = nan(optimal_num_clusters_combined, 1);
  801. p_values_entry_1Hz = nan(optimal_num_clusters_combined, 1);
  802. p_values_entry_fMRI = nan(optimal_num_clusters_fMRI, 1);
  803. p_values_exit_10Hz = nan(optimal_num_clusters_combined, 1);
  804. p_values_exit_1Hz = nan(optimal_num_clusters_combined, 1);
  805. p_values_exit_fMRI = nan(optimal_num_clusters_fMRI, 1);
  806. for i = 1:optimal_num_clusters_combined
  807. p_values_entry_10Hz(i) = safe_paired_ttest(entry_rate_10Hz_on(i, :), entry_rate_10Hz_off(i, :));
  808. p_values_entry_1Hz(i) = safe_paired_ttest(entry_rate_1Hz_on(i, :), entry_rate_1Hz_off(i, :));
  809. p_values_exit_10Hz(i) = safe_paired_ttest(exit_rate_10Hz_on(i, :), exit_rate_10Hz_off(i, :));
  810. p_values_exit_1Hz(i) = safe_paired_ttest(exit_rate_1Hz_on(i, :), exit_rate_1Hz_off(i, :));
  811. end
  812. for i = 1:optimal_num_clusters_fMRI
  813. p_values_entry_fMRI(i) = safe_paired_ttest(entry_rate_fMRI_on(i, :), entry_rate_fMRI_off(i, :));
  814. p_values_exit_fMRI(i) = safe_paired_ttest(exit_rate_fMRI_on(i, :), exit_rate_fMRI_off(i, :));
  815. end
  817. fprintf('P-values for entry rates - FL 10 Hz:\n'); disp(p_values_entry_10Hz);
  818. fprintf('P-values for entry rates - FL 1 Hz:\n'); disp(p_values_entry_1Hz);
  819. fprintf('P-values for entry rates - fMRI:\n'); disp(p_values_entry_fMRI);
  820. fprintf('P-values for exit rates - FL 10 Hz:\n'); disp(p_values_exit_10Hz);
  821. fprintf('P-values for exit rates - FL 1 Hz:\n'); disp(p_values_exit_1Hz);
  822. fprintf('P-values for exit rates - fMRI:\n'); disp(p_values_exit_fMRI);
  824. mean_entry_rates_10Hz_on = mean(entry_rate_10Hz_on, 2, 'omitnan');
  825. mean_entry_rates_10Hz_off = mean(entry_rate_10Hz_off, 2, 'omitnan');
  826. mean_entry_rates_1Hz_on = mean(entry_rate_1Hz_on, 2, 'omitnan');
  827. mean_entry_rates_1Hz_off = mean(entry_rate_1Hz_off, 2, 'omitnan');
  828. mean_entry_rates_fMRI_on = mean(entry_rate_fMRI_on, 2, 'omitnan');
  829. mean_entry_rates_fMRI_off = mean(entry_rate_fMRI_off, 2, 'omitnan');
  831. sem_entry_rates_10Hz_on = std(entry_rate_10Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  832. sem_entry_rates_10Hz_off = std(entry_rate_10Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  833. sem_entry_rates_1Hz_on = std(entry_rate_1Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  834. sem_entry_rates_1Hz_off = std(entry_rate_1Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  835. sem_entry_rates_fMRI_on = std(entry_rate_fMRI_on, 0, 2, 'omitnan') / sqrt(num_animals);
  836. sem_entry_rates_fMRI_off = std(entry_rate_fMRI_off, 0, 2, 'omitnan') / sqrt(num_animals);
  838. mean_exit_rates_10Hz_on = mean(exit_rate_10Hz_on, 2, 'omitnan');
  839. mean_exit_rates_10Hz_off = mean(exit_rate_10Hz_off, 2, 'omitnan');
  840. mean_exit_rates_1Hz_on = mean(exit_rate_1Hz_on, 2, 'omitnan');
  841. mean_exit_rates_1Hz_off = mean(exit_rate_1Hz_off, 2, 'omitnan');
  842. mean_exit_rates_fMRI_on = mean(exit_rate_fMRI_on, 2, 'omitnan');
  843. mean_exit_rates_fMRI_off = mean(exit_rate_fMRI_off, 2, 'omitnan');
  845. sem_exit_rates_10Hz_on = std(exit_rate_10Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  846. sem_exit_rates_10Hz_off = std(exit_rate_10Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  847. sem_exit_rates_1Hz_on = std(exit_rate_1Hz_on, 0, 2, 'omitnan') / sqrt(num_animals);
  848. sem_exit_rates_1Hz_off = std(exit_rate_1Hz_off, 0, 2, 'omitnan') / sqrt(num_animals);
  849. sem_exit_rates_fMRI_on = std(exit_rate_fMRI_on, 0, 2, 'omitnan') / sqrt(num_animals);
  850. sem_exit_rates_fMRI_off = std(exit_rate_fMRI_off, 0, 2, 'omitnan') / sqrt(num_animals);
  852. figure('Position', [0 0 400 900]);
  853. plot_grouped_bars_with_paired_points(3, 1, 1, mean_entry_rates_10Hz_on, mean_entry_rates_10Hz_off, sem_entry_rates_10Hz_on, sem_entry_rates_10Hz_off, entry_rate_10Hz_on, entry_rate_10Hz_off, 'Entry Rates - FL 10 Hz', 'Entry Rate (entries/window)', 1:optimal_num_clusters_combined, []);
