-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrepresentation.m
265 lines (247 loc) · 11.3 KB
/
representation.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
channel = readtable(fullfile("config", "eeg_channel_labels_64.csv"), TextType="string");
subjs_id = unique(readtable(fullfile("data", "group_task-wordencoding_events.csv")).subj_id, "stable");
load(fullfile("data", "grp_subjs206_nodemean_1000ms.mat"), "grp_data");
[~, len_time_point, len_trial, len_subj] = size(grp_data);
regions_id = 1:6;
% we keep the lower triangular correlation (no diagonal)
idx_keep_cors = tril(true(len_subj, len_subj), -1);
% debug
% regions_id = 1:2;
% len_trial = 2;
%% intersubject similarity
% acquire: trial-level
simi_inter_by_trial = ...
utils.preallocate(regions_id, 1:len_trial, ...
VariableNames=["region_id", "trial_id"], ...
ScalarFisherZ=false);
fprintf("Processing trial level intersubject similarity...\n")
for i_region = regions_id
chan_in_reg = channel.code(channel.("region" + string(i_region)) ~= 0);
fprintf("Region " + string(i_region) + "\n")
for i_trial = progress(1:len_trial)
% collapse channel and time (thus spatiotemporal pattern)
cur_cor_mat = corr(reshape(grp_data(chan_in_reg, :, i_trial, :), ...
length(chan_in_reg) * len_time_point, []));
simi_inter_by_trial.fisher_z{ ...
simi_inter_by_trial.region_id == i_region & ...
simi_inter_by_trial.trial_id == i_trial} = ...
atanh(cur_cor_mat(idx_keep_cors));
end
end
parquetwrite( ...
fullfile("data", "type-inter_acq-trial_rs.parquet"), ...
simi_inter_by_trial)
clearvars simi_inter_by_trial
% acquire: whole-time-series
simi_inter_by_whole = ...
utils.preallocate(regions_id, ...
VariableNames="region_id", ...
ScalarFisherZ=false);
fprintf("Processing whole time series intersubject similarity...\n")
for i_region = progress(regions_id)
chan_in_reg = channel.code(channel.("region" + string(i_region)) ~= 0);
cur_cor_mat = corr(reshape(grp_data(chan_in_reg, :, :, :), ...
length(chan_in_reg) * len_time_point * len_trial, []), ...
rows="pairwise");
simi_inter_by_whole.fisher_z{ ...
simi_inter_by_whole.region_id == i_region} = ...
atanh(cur_cor_mat(idx_keep_cors));
end
parquetwrite( ...
fullfile("data", "type-inter_acq-whole_rs.parquet"), ...
simi_inter_by_whole)
clearvars simi_inter_by_whole
%% individual to group similarity
% acquire: trial-level
simi_grp_by_trial = ...
utils.preallocate(regions_id, 1:len_trial, subjs_id, ...
VariableNames=["region_id", "trial_id", "subj_id"]);
fprintf("Processing trial level individual to group similarity...\n")
for i_region = regions_id
chan_in_reg = channel.code(channel.("region" + string(i_region)) ~= 0);
fprintf("Region " + string(i_region) + "\n")
for i_trial = progress(1:len_trial)
% collapse channel and time (thus spatiotemporal pattern)
cur_dat = reshape(grp_data(chan_in_reg, :, i_trial, :), ...
length(chan_in_reg) * len_time_point, []);
simi_grp_by_trial.fisher_z(...
simi_grp_by_trial.region_id == i_region & ...
simi_grp_by_trial.trial_id == i_trial) = ...
utils.calc_simi_ind_to_grp(cur_dat, FisherZ=true);
end
end
parquetwrite( ...
fullfile("data", "type-group_acq-trial_rs.parquet"), ...
simi_grp_by_trial)
clearvars simi_grp_by_trial
% acquire: whole-time-series
simi_grp_by_whole = utils.preallocate(regions_id, subjs_id, ...
