forked from MahmoudAbdellahi/Lively_Vectors
-
Notifications
You must be signed in to change notification settings - Fork 0
/
lv_SO_Spindle_FeatureExtraction.m
382 lines (318 loc) · 14.8 KB
/
lv_SO_Spindle_FeatureExtraction.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
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
%% Targeting targeted memory reactivation: characteristics of cued reactivation in sleep
% Mahmoud E. A. Abdellahi, Anne C. M. Koopman, Matthias S. Treder & Penelope A. Lewis
% This script is designed to compute temporal measures
% by Miguel Navarrete
% CUBRIC
% 2017
function varargout = lv_SO_Spindle_FeatureExtraction(correctness_trls, spindle)
%% Formatting data for the extraction of SO and spindle features
for i=1:length(correctness_trls)
fs=200;
new_header{i,1}.fsample = fs;
st = ceil(size(correctness_trls{i, 1}.trial,3)/2);
new_header{i,1}.label = correctness_trls{i, 1}.label;
new_header{i,1}.sampleinfo = [st size(correctness_trls{i, 1}.trial,1)*size(correctness_trls{i, 1}.trial,3) + st];
new_header{i,1}.scoring = { int32( ones( 1, size(correctness_trls{i, 1}.trial,1)*size(correctness_trls{i, 1}.trial,3) ) ) };
the1_2_stim = repmat([1;2], ceil( length(correctness_trls{i, 1}.sampleinfo(:,1))) ,1);
the1_2_stim = the1_2_stim(1:length(correctness_trls{i, 1}.sampleinfo(:,1)) *2);
idx = st: size(correctness_trls{i, 1}.trial,3): new_header{i,1}.sampleinfo(2)+ st;
dummy = idx ; % generating dummy idx for second click
idx = idx(1:sum(the1_2_stim==1));
dummy = dummy(1:sum(the1_2_stim==2));
V = [];
V(the1_2_stim==1) = idx;
V(the1_2_stim==2) = dummy + 301; % to cut the samples according to the
% trial length (1.5sec. in sleep and sampling rate is 200samples/sec., so 301samples)
new_header{i,1}.stim = [V' the1_2_stim ];
new_header{i,1}.time{1, 1} = 0 : 1/fs : (size(correctness_trls{i, 1}.trial,1)*size(correctness_trls{i, 1}.trial,3) )/200;
new_header{i,1}.time{1, 1} = new_header{i,1}.time{1, 1}(1:size(correctness_trls{i, 1}.trial,1)*size(correctness_trls{i, 1}.trial,3));
temp = permute(correctness_trls{i, 1}.trial,[2 3 1]);
temp = reshape(temp,length(correctness_trls{i, 1}.label),[] );
new_header{i,1}.trial{1, 1} = temp;
end
%% Setting parameters
sbj = 1:length(correctness_trls); % codes of participants
channel = find(ismember(lower(new_header{1, 1}.label),lower({'Fz'})));
nm_fsample = 200;
motor = find(ismember(lower(new_header{1, 1}.label),lower({'C6','CP4','C5','CP3'}))); % motor
vt_timeLim = [-3.04, 3.04]; % long trial duration inside sleep_raw
nm_SMOOTHWINDOW = 0.2; % 200ms window for rms power calculation
nm_TROUGHTHRESHOLD = 0;
nm_PEAKTHRESHOLD = 0; % considering all SOs, for now and filter them later inside SO_Spindle classification
nm_STIMDISTANCE = 2; % pre-cue duration for feature extraction
if spindle ==1, nm_STIMDISTANCE = 1.5; end % pre-cue duration for feature extraction
% detection thresholds
vt_timeSpindles = [0.3,3];
vt_timeTroughs = [0.166667,1.666667];
vt_timeTroughs = vt_timeTroughs + [-1,+1].*diff(vt_timeTroughs)*0.05; % +/- 5% time range
vt_stageSpindles = [3,4];
vt_stageSO = [2,3,4];
vt_freqSpindleSlow = [9,11];
vt_freqSpindleFast = [11,17];
nm_SPINDLENUMOSC = 5;
vt_timeEpoch = vt_timeLim(1):1/nm_fsample:vt_timeLim(2);
ob_filtEEGLo= fn_designDCfirfilter(nm_fsample);
ob_filtEEGHi= fn_designLowPassEEGfirfilter(nm_fsample);
ob_filtFS = fn_designSpindlefirfilter(nm_fsample,'custom',vt_freqSpindleFast);
ob_filtSS = fn_designSpindlefirfilter(nm_fsample,'custom',vt_freqSpindleSlow);
ob_filtSO = fn_designSOfirfilter(nm_fsample);
vt_hilFreq = (0:0.01:nm_fsample/2);
vt_idxAmp1 = vt_hilFreq < 3;
vt_idxAmp2 = vt_hilFreq >= 3 & vt_hilFreq <= 4;
vt_idxAmp3 = vt_hilFreq > 4;
vt_idxAmp1 = ones(sum(vt_idxAmp1),1);
vt_idxAmp2 = (1:sum(vt_idxAmp2)).*(-1/sum(vt_idxAmp2))+1;
vt_idxAmp3 = zeros(sum(vt_idxAmp3),1);
vt_hilAmpl = vertcat(vt_idxAmp1,vt_idxAmp2(:),vt_idxAmp3);
vt_hilFreq = vt_hilFreq * 2 /nm_fsample;
if mod(numel(vt_hilFreq),2)
vt_hilFreq = vt_hilFreq(1:end-1);
vt_hilAmpl = vt_hilAmpl(1:end-1);
end
nm_hilOrder = nm_fsample/2;
vt_filtImag = firls(nm_hilOrder,vt_hilFreq,vt_hilAmpl,'hilbert');
vt_filtReal = firls(nm_hilOrder,vt_hilFreq,ones(size(vt_hilAmpl)));
nm_hilDelay = nm_hilOrder/2;
nm_smoothSamples = round(nm_SMOOTHWINDOW * nm_fsample);
vt_dataLabels = {'markerTime','trialsEEG','trialsSO','trialsFS','trialsSS'...