  854. plot_grouped_bars_with_paired_points(3, 1, 2, mean_entry_rates_1Hz_on, mean_entry_rates_1Hz_off, sem_entry_rates_1Hz_on, sem_entry_rates_1Hz_off, entry_rate_1Hz_on, entry_rate_1Hz_off, 'Entry Rates - FL 1 Hz', 'Entry Rate (entries/window)', 1:optimal_num_clusters_combined, []);
  855. plot_grouped_bars_with_paired_points(3, 1, 3, mean_entry_rates_fMRI_on, mean_entry_rates_fMRI_off, sem_entry_rates_fMRI_on, sem_entry_rates_fMRI_off, entry_rate_fMRI_on, entry_rate_fMRI_off, 'Entry Rates - fMRI', 'Entry Rate (entries/window)', 1:optimal_num_clusters_fMRI, []);
  856. sgtitle('Entry rates for FL 10 Hz, FL 1 Hz, and fMRI (CO2 on/off)');
  858. figure('Position', [0 0 400 900]);
  859. plot_grouped_bars_with_paired_points(3, 1, 1, mean_exit_rates_10Hz_on, mean_exit_rates_10Hz_off, sem_exit_rates_10Hz_on, sem_exit_rates_10Hz_off, exit_rate_10Hz_on, exit_rate_10Hz_off, 'Exit Rates - FL 10 Hz', 'Exit Rate (exits/window)', 1:optimal_num_clusters_combined, []);
  860. plot_grouped_bars_with_paired_points(3, 1, 2, mean_exit_rates_1Hz_on, mean_exit_rates_1Hz_off, sem_exit_rates_1Hz_on, sem_exit_rates_1Hz_off, exit_rate_1Hz_on, exit_rate_1Hz_off, 'Exit Rates - FL 1 Hz', 'Exit Rate (exits/window)', 1:optimal_num_clusters_combined, []);
  861. plot_grouped_bars_with_paired_points(3, 1, 3, mean_exit_rates_fMRI_on, mean_exit_rates_fMRI_off, sem_exit_rates_fMRI_on, sem_exit_rates_fMRI_off, exit_rate_fMRI_on, exit_rate_fMRI_off, 'Exit Rates - fMRI', 'Exit Rate (exits/window)', 1:optimal_num_clusters_fMRI, []);
  862. sgtitle('Exit rates for FL 10 Hz, FL 1 Hz, and fMRI (CO2 on/off)');
  864. %% Cross-correlation between FL and fMRI state transitions
  865. thres = 0.01;
  866. state_transitions_FL = zeros(num_windows_10Hz - 1, num_animals);
  867. state_transitions_fMRI = zeros(num_windows_fMRI - 1, num_animals);
  868. for animal = 1:num_animals
  869. state_transitions_FL(:, animal) = diff(idx_10Hz(:, animal)) ~= 0;
  870. state_transitions_fMRI(:, animal) = diff(idx_fMRI(:, animal)) ~= 0;
  871. end
  873. max_lag = 6;
  874. cross_corr_FL_to_fMRI = zeros(2 * max_lag + 1, num_animals);
  875. for animal = 1:num_animals
  876. [cross_corr, lags_xcorr] = xcorr(state_transitions_FL(:, animal), state_transitions_fMRI(:, animal), max_lag, 'coeff');
  877. cross_corr_FL_to_fMRI(:, animal) = cross_corr;
  878. end
  879. avg_cross_corr_FL_to_fMRI = mean(cross_corr_FL_to_fMRI, 2, 'omitnan');
  881. p_values_FL_to_fMRI = nan(2 * max_lag + 1, 1);
  882. for lag_idx = 1:(2 * max_lag + 1)
  883. p_values_FL_to_fMRI(lag_idx) = safe_one_sample_ttest(cross_corr_FL_to_fMRI(lag_idx, :));
  884. end
  886. figure;
  887. subplot(2, 1, 1); hold on;
  888. plot(lags_xcorr, avg_cross_corr_FL_to_fMRI, '-b', 'LineWidth', 1.5);
  889. sig_lags = find(p_values_FL_to_fMRI < thres);
  890. plot(lags_xcorr(sig_lags), avg_cross_corr_FL_to_fMRI(sig_lags), 'r*', 'MarkerSize', 10);
  891. xlabel('Lag (frames)'); ylabel('Cross-Correlation'); title('Cross-correlation between FL and fMRI state transitions'); grid on; hold off;
  893. cross_corr_FL_to_fMRI_on = zeros(2 * max_lag + 1, num_animals);
  894. cross_corr_FL_to_fMRI_off = zeros(2 * max_lag + 1, num_animals);
  895. for animal = 1:num_animals
  896. [starts_10Hz, ~, starts_fMRI] = get_co2_start_indices_for_animal(animal, ...
  897. co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, ...
  898. co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, ...
  899. co2_on_indices_fMRI, co2_on_indices_fMRI_sub3);
  901. timing_info_10Hz = build_timing_mask(num_windows_10Hz, starts_10Hz, starts_10Hz + co2_duration);
  902. timing_info_fMRI = build_timing_mask(num_windows_fMRI, starts_fMRI, starts_fMRI + co2_duration);
  904. cross_corr_FL_to_fMRI_on(:, animal) = xcorr(state_transitions_FL(timing_info_10Hz(1:end-1), animal), state_transitions_fMRI(timing_info_fMRI(1:end-1), animal), max_lag, 'coeff');
  905. cross_corr_FL_to_fMRI_off(:, animal) = xcorr(state_transitions_FL(~timing_info_10Hz(1:end-1), animal), state_transitions_fMRI(~timing_info_fMRI(1:end-1), animal), max_lag, 'coeff');
  906. end
