-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathMLi2_Find_Spindles.m
317 lines (268 loc) · 12.9 KB
/
MLi2_Find_Spindles.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
function [Spindle]=MLi2_Find_Spindles(lowerthresh_std, upperthresh_std,lowlimit_fq,highlimit_fq, eeg_chan, plot_it)
%8/2018 Crown
%used for /2018 MJFF update as LRRK2MLi2_Find_Spindles (still can be found
%in "old names" folder" but now will be used for EEG conference and SfN
%under MLi2 name
%Goal is to auto-detect spindles, grab their times and be able to plot them
%along with power and spectrogram then decide are they spindles yes or no
%and pull out data about their properties to refine process and get some
%measures to compare between groups
%inputs/needed:
%Sleep Times
%re-referenced ECOG
%Spindle properties to use for detection
if nargin<6
plot_it=true;
% plot_it=false;
end
if nargin<5
eeg_chan=find_files('*ECoG_Left_Mid*'); %edit this input for choosing sides and location
if isempty(eeg_chan)
eeg_chan=find_files('*ECoG_Left_Post*');
if isempty(eeg_chan)
Spindle.aborted=1;
return
end
end
end
if nargin<4
highlimit_fq = 16;
end
if nargin<3
lowlimit_fq = 8; %This is in litterature and then later I can say actually i just want 8-14 Hz or so
end
if nargin<2
upperthresh_std=5;
end
if nargin<1
lowerthresh_std=2.8;
end
Spindle.aborted=0; %initialize that its a good session and change to one if you need to cut the session
% eeg_chan=load(f{1});
mindur=0.5;
maxdur=1.5;
%Some values for the detection later
spin_min_duration_sec = 0.5; %.4s %in samples, at 200Hz
%spin_max_duration = 4000; %2s %maybe do this post-hoc
spin_minimum_inter_interval_period_sec = []; %300ms
% Values taken from Characterization of Topographically
% Specific Sleep Spindles in Mice by Kim et al. 2014
spec_freqs = 2:.25:20; %parameters for the PSDs
frex_min=9; %for spindles to be kept
frex_max=16;
%get basic variables out of the way
direct=strsplit(pwd,'\');
mouseday=direct{end};
load('Session_Info.mat');
ff=find_files('*Reref.mat');
main_signal=eeg_chan{1};
LFP=load(eeg_chan{1});
if isempty(ff)
disp('Data not yet rereferenced')
Spindle.aborted=1;
return
end
load(ff{1})
%what side you will use as reref depends on what side you have
bits=strsplit(main_signal,'-');
location=bits{2};
if contains(location,'Left')==1
Ref=ReRef.Allright;
end
if contains(location,'Right')==1
Ref=ReRef.Allleft;
end
ECOG(:,1)=LFP.LFP_uV(:,1);
ECOG(:,2)=LFP.LFP_uV(:,2)-Ref(:,2);
assert(isequal(LFP.LFP_uV(:,1),Ref(:,1)),'Time Stamps not equal')
%% Filter/Power
% Kim et al. PNAS bandpass-filtered from 7 to 15
%first find power across the whole signal, then restrict analysis to just
%sleep periods (for mean standard deviation etc. Stephen says make this
%for pre sleep only
d = designfilt('bandpassiir','FilterOrder',10, ...
'HalfPowerFrequency1',lowlimit_fq,'HalfPowerFrequency2',highlimit_fq, ...
'SampleRate',LFP.fs_final,'DesignMethod','butter');
%%% filter type
% freqz(d,[],LFP.fs_final)
filt_sig = filtfilt(d,double(ECOG(:,2))); %filtered to spindle band
as=abs(hilbert(filt_sig));
% power=envelope_cowen(as.^2);
% envpower=convn(as.^2,hanning(round(LFP.fs_final*.1)),'same'); %% hanning might be increasing power- doesnt sum to 1 divide by sum of hanning
smoothed_power(:,1)=ECOG(:,1);
hanwin=hanning(round(LFP.fs_final*.2)); %raising this to .2 to see if it will make the envelope less up and down-y
smoothed_power(:,2)=convn(as.^2,hanwin/sum(hanwin),'same'); %% hanning might be increasing power- doesnt sum to 1 divide by sum of hanning
%here I divide by the sum of the window
% smoothed_power(:,2)=envelope_cowen(envpower); %possibly oversmoothing!!!!!