'trialsFSrms','trialsSSrms','trialsSOphH','trialsSSphF'...
'features','detectionLogs'};
%% Feature exraction
for ff = 1 : numel(sbj)
fprintf('\n\n\n ---*** Processing participant %d ***--- \n\n', sbj(ff))
st_Header = new_header{ff,1};
vt_channels = st_Header.label;
st_Header.scoring{1, 1} = st_Header.scoring{1, 1}.*3; % all data from SWS
%% - Set signals
% Selcet & Obtain raw channels
if numel(st_Header.label) == 1
st_Header.label = st_Header.label{1};
end
mx_rawEEG = nan(size(st_Header.trial{1},2),numel(vt_channels));
for cc = 1:numel(vt_channels)
vt_idxCh = ismember(st_Header.label,vt_channels{cc});
if any(vt_idxCh)
vt_tempChannel = st_Header.trial{1}(vt_idxCh,:);
mx_rawEEG(:,cc) = vt_tempChannel(:);
else
warning('Channel %s not present ',...
vt_channels{cc})
mx_rawEEG(:,cc) = 0;
end
clear vt_tempChannel
end
% Obtain sampling frequency
nm_rawSampling = st_Header.fsample;
% Obtain markers
mx_markers = st_Header.stim;
vt_idxStim = mx_markers(:,2) == 17;
mx_markers(vt_idxStim,2) = 1;
vt_idxStim = mx_markers(:,2) == 18;
mx_markers(vt_idxStim,2) = 2;
vt_idxStim = mx_markers(:,2) == 1 | mx_markers(:,2) == 2;
mx_markers = mx_markers(vt_idxStim,:);
% Prepare markers by trials
mx_markers = fn_preparemarkers(mx_markers,[1 2]);
mx_markTime = mx_markers./nm_rawSampling;
% Obtain hypnogram
vt_hypnogram = st_Header.scoring{1};
% Clear memory
clear st_Header
%% - Resample Data
fprintf('\n Resampling data to %iHz: ',nm_fsample)
mx_rawEEG = resample(mx_rawEEG,nm_fsample,nm_rawSampling);
vt_rawTime = (0:numel(vt_hypnogram)-1)/nm_rawSampling;
vt_eegTime = (0:size(mx_rawEEG,1)-1)/nm_fsample;
vt_hypnogram = single(vt_hypnogram);
vt_hypnogram = interp1(vt_rawTime,vt_hypnogram,vt_eegTime,'nearest');
vt_hypnogram = int32(vt_hypnogram);
%% - Process channels
mx_dataClick_1 = cell(numel(vt_channels),numel(vt_dataLabels));
mx_dataClick_2 = cell(numel(vt_channels),numel(vt_dataLabels));
mx_spindleDet = cell(numel(vt_channels),2);
vt_soWaves = cell(numel(vt_channels),1);
vt_soAmpli = cell(numel(vt_channels),1);
vt_soThres = cell(numel(vt_channels),1);
vt_pWelch = cell(numel(vt_channels),1);
if spindle ==1, loopidx = motor; end
if spindle ==0, loopidx = channel; end
for cci = 1:length(loopidx ) % looping on channels
cc = loopidx(cci);
if sum(abs(mx_rawEEG(:,cc))) == 0
continue
end
fprintf('\n ** Filtering in EEG band | ch-%s: ',...
vt_channels{cc})
vt_signalEEG = fn_filterOffline(mx_rawEEG(:,cc),ob_filtEEGLo);
vt_signalEEG = fn_filterOffline(vt_signalEEG,ob_filtEEGHi);
[st_psd.power,st_psd.freq] = pwelch(vt_signalEEG,...