  908. avg_cross_corr_FL_to_fMRI_on = mean(cross_corr_FL_to_fMRI_on, 2, 'omitnan');
  909. avg_cross_corr_FL_to_fMRI_off = mean(cross_corr_FL_to_fMRI_off, 2, 'omitnan');
  911. p_values_FL_to_fMRI_on = nan(2 * max_lag + 1, 1);
  912. p_values_FL_to_fMRI_off = nan(2 * max_lag + 1, 1);
  913. for lag_idx = 1:(2 * max_lag + 1)
  914. p_values_FL_to_fMRI_on(lag_idx) = safe_one_sample_ttest(cross_corr_FL_to_fMRI_on(lag_idx, :));
  915. p_values_FL_to_fMRI_off(lag_idx) = safe_one_sample_ttest(cross_corr_FL_to_fMRI_off(lag_idx, :));
  916. end
  918. subplot(2, 1, 2); hold on;
  919. plot(lags_xcorr, avg_cross_corr_FL_to_fMRI_on, '-g', 'LineWidth', 1.5);
  920. plot(lags_xcorr, avg_cross_corr_FL_to_fMRI_off, '-r', 'LineWidth', 1.5);
  921. plot(lags_xcorr(p_values_FL_to_fMRI_on < thres), avg_cross_corr_FL_to_fMRI_on(p_values_FL_to_fMRI_on < thres), 'g*', 'MarkerSize', 8);
  922. plot(lags_xcorr(p_values_FL_to_fMRI_off < thres), avg_cross_corr_FL_to_fMRI_off(p_values_FL_to_fMRI_off < thres), 'r*', 'MarkerSize', 8);
  923. xlabel('Lag (frames)'); ylabel('Cross-Correlation'); title('Cross-correlation between FL and fMRI state transitions (CO2 on/off)'); grid on; hold off; ylim([0 0.1]);
  924. sgtitle('Cross-correlation between FL and fMRI state transitions');
  926. %% Conditional co-occurrence matrices across temporal lags
  927. lags = 0;
  928. [T, N] = size(idx_10Hz);
  929. K_FL = max(idx_10Hz(:));
  930. K_fMRI = max(idx_fMRI(:));
  931. L = numel(lags);
  933. CoocOn = zeros(K_fMRI, K_FL, L);
  934. CoocOff = zeros(K_fMRI, K_FL, L);
  935. P_on_fMRIgivenFL = nan(K_fMRI, K_FL, L);
  936. P_off_fMRIgivenFL = nan(K_fMRI, K_FL, L);
  937. P_on_FLgivenfMRI = nan(K_fMRI, K_FL, L);
  938. P_off_FLgivenfMRI = nan(K_fMRI, K_FL, L);
  940. for animal = 1:N
  941. if animal == 3
  942. on_idx = co2_on_indices_10Hz_sub3;
  943. else
  944. on_idx = co2_on_indices_10Hz;
  945. end
  947. on_mask = false(T, 1);
  948. on_mask(on_idx(on_idx <= T)) = true;
  949. off_mask = ~on_mask;
  951. for li = 1:L
  952. lag = lags(li);
  953. for t = 1:T
  954. tFL = t - lag;
  955. if tFL < 1 || tFL > T || t > size(idx_fMRI, 1)
  956. continue;
  957. end
  958. fL = idx_10Hz(tFL, animal);
  959. fR = idx_fMRI(t, animal);
  960. if fL > 0 && fR > 0
  961. if on_mask(t)
  962. CoocOn(fR, fL, li) = CoocOn(fR, fL, li) + 1;
  963. elseif off_mask(t)
  964. CoocOff(fR, fL, li) = CoocOff(fR, fL, li) + 1;
  965. end
  966. end
  967. end
  968. end
  969. end
  971. for li = 1:L
  972. C_on = CoocOn(:, :, li);
  973. C_off = CoocOff(:, :, li);
  974. P_on_fMRIgivenFL(:, :, li) = normalize_columns(C_on);
  975. P_off_fMRIgivenFL(:, :, li) = normalize_columns(C_off);
  976. P_on_FLgivenfMRI(:, :, li) = normalize_rows_2d(C_on);
  977. P_off_FLgivenfMRI(:, :, li) = normalize_rows_2d(C_off);
  978. end
  980. figure('Position', [100 100 1200 600]);
  981. for li = 1:L
  982. lag = lags(li);
  984. subplot(4, L, li);
  985. imagesc(P_on_fMRIgivenFL(:, :, li)); axis square; colorbar;
  986. title(sprintf('P(fMRI|FL) On, lag=%ds', lag));
  987. if li == 1, ylabel('fMRI state'); end
  988. xlabel('FL state');
  990. subplot(4, L, li + L);
  991. imagesc(P_off_fMRIgivenFL(:, :, li)); axis square; colorbar;
  992. title(sprintf('P(fMRI|FL) Off, lag=%ds', lag));
  993. xlabel('FL state');
  995. subplot(4, L, li + 2 * L);
  996. imagesc(P_on_FLgivenfMRI(:, :, li)); axis square; colorbar;
  997. title(sprintf('P(FL|fMRI) On, lag=%ds', lag));
  998. if li == 1, ylabel('fMRI state'); end
  999. xlabel('FL state');
  1001. subplot(4, L, li + 3 * L);
  1002. imagesc(P_off_FLgivenfMRI(:, :, li)); axis square; colorbar;
  1003. title(sprintf('P(FL|fMRI) Off, lag=%ds', lag));
  1004. xlabel('FL state');
  1005. end
  1006. sgtitle('Conditional co-occurrence matrices across lags and CO2 conditions');
  1008. example_FL = 5;
  1009. example_fMRI = 1;
  1010. figure('Position', [200 200 800 300]);
  1011. for li = 1:L
  1012. subplot(2, L, li);
  1013. bar(P_on_fMRIgivenFL(:, example_FL, li), 'b'); hold on;