VariableNames=["region_id", "subj_id"]);
fprintf("Processing whole time series individual to group similarity...\n")
for i_region = progress(regions_id)
chan_in_reg = channel.code(channel.("region" + string(i_region)) ~= 0);
cur_dat = reshape(grp_data(chan_in_reg, :, :, :), ...
length(chan_in_reg) * len_time_point * len_trial, []);
simi_grp_by_whole.fisher_z(...
simi_grp_by_whole.region_id == i_region) = ...
utils.calc_simi_ind_to_grp(cur_dat, FisherZ=true);
end
parquetwrite( ...
fullfile("data", "type-group_acq-whole_rs.parquet"), ...
simi_grp_by_whole)
clearvars simi_grp_by_whole
%% windowed results (separate time window)
% setup for windowed calculations
size_window = 26;
step = 5;
[window_start, window_end] = utils.setup_window(len_time_point, size_window, step);
fprintf("Processing stepped window similarity...\n")
for i_region = regions_id
cur_reg = "region" + string(i_region);
chan_in_reg = channel.code(channel.(cur_reg) ~= 0);
fprintf(cur_reg + "\n")
for i_win = progress(1:length(window_start))
keep_time_points = window_start(i_win):window_end(i_win);
cur_win_dat = grp_data(chan_in_reg, keep_time_points, :, :);
fisher_z_inter = cell(len_trial, 1);
fisher_z_group = cell(len_trial, 1);
parfor i_trial = 1:len_trial
% collapse channel and time (thus spatiotemporal pattern)
cur_dat = reshape(cur_win_dat(:, :, i_trial, :), ...
length(chan_in_reg) * size_window, []); %#ok<*PFBNS>
cur_cor_inter = corr(reshape(cur_win_dat(:, :, i_trial, :), ...
length(chan_in_reg) * size_window, []));
fisher_z_inter{i_trial} = atanh(cur_cor_inter(idx_keep_cors));
fisher_z_group{i_trial} = ...
utils.calc_simi_ind_to_grp(cur_dat, FisherZ=true);
end
cur_simi_inter_by_window = ...
utils.preallocate(i_region, 1:len_trial, i_win, ...
VariableNames=["region_id", "trial_id", "window_id"], ...
ScalarFisherZ=false);
cur_simi_inter_by_window.fisher_z = fisher_z_inter;
path_inter = fullfile("data", "type-inter_acq-window", ...
"region-" + cur_reg, "window-" + string(i_win));
if (~exist(path_inter, "dir")), mkdir(path_inter), end
parquetwrite(fullfile(path_inter, "rs.parquet"), cur_simi_inter_by_window)
cur_simi_grp_by_window = ...
utils.preallocate(i_region, 1:len_trial, i_win, subjs_id, ...
VariableNames=["region_id", "trial_id", "window_id", "subj_id"]);
cur_simi_grp_by_window.fisher_z = vertcat(fisher_z_group{:});
path_group = fullfile("data", "type-group_acq-window", ...
"region-" + cur_reg, "window-" + string(i_win));
if (~exist(path_group, "dir")), mkdir(path_group), end
parquetwrite(fullfile(path_group, "rs.parquet"), cur_simi_grp_by_window)
end
end
clearvars cur_simi_inter_by_window cur_simi_grp_by_window
%% windowed results (separate regions, deprecated due to memory consuming)
% setup for windowed calculations
size_window = 26;
step = 5;
[window_start, window_end] = utils.setup_window(len_time_point, size_window, step);
% type: intersubject similarity
fprintf("Processing stepped window intersubject similarity...\n")
for i_region = regions_id
cur_simi_inter_by_window = ...
utils.preallocate(i_region, 1:len_trial, 1:length(window_start), ...
VariableNames=["region_id", "trial_id", "window_id"], ...
ScalarFisherZ=false);
cur_reg = "region" + string(i_region);
chan_in_reg = channel.code(channel.(cur_reg) ~= 0);
fprintf(cur_reg + "\n")
for i_trial = progress(1:len_trial)
cur_trial_dat = grp_data(chan_in_reg, :, i_trial, :);
fisher_z = cell(length(window_start), 1);
parfor i_win = 1:length(window_start)
% collapse channel and time (thus spatiotemporal pattern)
cur_cor_mat = corr(reshape( ...
cur_trial_dat(:, window_start(i_win):window_end(i_win), :, :), ...
length(chan_in_reg) * size_window, [])); %#ok<*PFBNS>
fisher_z{i_win} = atanh(cur_cor_mat(idx_keep_cors));
end
cur_simi_inter_by_window.fisher_z( ...
cur_simi_inter_by_window.trial_id == i_trial) = ...
fisher_z;
end
parquetwrite( ...
fullfile("data", "deprecated", ...
