20240706 from cogitate-msp1/fmri@master

about.md

@@ -0,0 +1,18 @@
+# About
+
+General information regarding the fMRI analysis pipeline. A comprehensive description of the pipeline and running instructions can be found in `readme.md`
+
+## Information
+
+| | |
+| --- | --- |
+author_name | *Yamil Vidal<sup>1</sup>, David Richter<sup>1</sup>, Aya Khalaf<sup>2</sup>*
+author_email | *<[email protected]>, <[email protected]>, <[email protected]>*
+PI_name | *Floris de Lange<sup>1</sup>, Hal Blumenfeld<sup>2</sup>*
+affiliation | *<sup>1</sup>Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, 6500 HB Nijmegen, The Netherlands. <sup>2</sup>Department of Neurology, Yale University School of Medicine, New Haven, CT*
+PI_email | *<[email protected]>, <[email protected]>*
+programming_language | *Bash, Python, MATLAB*
+Is a readme file included with detailed instructions for running the code? | *Yes. readme.md*
+Is the environment file provided? | *Yes*
+Is there a config file provided to change runtime parameters? | *No*
+Does the code run on the sample dataset? | *Yes*

decoding/MSP1_roi_decoding_plots.py

@@ -0,0 +1,103 @@
+"""
+Plots ROI accuracy results on a brain surface
+
+Author: Aya Khalaf
+Email: [email protected]
+Date created: 03-04-2023
+"""
+
+import numpy as np
+import matplotlib.pyplot as plt
+from plotters import plot_time_series, plot_matrix, plot_rasters, plot_brain
+import config
+import os
+import pandas as pd
+# get the parameters dictionary
+param = config.param
+
+# =================================================================================
+# Select whether to apply plotting to category or orientation decoding problems - options 'category' and 'orientation'
+decoding_problem = 'category'
+#decoding_problem = 'orientation'
+#Select within condition decoding or generalization across conditions - options 'within_condition' and 'generalization'
+# For orientation decoding, approach should be set to 'within_condition'
+approach = 'generalization'
+approach = 'within_condition'
+# if approach = 'within_condition', the line below selects  relevant or irrelevant condition - options 'relevant' and 'irrelevant'
+# if approach = 'generalization', the line below selects generalization direction  - options 'relevant-irrelevant' and 'irrelevant-relevant'
+condition = 'relevant'
+condition = 'relevant-irrelevant'
+classifier = 'SVM'
+# Stimulus categories - options include 'face', 'object', 'letter' , and 'falseFont'
+stimulus_categories = ['face', 'object']
+stimulus_categories = ['letter', 'falseFont']
+#stimulus_categories = ['letter', 'falsefont']
+# Number of voxels per ROI (ROI size)
+n_voxels = 300
+# Chance level (50% for category decoding (binary) and 33.33% for orientation decoding (3-class) )
+chance_level = 0.5
+csv_dir = os.path.join('/mnt/beegfs/XNAT/COGITATE/fMRI/phase_2/processed/bids/derivatives/decoding/nibetaseries/roi_decoding/',approach, classifier, decoding_problem, stimulus_categories[0] + '_' + stimulus_categories[1] + '_' + str(n_voxels))
+if decoding_problem == 'orientation':
+    condition = 'relevant+irrelevant'
+    chance_level = 0.33
+    stimulus_category = 'face'
+    #stimulus_category = 'object'
+    #stimulus_category = 'letter'
+    #stimulus_category = 'falseFont'
+    csv_dir = os.path.join('/mnt/beegfs/XNAT/COGITATE/fMRI/phase_2/processed/bids/derivatives/decoding/nibetaseries/roi_decoding/', approach, classifier, decoding_problem, stimulus_category + '_' + str(n_voxels))
+
+if decoding_problem == 'category':
+            vmin=0.5
+            vmax=0.85
+            if ((condition=='relevant') | (condition=='irrelevant-relevant')):
+                cmaps = 'Oranges'
+                cmap_start = 0.1
+                cmap_end = 1
+            elif ((condition == 'irrelevant') | (condition == 'relevant-irrelevant')):
+                cmaps = 'Purples'
+                cmap_start = 0.5
+                cmap_end = 1
+elif decoding_problem == 'orientation':
+    cmaps = 'Purples'
+    vmin=0.33
+    vmax=0.40
+    cmap_start = 0.5
+    cmap_end = 1
+
+# Plotting brain surface:
+acc_values = {'G_front_middle': 0.9, 'Pole_temporal': 0.7, 'G_occipital_middle': 0.6}
+#IIT_roi_list = ['GNW','IIT','IIT_excluded', 'IIT_extended', 'G_and_S_frontomargin', 'G_and_S_transv_frontopol', 'G_front_sup', 'G_rectus', 'G_subcallosal', 'S_orbital_lateral', 'S_orbital_med-olfact', 'S_orbital-H_Shaped', 'S_suborbital']
+GNW_roi_list = ['G_and_S_cingul-Ant', 'G_and_S_cingul-Mid-Ant', 'G_and_S_cingul-Mid-Post', 'G_front_inf-Opercular', 'G_front_inf-Orbital', 'G_front_inf-Triangul', 'G_front_middle', 'Lat_Fis-ant-Horizont', 'Lat_Fis-ant-Vertical', 'S_front_inf', 'S_front_middle', 'S_front_sup']
+# IIT Basic ROI list
+IIT_roi_list_1 = ['G_temporal_inf', 'Pole_temporal', 'G_cuneus', 'G_occipital_sup', 'G_oc-temp_med-Lingual', 'Pole_occipital', 'S_calcarine', 'G_and_S_occipital_inf', 'G_occipital_middle', 'G_oc-temp_lat-fusifor', 'G_oc-temp_med-Parahip', 'S_intrapariet_and_P_trans', 'S_oc_middle_and_Lunatus', 'S_oc_sup_and_transversal', 'S_temporal_sup']
+roi_list = GNW_roi_list + IIT_roi_list_1
+
+csv_file = os.path.join(csv_dir, 'accuracy_stats_' + condition + '.csv')
+save_file= os.path.join(csv_dir, condition + '.eps')
+data_df = pd.read_csv(csv_file)
+rois = (data_df['ROI']).tolist()
+accuracies= (data_df['Average Accuracy']).array
+Significance = (data_df['Significance']).array
+#Significance[10]=0
+#Significance[36]=0
+#G_orbital
+#Significance[18]=0
+#Significance[16]=0
+rois_dict = {}
+k=0
+for roi in rois:
+    if roi in roi_list:
+        if Significance[k]:
+            rois_dict[roi] =accuracies[k]
+    k=k+1
+
+
+plot_brain( roi_map=rois_dict, subject='fsaverage', surface='inflated', hemi='lh', sulc_map='curv', parc='aparc.a2009s',
+                         views=['lateral',(90, -30, 0)],cmap_start=cmap_start, cmap_end=cmap_end,
+                         cmap=cmaps, colorbar=True, colorbar_title='ACC', vmin=vmin, vmax=vmax, outline_overlay=True, overlay_method='overlay',
+                         brain_cmap='Greys', brain_alpha=1, save_file=os.path.join(csv_dir, condition + '_lh_outline_30deg.png'))
+
+#roi_map_edge_color = [0, 0, 0] to add borders around rois
+
+
+