  1014. bar(P_off_fMRIgivenFL(:, example_FL, li), 'r'); hold off;
  1015. ylim([0 1]);
  1016. title(sprintf('P(fMRI|FL=%d), lag=%ds', example_FL, lags(li)));
  1017. if li == 1, ylabel('Probability'); end
  1018. xlabel('fMRI state'); legend('On', 'Off');
  1020. subplot(2, L, li + L);
  1021. bar(P_on_FLgivenfMRI(example_fMRI, :, li), 'b'); hold on;
  1022. bar(P_off_FLgivenfMRI(example_fMRI, :, li), 'r'); hold off;
  1023. ylim([0 1]);
  1024. title(sprintf('P(FL|fMRI=%d), lag=%ds', example_fMRI, lags(li)));
  1025. xlabel('FL state');
  1026. end
  1027. sgtitle(sprintf('Example conditional distributions for FL=%d and fMRI=%d', example_FL, example_fMRI));
  1029. %% Local functions
  1030. function cmap = load_colormap_safe(filename, varname, fallback)
  1031. if exist(filename, 'file')
  1032. S = load(filename);
  1033. if isfield(S, varname)
  1034. cmap = S.(varname);
  1035. return;
  1036. end
  1037. end
  1038. cmap = fallback;
  1039. end
  1041. function [starts_10Hz, starts_1Hz, starts_fMRI] = get_co2_start_indices_for_animal(animal, co2_on_indices_10Hz, co2_on_indices_10Hz_sub3, co2_on_indices_1Hz, co2_on_indices_1Hz_sub3, co2_on_indices_fMRI, co2_on_indices_fMRI_sub3)
  1042. if animal == 3
  1043. starts_10Hz = co2_on_indices_10Hz_sub3;
  1044. starts_1Hz = co2_on_indices_1Hz_sub3;
  1045. starts_fMRI = co2_on_indices_fMRI_sub3;
  1046. else
  1047. starts_10Hz = co2_on_indices_10Hz;
  1048. starts_1Hz = co2_on_indices_1Hz;
  1049. starts_fMRI = co2_on_indices_fMRI;
  1050. end
  1051. end
  1053. function mask = build_timing_mask(num_windows, starts, ends_)
  1054. mask = false(num_windows, 1);
  1055. for i = 1:length(starts)
  1056. s = max(1, round(starts(i)));
  1057. e = min(num_windows, round(ends_(i)));
  1058. if s <= e
  1059. mask(s:e) = true;
  1060. end
  1061. end
  1062. end
  1064. function [idx_out, removed_states] = remove_single_subject_states(idx_in)
  1065. idx_out = idx_in;
  1066. removed_states = [];
  1067. states = unique(idx_in(:));
  1068. states = states(~isnan(states));
  1069. for state = states'
  1070. subject_count = sum(any(idx_in == state, 1));
  1071. if subject_count == 1
  1072. idx_out(idx_out == state) = NaN;
  1073. removed_states(end + 1) = state; %#ok<AGROW>
  1074. end
  1075. end
  1076. end
  1078. function idx_out = replace_nan_states_with_closest(idx_in, distance_matrix)
  1079. idx_out = idx_in;
  1080. for animal = 1:size(idx_out, 2)
  1081. for t = 1:size(idx_out, 1)
  1082. if isnan(idx_out(t, animal))
  1083. non_nan_states = idx_out(~isnan(idx_out(:, animal)), animal);
  1084. if isempty(non_nan_states)
  1085. continue;
  1086. end
  1087. min_distance = inf;
  1088. similar_state = NaN;
  1089. for s = non_nan_states'
  1090. distances = distance_matrix(s, :);
  1091. [min_dist, state] = min(distances);
  1092. if min_dist < min_distance
  1093. min_distance = min_dist;
  1094. similar_state = state;
  1095. end
  1096. end
  1097. idx_out(t, animal) = similar_state;
  1098. end
  1099. end
  1100. end
  1101. end
  1103. function counts = count_states_in_range(state_vector, start_idx, end_idx, num_states)
  1104. counts = zeros(num_states, 1);
  1105. start_idx = max(1, round(start_idx));
  1106. end_idx = min(length(state_vector), round(end_idx));
  1107. if start_idx > end_idx
  1108. return;
  1109. end
  1110. for s = 1:num_states
  1111. counts(s) = sum(state_vector(start_idx:end_idx) == s);
  1112. end
  1113. end
  1115. function trans_counts = count_transitions_in_range(state_vector, start_idx, end_idx, num_states)
  1116. trans_counts = zeros(num_states, num_states);
  1117. start_idx = max(1, round(start_idx));
  1118. end_idx = min(length(state_vector), round(end_idx));
  1119. if start_idx >= end_idx
  1120. return;
  1121. end
  1122. for t = start_idx:(end_idx - 1)
  1123. current_state = state_vector(t);
  1124. next_state = state_vector(t + 1);
  1125. if ~isnan(current_state) && ~isnan(next_state)
  1126. trans_counts(current_state, next_state) = trans_counts(current_state, next_state) + 1;