10*nm_fsample,2*nm_fsample,...
[],nm_fsample);
vt_pWelch{cc} = st_psd;
% Filter in the fast spindle band
vt_signalFS = fn_filterOffline(mx_rawEEG(:,cc),ob_filtFS);
vt_rmsFS = fn_rmstimeseries(vt_signalFS,nm_smoothSamples);
% Filter in the slow spindle band
vt_signalSS = fn_filterOffline(mx_rawEEG(:,cc),ob_filtSS);
vt_rmsSS = fn_rmstimeseries(vt_signalSS,nm_smoothSamples);
% Filter in SO frequency band
vt_signalSO = fn_filterOffline(mx_rawEEG(:,cc),ob_filtSO);
% Apply Hilbert filter
vt_Imag = filter(vt_filtImag,1,vt_signalSO);
vt_Real = filter(vt_filtReal,1,vt_signalSO);
vt_hilb = vt_Real + 1i*vt_Imag;
vt_hilb(1:end-nm_hilDelay) = vt_hilb(nm_hilDelay+1:end);
vt_hilb(end-nm_hilDelay:end)= nan;
vt_hilbFilter = vt_hilb;
% Obtain SO phase
vt_phaseSO = angle(hilbert(vt_signalSO));
%% Detect SO events
fprintf('\n ** Detect SO events: ')
st_Cnf = struct;
st_Cnf.freqband = [];
st_Cnf.fsampling = nm_fsample;
st_Cnf.threshold = nm_TROUGHTHRESHOLD;
st_Cnf.timebounds = vt_timeTroughs;
st_Cnf.hypnogram = vt_hypnogram;
st_Cnf.method = 'threshold';
st_Cnf.stage = vt_stageSO;
st_Cnf.minthresh = [];
st_Cnf.toFilter = [];
[vt_SOwaves,nm_thr] = fn_detectsleepSO(vt_signalSO,st_Cnf);
vt_soWaves{cc} = single(vt_SOwaves);
vt_soAmpli{cc} = single(vt_signalSO(vt_SOwaves));
vt_soThres{cc} = single(nm_thr);
clear st_Cnf
%% Detect spindle events
fprintf('\n ** Processing fast spindles: ')
st_Cnf = struct;
st_Cnf.fsampling = nm_fsample;
st_Cnf.minnumosc = nm_SPINDLENUMOSC;
st_Cnf.timebounds = vt_timeSpindles;
st_Cnf.hypnogram = vt_hypnogram;
st_Cnf.stage = vt_stageSpindles;
st_Cnf.rms = vt_rmsFS;
st_Cnf.rawEEG = vt_signalEEG;
st_Cnf.freqband = vt_freqSpindleFast;
st_Cnf.method = 'fixed';
mx_spindleFast = fn_detectsleepSpindles(vt_signalFS,st_Cnf);
clear st_Cnf
fprintf('\n ** Processing slow spindles: ')
st_Cnf = struct;
st_Cnf.fsampling = nm_fsample;
st_Cnf.minnumosc = nm_SPINDLENUMOSC;
st_Cnf.timebounds = vt_timeSpindles;
st_Cnf.hypnogram = vt_hypnogram;
st_Cnf.stage = vt_stageSpindles;
st_Cnf.rms = vt_rmsSS;
st_Cnf.rawEEG = vt_signalEEG;
st_Cnf.freqband = vt_freqSpindleSlow;
st_Cnf.window = nm_SMOOTHWINDOW;
st_Cnf.method = 'fixed';
mx_spindleSlow = fn_detectsleepSpindles(vt_signalSS,st_Cnf);
clear st_Cnf
mx_spindleDet{cc,1} = mx_spindleFast;
mx_spindleDet{cc,2} = mx_spindleSlow;
%% extracting spindle feature
if spindle ==1
trl_mx = new_header{ff,1}.stim; trl_lim_s = nm_fsample.* [-3.04 3.04]; detected_Fspdl=[]; trl_mx2=[];
trl_mx = trl_mx(trl_mx(:,2)==1 , 1);
for i=1:length(trl_mx), trl_mx2(i,:) = trl_mx(i) + [trl_lim_s(1):trl_lim_s(2)]; end
trl_mx2 = trl_mx2'; %timextrls
trl_time = -3.04:1/nm_fsample:3.04;
for i=1:size(mx_spindleFast,1)
detected_Fspdl = [detected_Fspdl mx_spindleFast(i,1):mx_spindleFast(i,2)];
end
trl_V = trl_mx2(:); trl_V = trl_V'; spdl_pos = zeros(1,length(trl_V));
spdl_pos(detected_Fspdl)=1;
spdl_pos = reshape(spdl_pos, [size(trl_mx2,1),size(trl_mx2,2)]);