%now you have smoothed power as a time series with time in first col
%% Concatinate the Power for the sleep intevals of PRE sleep
SleepPeriods.Pre=load([mouseday '-Sleep_Times_Pre.mat']);
SleepPeriods.Post=load([mouseday '-Sleep_Times_Post.mat']);
SleepPower=zeros(1,2);
for isleep=1:Rows(SleepPeriods.Pre.Actual_Sleep_Times_Pre)
SleepIntervalIX= smoothed_power(:,1)> SleepPeriods.Pre.Actual_Sleep_Times_Pre(isleep,1) & smoothed_power(:,1)< SleepPeriods.Pre.Actual_Sleep_Times_Pre(isleep,2);
SleepPower=[SleepPower ; smoothed_power(SleepIntervalIX,:)];
end
% only if you want to include post in your calculation
% for isleep=1:Rows(SleepPeriods.Post.Actual_Sleep_Times_Post)
% SleepIntervalIX= smoothed_power(:,1)> SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,1) & smoothed_power(:,1)< SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,2);
% SleepPower=[SleepPower ; smoothed_power(SleepIntervalIX,:)];
% end
maxpow = prctile(SleepPower(:,2),99);
minpow = prctile(SleepPower(:,2),1);
trimmedpowerIX=SleepPower(:,2)>minpow & SleepPower(:,2)<maxpow;
powstd=std(SleepPower(trimmedpowerIX,2));
mean_power=mean(SleepPower(trimmedpowerIX,2));
%% Thresholding JP did - spindle has to be between 1 and 3 stds, This can be adjusted as you see fit
lower_spin_thresh = mean_power + powstd*lowerthresh_std; %these will be in standard deviations
upper_spin_thresh = mean_power + powstd*upperthresh_std;
count1=0;
for isleep=1:Rows(SleepPeriods.Pre.Actual_Sleep_Times_Pre)
%or post
SleepIX=ECOG(:,1)>SleepPeriods.Pre.Actual_Sleep_Times_Pre(isleep,1) & ECOG(:,1)<SleepPeriods.Pre.Actual_Sleep_Times_Pre(isleep,2);
% SleepIX=ECOG(:,1)>SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,1) & ECOG(:,1)<SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,2);
% figure
% plot(smoothed_power(SleepIX,1),smoothed_power(SleepIX,2))
% % refline(0,upper_spin_thresh)
% % refline(0,lower_spin_thresh)
% refline(0,lower_spin_thresh)
% refline(0,upper_spin_thresh)
[spin_above_times, ~] = find_intervals(smoothed_power(SleepIX,:), upper_spin_thresh, lower_spin_thresh, ...
spin_min_duration_sec, spin_minimum_inter_interval_period_sec);
if ~isempty(spin_above_times)
% sigmatimes{isleep}=spin_above_times;
for ispin= 1:Rows(spin_above_times)
spinintervalIX=ECOG(:,1)>spin_above_times(ispin,1) & ECOG(:,1)<spin_above_times(ispin,2);
putative_spindle=ECOG(spinintervalIX,:);
plotinterval=ECOG(:,1)> spin_above_times(ispin,1)-1 & ECOG(:,1)<spin_above_times(ispin,2)+1;
[pxx,frex]= pburg(double(ECOG(spinintervalIX,2)),40,spec_freqs,LFP.fs_final);
% [pks,locs]=findpeaks(pxx,frex); % I think locs are now in frequency
[m ix]=max(pxx); % I think locs are now in frequency
peakfrex= frex(ix);
if peakfrex>frex_max || peakfrex<frex_min %make sure peak it where you want it
continue
end
spinlength=spin_above_times(ispin,2)-spin_above_times(ispin,1);
if spinlength<mindur ||spinlength>maxdur
continue
end
if max(putative_spindle(:,2)) > 500 || min(putative_spindle(:,2)) < -500
continue
end
count1=count1+1;
if plot_it==true
figure
subplot(3,1,1)
yyaxis left
plot(ECOG(plotinterval,1),ECOG(plotinterval,2),'k')
hold on
plot(ECOG(plotinterval,1),filt_sig(plotinterval),'b')
vline(min(ECOG(spinintervalIX,1)))
vline(max(ECOG(spinintervalIX,1)))
axis tight
yyaxis right
plot(smoothed_power(plotinterval,1),smoothed_power(plotinterval,2),'r') % now 1= 1 second and this should be easy to estimate Hz from
refline(0,lower_spin_thresh)
refline(0,upper_spin_thresh)
ylabel('uV')
xlabel('Time(s)')
title(sprintf('Duration %2.2f secs %1.1f frex',spinlength,peakfrex))
subplot(3,1,2)
Spectrogram_spindle_LC(ECOG(plotinterval,2), LFP.fs_final, 3:0.2:20, 'wavelet');
subplot(3,1,3)
pburg(double(ECOG(spinintervalIX,2)),40,spec_freqs,LFP.fs_final);
end
%Meta data for PRE Sleep