  1127. end
  1128. end
  1129. end
  1131. function M = normalize_columns(M)
  1132. denom = sum(M, 1, 'omitnan');
  1133. denom(denom == 0) = NaN;
  1134. M = M ./ denom;
  1135. end
  1137. function P = normalize_rows_2d(M)
  1138. denom = sum(M, 2, 'omitnan');
  1139. denom(denom == 0) = NaN;
  1140. P = M ./ denom;
  1141. P(isnan(P)) = 0;
  1142. end
  1144. function pct = safe_percent_by_column(counts)
  1145. denom = sum(counts, 1);
  1146. denom(denom == 0) = NaN;
  1147. pct = counts ./ denom * 100;
  1148. pct(isnan(pct)) = 0;
  1149. end
  1151. function T = normalize_transition_tensor(count_tensor)
  1152. T = zeros(size(count_tensor));
  1153. for animal = 1:size(count_tensor, 3)
  1154. for i = 1:size(count_tensor, 1)
  1155. row_sum = sum(count_tensor(i, :, animal));
  1156. if row_sum > 0
  1157. T(i, :, animal) = count_tensor(i, :, animal) / row_sum;
  1158. end
  1159. end
  1160. end
  1161. end
  1163. function p = safe_paired_ttest(x, y)
  1164. x = x(:);
  1165. y = y(:);
  1166. valid = ~isnan(x) & ~isnan(y);
  1167. if nnz(valid) < 2 || numel(unique(x(valid) - y(valid))) < 1
  1168. p = NaN;
  1169. return;
  1170. end
  1171. [~, p] = ttest(x(valid), y(valid));
  1172. end
  1174. function p = safe_one_sample_ttest(x)
  1175. x = x(:);
  1176. x = x(~isnan(x));
  1177. if numel(x) < 2 || numel(unique(x)) < 2
  1178. p = NaN;
  1179. return;
  1180. end
  1181. [~, p] = ttest(x);
  1182. end
  1184. function plot_grouped_bars_with_paired_points(nr, nc, idx_subplot, mean_on, mean_off, sem_on, sem_off, data_on, data_off, plot_title, ylab, xtick_vals, y_limits)
  1185. subplot(nr, nc, idx_subplot);
  1186. bar_data = [mean_on, mean_off];
  1187. bar_handle = bar(bar_data, 'grouped');
  1188. hold on;
  1189. bar_handle(1).FaceColor = [0.2 0.6 0.8];
  1190. bar_handle(2).FaceColor = [0.8 0.4 0.4];
  1192. x_on = bar_handle(1).XEndPoints;
  1193. x_off = bar_handle(2).XEndPoints;
  1195. errorbar(x_on, mean_on, sem_on, 'k', 'linestyle', 'none', 'LineWidth', 1);
  1196. errorbar(x_off, mean_off, sem_off, 'k', 'linestyle', 'none', 'LineWidth', 1);
  1198. jitter_amount = 0.05;
  1199. marker_size = 22;
  1200. line_color = [0.7 0.7 0.7];
  1202. for s = 1:numel(mean_on)
  1203. jitter_vec = (rand(size(data_on, 2), 1) - 0.5) * 2 * jitter_amount;
  1204. x_on_jit = x_on(s) + jitter_vec;
  1205. x_off_jit = x_off(s) + jitter_vec;
  1207. y_on = data_on(s, :)';
  1208. y_off = data_off(s, :)';
  1210. valid_pairs = ~isnan(y_on) & ~isnan(y_off);
  1211. for a = find(valid_pairs)'
  1212. plot([x_on_jit(a), x_off_jit(a)], [y_on(a), y_off(a)], '-', 'Color', line_color, 'LineWidth', 0.8);
  1213. end
  1215. scatter(x_on_jit(~isnan(y_on)), y_on(~isnan(y_on)), marker_size, 'filled', 'MarkerFaceColor', [0 0.25 0.55], 'MarkerFaceAlpha', 0.7);
  1216. scatter(x_off_jit(~isnan(y_off)), y_off(~isnan(y_off)), marker_size, 'filled', 'MarkerFaceColor', [0.55 0 0], 'MarkerFaceAlpha', 0.7);
  1217. end
  1219. title(plot_title);
  1220. xlabel('State');
  1221. ylabel(ylab);
  1222. xticks(xtick_vals);
  1223. legend({'CO2 On', 'CO2 Off'});
  1224. if ~isempty(y_limits)
  1225. ylim(y_limits);
  1226. end
  1227. hold off;
  1228. end
  1230. function plot_transition_matrix_with_sig(nr, nc, idx_subplot, M, p_values, threshold, ttl, xl, yl, climits, cmap)
  1231. subplot(nr, nc, idx_subplot);
  1232. imagesc(M);
  1233. colorbar;
  1234. axis square;
  1235. colormap(cmap);
  1236. caxis(climits);
  1237. title(ttl);
  1238. xlabel(xl);
  1239. ylabel(yl);
  1240. hold on;
  1241. for i = 1:size(M, 1)
  1242. for j = 1:size(M, 2)
  1243. if ~isnan(p_values(i, j)) && p_values(i, j) < threshold
  1244. text(j, i, '*', 'HorizontalAlignment', 'Center', 'VerticalAlignment', 'Middle', 'Color', 'k', 'FontSize', 14);
  1245. end
  1246. end
  1247. end
  1248. hold off;
  1249. end
  1251. function plot_matrix_with_sig(nr, nc, idx_subplot, M, p_values, threshold, ttl, xl, yl, climits, cmap)
  1252. subplot(nr, nc, idx_subplot);
  1253. imagesc(M);