spdl_pos = spdl_pos';
% extracting spindle presence
ft_limits = nearest(trl_time,-nm_STIMDISTANCE):nearest(trl_time,0); % extracting the spindle features in that range
masked_spindles = spdl_pos;
masked_spindles = masked_spindles(:,ft_limits);
for i=1:size(masked_spindles,1)
spindle_features{cci , ff}(i,:) = double(any(masked_spindles(i,:))); % has_spindle, yes: 1, no: 0
end
continue;
end
%% Extract time trials
for nm_markType = 1:2
fprintf('\n > Extract time trials for %i-click: ',nm_markType)
st_Cnf = struct;
st_Cnf.markerTime = mx_markTime(:,nm_markType);
st_Cnf.fsampling = nm_fsample;
st_Cnf.window = vt_timeLim;
st_Cnf.timetrial = vt_timeEpoch(1:2:end);
vt_timeTrialsEEG = fn_extracttimetrials(vt_signalEEG,st_Cnf);
vt_timeTrialsSO = fn_extracttimetrials(vt_signalSO,st_Cnf);
vt_timeTrialsFS = fn_extracttimetrials(vt_signalFS,st_Cnf);
vt_timeTrialsSS = fn_extracttimetrials(vt_signalSS,st_Cnf);
vt_timeTrialsFSrms = fn_extracttimetrials(vt_rmsFS,st_Cnf);
vt_timeTrialsSSrms = fn_extracttimetrials(vt_rmsSS,st_Cnf);
switch nm_markType
case 1
mx_dataClick_1{cc,1} = mx_markTime(:,nm_markType);
mx_dataClick_1{cc,2} = vt_timeTrialsEEG;
mx_dataClick_1{cc,3} = vt_timeTrialsSO;
mx_dataClick_1{cc,4} = vt_timeTrialsFS;
mx_dataClick_1{cc,5} = vt_timeTrialsSS;
mx_dataClick_1{cc,6} = vt_timeTrialsFSrms;
mx_dataClick_1{cc,7} = vt_timeTrialsSSrms;
case 2
mx_dataClick_2{cc,1} = mx_markTime(:,nm_markType);
mx_dataClick_2{cc,2} = vt_timeTrialsEEG;
mx_dataClick_2{cc,3} = vt_timeTrialsSO;
mx_dataClick_2{cc,4} = vt_timeTrialsFS;
mx_dataClick_2{cc,5} = vt_timeTrialsSS;
mx_dataClick_2{cc,6} = vt_timeTrialsFSrms;
mx_dataClick_2{cc,7} = vt_timeTrialsSSrms;
end
end
vt_timeTrial = vt_timeEpoch(1:2:end);
%% Compute event features
for nm_markType = 1:2
fprintf('\n > Compute event features for %i-click: ',nm_markType)
st_Cnf = struct;
st_Cnf.SOlocations = vt_SOwaves;
st_Cnf.markerTime = mx_markTime;
st_Cnf.currentMark = nm_markType;
st_Cnf.threshold = nm_PEAKTHRESHOLD;
st_Cnf.stimDistance = round(nm_STIMDISTANCE * nm_fsample);
st_Cnf.peakDistance = round(nm_STIMDISTANCE * nm_fsample);
st_Cnf.fSampling = nm_fsample;
st_Cnf.FSlocations = mx_spindleFast;
st_Cnf.SSlocations = mx_spindleSlow;
st_Cnf.FSFilter = vt_signalFS;
st_Cnf.SSFilter = vt_signalSS;
st_Cnf.FSrms = vt_rmsFS;
st_Cnf.SSrms = vt_rmsSS;
st_Cnf.hilbDelay = nm_hilDelay;
st_Cnf.hilbFilter = vt_hilbFilter;
st_Cnf.hilbPhase = vt_phaseSO;
st_Cnf.rawEEG = vt_signalEEG;
st_Cnf.spindFreq = vt_freqSpindleFast;
st_Cnf.timeSpind = [0.25,1.75];
[st_Features,st_Logs] = fn_obtaintrialfeatures(...
vt_signalSO,st_Cnf);
clear st_Cnf
switch nm_markType
case 1
mx_dataClick_1{cc,10} = st_Features;
mx_dataClick_1{cc,11} = st_Logs;
case 2
mx_dataClick_2{cc,10} = st_Features;
mx_dataClick_2{cc,11} = st_Logs;
end
end
end
if spindle ==0
SO_features{ff,1} = mx_dataClick_1;
end
end
if spindle ==1
varargout{1} = spindle_features;
else
varargout{1} = SO_features;
end
end