First upload of Neurosynth_lexical_plot

CanlabCore/@fmri_data/neurosynth_lexical_plot.m

@@ -0,0 +1,98 @@
+function [image_by_feature_correlations, top_feature_tables]=neurosynth_lexical_plot(fmri_data)
+    % NEUROSYNTH_LEXICAL_PLOT Plots the correlations between fMRI data and lexical features.
+    %
+    % This function takes fMRI data and calculates the correlation between
+    % the data and lexical features using the Neurosynth toolbox. It then
+    % creates plots to visualize these correlations.
+    %
+    % USAGE:
+    %   [image_by_feature_correlations, top_feature_tables] = neurosynth_lexical_plot(fmri_data, varargin)
+    %
+    % INPUTS:
+    %   fmri_data : fMRI data to be analyzed.
+    %
+    % OUTPUTS:
+    %   image_by_feature_correlations : Correlations between images and features.
+    %   top_feature_tables            : Tables containing the top features and their scores.
+    %
+    % EXAMPLES:
+    %   [correlations, tables] = neurosynth_lexical_plot(fmri_data);
+    %
+    % Authors: Michael Sun, Ph.D. 
+    % Date: 05/30/2024
+
+
+    % Future: Add a between option for separate image analysis instead of
+    % replicates
+    % if ~isempty(varargin) & strcmp(varargin{1}, 'between')
+    %     [image_by_feature_correlations, top_feature_tables] = neurosynth_feature_labels(fmri_data, true, 'noverbose', 'display_output');
+    % else
+    %     [image_by_feature_correlations, top_feature_tables] = neurosynth_feature_labels(fmri_data, 'images_are_replicates', true, 'noverbose');
+    % end
+
+    [image_by_feature_correlations, top_feature_tables] = neurosynth_feature_labels(fmri_data, 'images_are_replicates', true, 'noverbose');
+    % Aggregate results for plot
+    lowwords = [top_feature_tables{1}.words_low(:)];
+    % disp(lowwords);
+
+    highwords = [top_feature_tables{1}.words_high(:)];
+    % disp(highwords);
+
+    r_low = top_feature_tables{1}.t_low;
+    r_high = top_feature_tables{1}.t_high;
+
+    r_to_plot = [r_high; r_low];
+    textlabels = [highwords; lowwords];
+    textlabels = strcat(textlabels, '  (', num2str(round(r_to_plot,1)), ')'); 
+
+    figure;
+    subplot(1,3,1)
+
+    % Old version used Tor's Wedge Plot, but this doesn't display well most
+    % of the time.
+    % tor_wedge_plot(r_to_plot, textlabels, 'outer_circle_radius', .3, 'nofigure', 'colors', {[1 .7 0] [.4 0 .8]}, 'bicolor', 'labelstyle', 'close');
+
+    hitbl=table(highwords, r_high, 'VariableNames', ["Word", "t"]);
+    lotbl=table(lowwords, -1*r_low, 'VariableNames', ["Word", "t"]);
+
+    % Horizontal Bar plot looks better:
+    % Sort the values in descending order and reorder the text labels accordingly
+    [sorted_values, sort_idx] = sort(r_to_plot, 'ascend');
+    sorted_labels = textlabels(sort_idx);
+
+    % Create a horizontal bar plot
+    h = barh(sorted_values, 'FaceColor', 'flat');
+
+    % Color the bars: green for positive values, red for negative values
+    colors = repmat([0, 1, 0], length(sorted_values), 1);  % Initialize all as green
+    colors(sorted_values < 0, :) = repmat([1, 0, 0], sum(sorted_values < 0), 1);  % Change to red for negative values
+    h.CData = colors;
+
+    % Remove the y-axis tick labels
+    set(gca, 'YTick', []);
+
+    % Add text labels at the end of the bars
+    for i = 1:length(sorted_values)
+        if sorted_values(i) >= 0
+            text(sorted_values(i) + 0.1, i, sorted_labels{i}, 'VerticalAlignment', 'middle');
+        else
+            text(sorted_values(i) - 0.1, i, sorted_labels{i}, 'VerticalAlignment', 'middle', 'HorizontalAlignment', 'right');
+        end
+    end
+
+    % Optionally, you can adjust other properties like axis labels, title, etc.
+    xlabel('t-score');
+    % title('Horizontal Bar Plot with Sorted Bars and Custom Labels');
+
+    subplot(1,3,2);
+    wordcloud(hitbl,'Word','t', 'Color', 'red', 'HighlightColor', 'red');
+    title("High Word Cloud");
+
+    subplot(1,3,3);
+    wordcloud(lotbl,'Word','t', 'Color', 'blue', 'HighlightColor', 'blue');
+    title("Low Word Cloud");
+
+    % Control Sizing
+    set(gcf, 'position', [10, 10, 2000, 1600]);
+    drawnow; snapnow;
+end