  1254. colorbar;
  1255. axis square;
  1256. colormap(cmap);
  1257. caxis(climits);
  1258. title(ttl);
  1259. xlabel(xl);
  1260. ylabel(yl);
  1261. hold on;
  1262. for i = 1:size(M, 1)
  1263. for j = 1:size(M, 2)
  1264. if ~isnan(p_values(i, j)) && p_values(i, j) < threshold
  1265. text(j, i, '*', 'HorizontalAlignment', 'Center', 'VerticalAlignment', 'Middle', 'Color', 'k', 'FontSize', 14);
  1266. end
  1267. end
  1268. end
  1269. hold off;
  1270. end
  1272. function y = percentile_filter(x, lower_pct, upper_pct)
  1273. x = x(~isnan(x));
  1274. if isempty(x)
  1275. y = x;
  1276. return;
  1277. end
  1278. lo = prctile(x, lower_pct);
  1279. hi = prctile(x, upper_pct);
  1280. y = x(x >= lo & x <= hi);
  1281. end
  1283. function plot_distribution_comparison(x_on, x_off, ttl)
  1284. if exist('daviolinplot', 'file') == 2
  1285. daviolinplot([x_on; x_off], 'groups', [ones(size(x_on)); 2 * ones(size(x_off))]);
  1286. title(ttl);
  1287. return;
  1288. end
  1289. boxplot([x_on; x_off], [ones(size(x_on)); 2 * ones(size(x_off))]);
  1290. set(gca, 'XTickLabel', {'CO2 On', 'CO2 Off'});
  1291. ylabel('Similarity');
  1292. title(ttl);
  1293. end
  1295. function d = cohens_d_two_sample(x, y)
  1296. x = x(~isnan(x));
  1297. y = y(~isnan(y));
  1298. if numel(x) < 2 || numel(y) < 2
  1299. d = NaN;
  1300. return;
  1301. end
  1302. nx = numel(x);
  1303. ny = numel(y);
  1304. sx = std(x, 0);
  1305. sy = std(y, 0);
  1306. s_pooled = sqrt(((nx - 1) * sx^2 + (ny - 1) * sy^2) / (nx + ny - 2));
  1307. if s_pooled == 0
  1308. d = NaN;
  1309. else
  1310. d = (mean(x) - mean(y)) / s_pooled;
  1311. end
  1312. end
  1314. function dwell_times = calculate_dwell_times(state_sequence, timing_mask, num_states)
  1315. dwell_times = nan(num_states, 1);
  1316. state_sequence = state_sequence(:);
  1317. timing_mask = timing_mask(:);
  1318. for state = 1:num_states
  1319. run_lengths = [];
  1320. in_run = false;
  1321. current_len = 0;
  1322. for t = 1:length(state_sequence)
  1323. is_selected_state = timing_mask(t) && state_sequence(t) == state;
  1324. if is_selected_state
  1325. current_len = current_len + 1;
  1326. in_run = true;
  1327. elseif in_run
  1328. run_lengths(end + 1) = current_len; %#ok<AGROW>
  1329. current_len = 0;
  1330. in_run = false;
  1331. end
  1332. end
  1333. if in_run
  1334. run_lengths(end + 1) = current_len; %#ok<AGROW>
  1335. end
  1336. if ~isempty(run_lengths)
  1337. dwell_times(state) = mean(run_lengths);
  1338. end
  1339. end
  1340. end
  1342. function entry_rates = calculate_entry_rates(state_sequence, timing_mask, num_states)
  1343. entry_rates = nan(num_states, 1);
  1344. state_sequence = state_sequence(:);
  1345. timing_mask = timing_mask(:);
  1346. total_selected = sum(timing_mask);
  1347. if total_selected == 0
  1348. return;
  1349. end
  1350. for state = 1:num_states
  1351. entries = 0;
  1352. for t = 1:length(state_sequence)
  1353. if timing_mask(t) && state_sequence(t) == state
  1354. if t == 1 || ~timing_mask(t - 1) || state_sequence(t - 1) ~= state
  1355. entries = entries + 1;
  1356. end
  1357. end
  1358. end
  1359. entry_rates(state) = entries / total_selected;
  1360. end
  1361. end
  1363. function exit_rates = calculate_exit_rates(state_sequence, timing_mask, num_states)
  1364. exit_rates = nan(num_states, 1);
  1365. state_sequence = state_sequence(:);
  1366. timing_mask = timing_mask(:);
  1367. total_selected = sum(timing_mask);
  1368. if total_selected == 0
  1369. return;
  1370. end
  1371. for state = 1:num_states
  1372. exits = 0;
  1373. for t = 1:length(state_sequence)
  1374. if timing_mask(t) && state_sequence(t) == state
  1375. is_exit = (t == length(state_sequence)) || ~timing_mask(min(t + 1, length(timing_mask))) || state_sequence(min(t + 1, length(state_sequence))) ~= state;
  1376. if is_exit
  1377. exits = exits + 1;
  1378. end
  1379. end
  1380. end
  1381. exit_rates(state) = exits / total_selected;
  1382. end
  1383. end