"type-inter_acq-window_region-" + cur_reg + "_rs.parquet"), ...
cur_simi_inter_by_window)
end
clearvars cur_simi_inter_by_window
% type: individual to group similarity
fprintf("Processing stepped window intersubject similarity...\n")
for i_region = regions_id
cur_simi_grp_by_window = ...
utils.preallocate(i_region, 1:len_trial, 1:length(window_start), subjs_id, ...
VariableNames=["region_id", "trial_id", "window_id", "subj_id"]);
cur_reg = "region" + string(i_region);
chan_in_reg = channel.code(channel.(cur_reg) ~= 0);
fprintf(cur_reg + "\n")
for i_trial = progress(1:len_trial)
cur_trial_dat = grp_data(chan_in_reg, :, i_trial, :);
fisher_z = cell(length(window_start), 1);
parfor i_win = 1:length(window_start)
% collapse channel and time (thus spatiotemporal pattern)
cur_dat = reshape( ...
cur_trial_dat(:, window_start(i_win):window_end(i_win), :, :), ...
length(chan_in_reg) * size_window, []);
fisher_z{i_win} = utils.calc_simi_ind_to_grp(cur_dat, FisherZ=true);
end
cur_simi_grp_by_window.fisher_z(...
cur_simi_grp_by_window.trial_id == i_trial) = ...
vertcat(fisher_z{:});
end
parquetwrite( ...
fullfile("data", "dep recated", ...
"type-group_acq-window_region-" + cur_reg + "_rs.parquet"), ...
cur_simi_grp_by_window)
end
clearvars cur_simi_grp_by_window
%% windowed results (MATLAB native, deprecated, hard for further analysis)
% setup for windowed calculations
size_window = 26;
step = 5;
[window_start, window_end] = utils.setup_window(len_time_point, size_window, step);
% type: intersubject similarity
fprintf("Processing stepped window intersubject similarity...\n")
simi_inter_by_window = nan(length(regions_id), len_trial, length(window_start), ...
sum(idx_keep_cors, "all"));
for i_region = regions_id
chan_in_reg = channel.code(channel.("region" + string(i_region)) ~= 0);
fprintf("Region " + string(i_region) + "\n")
for i_trial = progress(1:len_trial)
for i_win = 1:length(window_start)
% collapse channel and time (thus spatiotemporal pattern)
cur_cor_mat = corr(reshape( ...
grp_data(chan_in_reg, window_start(i_win):window_end(i_win), i_trial, :), ...
length(chan_in_reg) * size_window, []));
simi_inter_by_window(i_region, i_trial, i_win, :) = ...
atanh(cur_cor_mat(idx_keep_cors));
end
end
end
save(fullfile("data", "deprecated", "type-inter_acq-window_rs.mat"), ...
"simi_inter_by_window", "-v7.3")
clearvars simi_inter_by_window
% type: individual to group similarity
simi_grp_by_window = nan(length(regions_id), len_trial, length(window_start), len_subj);
fprintf("Processing stepped window intersubject similarity...\n")
for i_region = regions_id
chan_in_reg = channel.code(channel.("region" + string(i_region)) ~= 0);
fprintf("Region " + string(i_region) + "\n")
for i_trial = progress(1:len_trial)
for i_win = 1:length(window_start)
% collapse channel and time (thus spatiotemporal pattern)
cur_dat = reshape( ...
grp_data(chan_in_reg, window_start(i_win):window_end(i_win), i_trial, :), ...
length(chan_in_reg) * size_window, []);
simi_grp_by_window(i_region, i_trial, i_win, :) = ...
utils.calc_simi_ind_to_grp(cur_dat, FisherZ=true);
end
end
end
save(fullfile("data", "deprecated", "type-group_acq-window_rs.mat"), ...
"simi_grp_by_window", "-v7.3")
clearvars simi_grp_by_window