decoding/MSP1_searchlight_decoding_plots.py

@@ -0,0 +1,97 @@
+"""
+Plots searchlight accuracy maps on a brain surface
+
+Author: Aya Khalaf
+Email: [email protected]
+Date created: 03-04-2023
+"""
+
+from plotters import plot_brain
+import matplotlib
+import matplotlib.pyplot as plt
+import os
+import config
+
+# Select whether to apply plotting to category or orientation decoding problems - options 'category' and 'orientation'
+decoding_problem = 'category'
+decoding_problem = 'orientation'
+# Select within condition decoding or generalization across conditions - options 'within_condition' and 'generalization'
+# For orientation decoding, approach should be set to 'within_condition'
+approach = 'generalization'
+approach = 'within_condition'
+# if approach = 'within_condition', the line below selects  relevant or irrelevant condition - options 'relevant' and 'irrelevant'
+# if approach = 'generalization', the line below selects generalization direction  - options 'relevant-irrelevant' and 'irrelevant-relevant'
+# For orientation decoding, no need to change 'condition' as it will be specified at line 40.
+condition = 'irrelevant-relevant'
+#condition ='relevant'
+
+classifier = 'SVM'
+
+# Stimulus categories  - options include 'face', 'object', 'letter' , and 'falseFont'
+# For orientation decoding, do not change 'stimulus_categories' here and change line 42 instead.
+stimulus_categories = ['face', 'object']
+#stimulus_categories = ['letter', 'falseFont']
+#stimulus_categories = ['face', 'baseline']
+#stimulus_categories = ['object', 'baseline']
+#stimulus_categories = ['letter', 'baseline']
+#stimulus_categories = ['falseFont', 'baseline']
+#stimulus_categories = ['falseFonts', 'baseline']
+# Radius of searchlight sphere
+searchlight_radius = 4
+# Chance level (50% for category decoding (binary) and 33.33% for orientation decoding (3-class) )
+chance_level = 0.5
+
+data_dir = os.path.join('/mnt/beegfs/XNAT/COGITATE/fMRI/phase_2/processed/bids/derivatives/decoding/nibetaseries/searchlight_decoding',approach, condition, classifier, str(searchlight_radius) + 'mm', 'category/' + stimulus_categories[0] + '_' + stimulus_categories[1])
+
+if decoding_problem == 'orientation':
+    condition = 'relevant+irrelevant'
+    chance_level = 0.33
+    stimulus_category = 'face'
+    #stimulus_category = 'object'
+    #stimulus_category = 'letter'
+    #stimulus_category = 'falseFont'
+    data_dir = os.path.join( '/mnt/beegfs/XNAT/COGITATE/fMRI/phase_2/processed/bids/derivatives/decoding/nibetaseries/searchlight_decoding',approach, condition, classifier, str(searchlight_radius) + 'mm', 'Orientation/' + stimulus_category)
+
+
+if decoding_problem == 'category':
+    if ((condition == 'relevant') | (condition == 'irrelevant-relevant')):
+        cmaps = 'Oranges'
+        cmap_start = 0.1
+        cmap_end = 1
+    elif ((condition == 'irrelevant') | (condition == 'relevant-irrelevant')):
+        cmaps = 'Purples'
+        cmap_start = 0.5
+        cmap_end = 1
+    #plot_brain( surface='inflated', cmap=cmaps, cmap_start=0.5, cmap_end=1, overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')], overlay_threshold=0, vmin=chance_level, vmax=0.75, outline_overlay=True, save_file=os.path.join(data_dir, condition + '_' + 'lh_outline.png'))
+    ## for the right side
+    #plot_brain( surface='inflated', cmap=cmaps, cmap_start=0.5, cmap_end=1, hemi='rh', overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')], overlay_threshold=0, vmin=chance_level, vmax=0.75, outline_overlay=True, views=['medial', 'lateral'], save_file=os.path.join(data_dir, condition + '_' + 'rh_outline.png'))