dfc_state_analysis.m, under CC-BY-4.0 · at the source

Overview

Authors: Irmak Gezginer1,2, Yi Chen1,2, Valerio Zerbi3,4, Zhenyue Chen5, Daniel Razansky1,2
ORCID iDs: Irmak Gezginer, Yi Chen, Valerio Zerbi, Zhenyue Chen, Daniel Razansky
  1. Institute for Biomedical Engineering and Institute of Pharmacology and Toxicology, Faculty of Medicine, University of Zurich,Zurich, Switzerland
  2. Department of Information Technology and Electrical Engineering, Institute for Biomedical Engineering, ETH Zurich,Zurich, Switzerland
  3. Department of Psychiatry, Faculty of Medicine, University of Geneva,Geneva, Switzerland
  4. Department of Basic Neurosciences, Faculty of Medicine, University of Geneva,Geneva, Switzerland
  5. Institute of Precision Optical Engineering, School of Physics Science and Engineering, Tongji University,Shanghai, China
Institutions: University of Zurich (Switzerland); Institute for Biomedical Engineering (Switzerland); University of Geneva (Switzerland); Tongji University (China)
Journal: Nature communications, volume 17, issue 1, article 5158
Dates: received 18 February 2025; accepted 31 March 2026; published online 13 April 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1038/s41467-026-71742-z · PMID 41974737 · PMCID PMC13249851 · OpenAlex W7154072756
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: fMRI (modality), mouse (organism)
Methods: Spectral & time-frequency, Connectivity, Statistics, Machine learning, Preprocessing, fMRI & imaging, Single-unit activity, calcium imaging, Smoothing, state filtering, decompositions
Keywords: Functional magnetic resonance imaging, Fluorescence imaging, Neuro-vascular interactions
MeSH: Brain*, Hemodynamics*, Hypercapnia*, Neurons*, Neurovascular Coupling*, Animals, Calcium, Carbon Dioxide, Magnetic Resonance Imaging, Male, Mice, Mice, Inbred C57BL, Oxygen (* major topic)
Topic: Neuroscience of respiration and sleep (Endocrine and Autonomic Systems, Neuroscience), according to OpenAlex
Funding: Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (Swiss National Science Foundation) (310030_192757)
Citations: cited by 1 paper (Europe PMC); 50 references in the paper

Abstract

Neurovascular coupling (NVC) underpins the interpretation of hemodynamic signals as proxies for neural activity, yet its response to metabolic perturbations remains poorly understood. Here, we leverage concurrent fluorescence calcium imaging and functional magnetic resonance imaging in mice expressing genetically encoded calcium indicators to dissect how elevated CO₂ levels reshape the interplay between neural and vascular responses. Our findings indicate that hypercapnia induces opposing trends in calcium and blood-oxygen-level-dependent (BOLD) responses, accompanied by global desynchronization of brain activity. Additionally, 5% CO₂ suppressed sensory-evoked BOLD and hemoglobin responses, while neuronal and astrocytic activity remained unaffected. Dynamic functional connectivity and co-activation pattern analyses further reveal a dissociation in the coordination between BOLD hemodynamic responses and their underlying neural dynamics under hypercapnic conditions. Notably, we observed that hemodynamic responses, normally driven by neuronal signaling, get attenuated with BOLD signals no longer reflecting neural activity patterns in the brain. These findings demonstrate that hypercapnia can override conventional NVC relationships, compelling a reassessment of how BOLD contrast is interpreted under hypercapnic stress in preclinical and clinical settings. This holds particular relevance for chronic hypercapnia-related conditions where a deeper understanding of NVC disruption may inform improved diagnostic and therapeutic strategies.

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 1 match between paragraphs and lines of code.