Spindle.Pre(count1).raw_spin= putative_spindle;
Spindle.Pre(count1).length_spin_sec=spinlength;
Spindle.Pre(count1).filt_sig=filt_sig(spinintervalIX,1);
Spindle.Pre(count1).mean_sig_pow=mean(smoothed_power(spinintervalIX,2));
Spindle.Pre(count1).time=ECOG(spinintervalIX,1);
Spindle.Pre(count1).values=ECOG(spinintervalIX,2);
Spindle.Pre(count1).peakfrex=peakfrex;
Spindle.Pre(count1).DayType=Session_Info.RecordingType{1};
Spindle.Pre(count1).Mouse=mouseday(1:3);
Spindle.Pre(count1).Day=Session_Info.Recordingday;
ixx=Session_Info.drugs.Mouse==mouseday(1:3);
Spindle.Pre(count1).drugmouse=Session_Info.drugs.GotDrug(ixx);
Spindle.Pre(count1).MouseType=mouseday(1);
Spindle.Pre(count1).SleepSession='Pre';
Spindle.Pre(count1).psd=pxx;
Spindle.Pre(count1).psd_frex=frex;
Spindle.Pre(count1).location=bits{2};
end
end
end
%now for post sleep
%%
count2=0;
for isleep=1:Rows(SleepPeriods.Post.Actual_Sleep_Times_Post)
%or post
SleepIX=ECOG(:,1)>SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,1) & ECOG(:,1)<SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,2);
% SleepIX=ECOG(:,1)>SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,1) & ECOG(:,1)<SleepPeriods.Post.Actual_Sleep_Times_Post(isleep,2);
%
% figure
% plot(smoothed_power(SleepIX,1),smoothed_power(SleepIX,2))
% refline(0,upper_spin_thresh)
% refline(0,lower_spin_thresh)
[spin_above_times, ~] = find_intervals(smoothed_power(SleepIX,:), upper_spin_thresh, lower_spin_thresh, ...
spin_min_duration_sec, spin_minimum_inter_interval_period_sec);
if ~isempty(spin_above_times)
% sigmatimes{isleep}=spin_above_times;
for ispin= 1:Rows(spin_above_times)
spinintervalIX=ECOG(:,1)>spin_above_times(ispin,1) & ECOG(:,1)<spin_above_times(ispin,2);
putative_spindle=ECOG(spinintervalIX,:);
plotinterval=ECOG(:,1)> spin_above_times(ispin,1)-1 & ECOG(:,1)<spin_above_times(ispin,2)+1;
[pxx,frex]= pburg(double(ECOG(spinintervalIX,2)),40,spec_freqs,LFP.fs_final);
[m, ix]=max(pxx); % I think locs are now in frequency
peakfrex= frex(ix);
if peakfrex>frex_max || peakfrex<frex_min %make sure peak it where you want it
continue
end
spinlength=spin_above_times(ispin,2)-spin_above_times(ispin,1);
if spinlength<mindur ||spinlength>maxdur
continue
end
count2=count2+1;
if plot_it==true
figure
subplot(3,1,1)
yyaxis left
plot(ECOG(plotinterval,1),ECOG(plotinterval,2),'k')
hold on
plot(ECOG(plotinterval,1),filt_sig(plotinterval),'b')
vline(min(ECOG(spinintervalIX,1)))
vline(max(ECOG(spinintervalIX,1)))
axis tight
yyaxis right
plot(smoothed_power(plotinterval,1),smoothed_power(plotinterval,2),'r') % now 1= 1 second and this should be easy to estimate Hz from
refline(0,lower_spin_thresh)
refline(0,upper_spin_thresh)
ylabel('uV')
xlabel('Time(s)')
title(sprintf('Duration %2.2f secs %1.1f frex',spinlength,peakfrex))
subplot(3,1,2)
Spectrogram_spindle_LC(ECOG(plotinterval,2), LFP.fs_final, 3:0.2:20, 'wavelet');
subplot(3,1,3)
pburg(double(ECOG(spinintervalIX,2)),40,spec_freqs,LFP.fs_final);
end
Spindle.Post(count2).raw_spin= putative_spindle;
Spindle.Post(count2).length_spin_sec=spinlength;
Spindle.Post(count2).filt_sig=filt_sig(spinintervalIX,1);
Spindle.Post(count2).mean_sig_pow=mean(smoothed_power(spinintervalIX,2));
Spindle.Post(count2).time=ECOG(spinintervalIX,1);
Spindle.Post(count2).values=ECOG(spinintervalIX,2);
Spindle.Post(count2).peakfrex=peakfrex;
Spindle.Post(count2).DayType=Session_Info.RecordingType{1};
Spindle.Post(count2).Mouse=mouseday(1:3);
Spindle.Post(count2).Day=Session_Info.Recordingday;
ixx=Session_Info.drugs.Mouse==mouseday(1:3);
Spindle.Post(count2).drugmouse=Session_Info.drugs.GotDrug(ixx);
Spindle.Post(count2).MouseType=mouseday(1);
Spindle.Post(count2).SleepSession='Post';
Spindle.Post(count2).psd=pxx;
Spindle.Post(count2).psd_frex=frex;
Spindle.Post(count2).location=bits{2};
end
end
end