+    plot_brain( surface='inflated', cmap=cmaps,  overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')], overlay_threshold=0, vmin=chance_level, vmax=0.75, cmap_start=cmap_start, cmap_end=cmap_end, outline_overlay=True, views=['lateral',(90, -30, 0)], save_file=os.path.join(data_dir, condition + '_' + 'lh_outline_30deg.png'))
+    ## for the right side
+    plot_brain( surface='inflated', cmap=cmaps,  hemi='rh', overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')], overlay_threshold=0, vmin=chance_level, vmax=0.75, cmap_start=cmap_start, cmap_end=cmap_end, outline_overlay=True, views=[(-90, 30, 0), 'lateral'], save_file=os.path.join(data_dir, condition + '_' + 'rh_outline_30deg.png'))
+
+elif decoding_problem == 'orientation':
+    param = config.param
+    colors = [(1, 1, 1), param['colors']['face']]
+    cmaps = matplotlib.colors.LinearSegmentedColormap.from_list('Custom', colors, N=128)
+
+    """
+    plot_brain(surface='inflated', cmap=cmaps, cmap_start=0.5, cmap_end=1,
+               overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')],
+               overlay_threshold=0, vmin=chance_level, vmax=0.45, outline_overlay=True, views=['lateral', (180, 20, -10)],
+               save_file=os.path.join(data_dir, 'orientation_lh_inf_outline_posterior.png'))
+    # for the right side
+    plot_brain(surface='inflated', cmap=cmaps,  cmap_start=0.5, cmap_end=1, hemi='rh',
+               overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')],
+               overlay_threshold=0, vmin=chance_level, vmax=0.45, outline_overlay=True, views=[(180, 20, -10), 'lateral'],
+               save_file=os.path.join(data_dir, 'orientation_rh_inf_outline_posterior.png'))
+"""
+plot_brain(surface='inflated', cmap=cmaps, cmap_start=0.5, cmap_end=1,
+           overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')],
+           overlay_threshold=0, vmin=chance_level, vmax=0.45, outline_overlay=True, views=['lateral', 'medial'],
+           save_file=os.path.join(data_dir, 'orientation_lh_inf_outline.png'))
+# for the right side
+plot_brain(surface='inflated', cmap=cmaps, cmap_start=0.5, cmap_end=1, hemi='rh',
+           overlays=[os.path.join(data_dir, 'searchlight_group_accuracy_map_nonparametric.nii')],
+           overlay_threshold=0, vmin=chance_level, vmax=0.45, outline_overlay=True, views=['medial', 'lateral'],
+           save_file=os.path.join(data_dir, 'orientation_rh_inf_outline.png'))
+

decoding/config.py

@@ -0,0 +1,63 @@
+param = {
+  "font": "helvetica",
+  "font_size": 22,
+  "figure_size_mm": [183, 108],
+  "fig_res_dpi": 300,
+  "colors": {
+    "iit": [
+      0.00392156862745098,
+      0.45098039215686275,
+      0.6980392156862745],
+    "gnw": [
+      0.00784313725490196,
+      0.6196078431372549,
+      0.45098039215686275
+    ],
+    "task relevant": [
+      0.8352941176470589,
+      0.3686274509803922,
+      0.0
+    ],
+    "Irrelevant": [
+      0.5450980392156862,
+      0.16862745098039217,
+      0.8862745098039215
+    ],
+    "face": [
+      0/255,
+      53/255,
+      68/255
+    ],
+    "object": [
+      173/255,
+      80/255,
+      29/255
+    ],
+    "letter": [
+      57/255,
+      115/255,
+      132/255
+    ],
+    "false": [
+      97/255,
+      15/255,
+      0/255
+    ],
+    "500ms": [
+      1.0,
+      0.48627450980392156,
+      0.0
+    ],
+    "1000ms": [
+      0.6235294117647059,
+      0.2823529411764706,
+      0.0
+    ],
+    "1500ms": [
+      1.0,
+      0.7686274509803922,
+      0.0
+    ],
+    "cmap": "RdYlBu_r"
+  }
+}