Zenodo 19099592

License: CC-BY-4.0
State: the link answers, verified on 29 September 2026
Evidence: files inventoried
Languages: MATLAB (2)
Size: 2 files, 2 scripts
Software Heritage: not checked
Found in: “Code availability”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: Statistics and Machine Learning Toolbox (2 files), Image Processing Toolbox (1 file), Signal Processing Toolbox (1 file)
Availability: 1 check, the latest on 29 September 2026: the link answers (HTTP 200)
  • 29 September 2026: the link answers (HTTP 200)
2 files
At the source:

Code availability

Examples of custom codes used in this manuscript are deposited and publicly available on Zenodo at 10.5281/zenodo.19099592.

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;
  • 2 scripts, each with its path and the digest of its content;
  • 1 match 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

The data that support the findings of this study are available from the corresponding author. All data supporting the findings of this study are found within the paper and its Supplementary Information. 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, 29 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 5 authors, 3 keywords, 13 MeSH terms, 1 funder, 49 references.

Cite

This paper

Gezginer, I., Chen, Y., Zerbi, V., Chen, Z., & Razansky, D. (2026). Hypercapnia dissociates neuronal and hemodynamic responses impairing neurovascular coupling and functional brain connectivity. Nature communications, 17(1), 5158. https://doi.org/10.1038/s41467-026-71742-z

BibTeX

@article{gezginer2026hypercapnia,
author = {Gezginer, Irmak and Chen, Yi and Zerbi, Valerio and Chen, Zhenyue and Razansky, Daniel},
title = {{Hypercapnia dissociates neuronal and hemodynamic responses impairing neurovascular coupling and functional brain connectivity}},
journal = {Nature communications},
year = {2026},
month = apr,
volume = {17},
number = {1},
pages = {5158},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/s41467-026-71742-z},
url = {https://doi.org/10.1038/s41467-026-71742-z},
pmid = {41974737},
pmcid = {PMC13249851}
}

RIS

TY - JOUR
AU - Gezginer, Irmak
AU - Chen, Yi
AU - Zerbi, Valerio
AU - Chen, Zhenyue
AU - Razansky, Daniel
TI - Hypercapnia dissociates neuronal and hemodynamic responses impairing neurovascular coupling and functional brain connectivity
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/04/13
VL - 17
IS - 1
SP - 5158
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/s41467-026-71742-z
UR - https://doi.org/10.1038/s41467-026-71742-z
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41467-026-71742-z",
"type": "article-journal",
"title": "Hypercapnia dissociates neuronal and hemodynamic responses impairing neurovascular coupling and functional brain connectivity",
"container-title": "Nature communications",
"author": [
{
"family": "Gezginer",
"given": "Irmak"
},
{
"family": "Chen",
"given": "Yi"
},
{
"family": "Zerbi",
"given": "Valerio"
},
{
"family": "Chen",
"given": "Zhenyue"
},
{
"family": "Razansky",
"given": "Daniel"
}
],
"container-title-short": "Nat Commun",
"volume": "17",
"issue": "1",
"page": "5158",
"DOI": "10.1038/s41467-026-71742-z",
"PMID": "41974737",
"PMCID": "PMC13249851",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41467-026-71742-z",
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
13
]
]
}
}

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/s41467-026-71746-9
Double-helix optical point spread function enables real-time mesoscopic 3D functional microangiography in the living mouse brain and skull.
Journal: Nature communications
In common: mouse, 3 authors
[2] doi:10.64898/2026.03.12.710517 [code]
Cortical excitability inversely modulates fMRI connectivity via low-frequency neuronal coupling
Journal: bioRxiv (preprint)
In common: Image Processing Toolbox, Signal Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, 1 reference, author Valerio Zerbi
[3] doi:10.1038/s41592-026-03154-2 [code]
Simultaneous single-cell calcium imaging of neuronal population activity and brain-wide BOLD fMRI.
Journal: Nature methods
In common: Image Processing Toolbox, Signal Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, mouse, 2 references
[4] doi:10.1162/imag.a.1254 [code]
The individuality of single-frame functional brain connectivity.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: Signal Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, 3 references
[5] doi:10.1016/j.bpsgos.2026.100753 [code]
Dynamic Repertoire of Brain Networks in Mindfulness-Based Cognitive Therapy During Rumination: A Randomized Controlled Trial.
Journal: Biological psychiatry global open science
In common: Image Processing Toolbox, Signal Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, 2 references
[6] doi:10.1093/braincomms/fcag279 [code]
Network flexibility facilitates treatment-induced recovery in post-stroke aphasia.
Journal: Brain communications
In common: Image Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, 3 references
[7] doi:10.1162/imag.a.1278 [code]
Young and old adult brains experience opposite effects of acute sleep restriction on the functional connectivity network.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: Image Processing Toolbox, Signal Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, 2 references
[8] doi:10.1162/imag.a.1357
Biphasic acclimation for simultaneous wide-field fluorescent Ca&lt;sup&gt;2+&lt;/sup&gt; imaging and fMRI of awake mice.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: fMRI, mouse, 4 references
[9] doi:10.1162/imag.a.1202
Recurring transient brain-wide co-activation patterns from EEG spatially resembling time-averaged resting-state networks.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: fMRI, 4 references
[10] doi:10.1016/j.bbih.2026.101274 [code]
Brain dynamics of attentional, default-mode and limbic networks are disrupted at rest in post-COVID-19 syndrome.
Journal: Brain, behavior, & immunity - health
In common: Signal Processing Toolbox, Statistics and Machine Learning Toolbox, fMRI, 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.