-
Notifications
You must be signed in to change notification settings - Fork 0
/
dfs_decodequant_movement.m
277 lines (254 loc) · 15.9 KB
/
dfs_decodequant_movement.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
%% Sungod Manu: Plot Decoding error during movement times
animals = {'jaq','roquefort','despereaux','montague'}; %
for a = 1:length(animals)
epochfilter{1} = ['$ripthresh>=0 & (isequal($environment,''goal'')) & $forageassist==0']; %& $epoch==2
%epochfilter{1} = ['$session==6 & isequal($environment,''goal'')'];
%epochfilter{1} = ['$ripthresh>=0 & (isequal($environment,''goal'') | isequal($environment,''hybrid2'') | isequal($environment,''hybrid3''))'];
% resultant excludeperiods will define times when velocity is high
timefilter{1} = {'ag_get2dstate', '($immobility == 1)','immobility_velocity',4,'immobility_buffer',0};
iterator = 'epochbehaveanal';
f(a) = createfilter('animal',animals{a},'epochs',epochfilter,'excludetime', timefilter, 'iterator', iterator);
%args:
%converter = {[
f(a) = setfilterfunction(f(a), 'dfa_decodequant_movement', {'pos','ca1rippleskons','ripdecodesv3','trials'},'animal',animals{a},'useclassifier',1,'vers','v3');
f(a) = runfilter(f(a));
end
%save('/media/anna/whirlwindtemp2/ffresults/ctrl_movementquant_full2state_all_withtrialwise.mat','f','-v7.3')
animcol = [27 92 41; 25 123 100; 33 159 169; 123 225 191]./255; %ctrlcols
%load('/media/anna/whirlwindtemp2/ffresults/ctrl_movementquant_full2state_gooddecode.mat')
%load('/media/anna/whirlwindtemp2/ffresults/ctrl_movementquant_full2state_all_withtrialwise.mat')
%% 1. heatmaps of error distributions per animal across days and summary of all days
all = figure(); set(gcf,'Position',[369 473 1387 408]);
animcols = get(0,'DefaultAxesColorOrder');
for a = 1:length(animals)
eps = find(arrayfun(@(x) ~isempty(x.errorhist_bybin),f(a).output{1}));
erange = f(a).output{1}(eps(1)).errorhist_edges(1:end-1);
figure;
for e = 1:length(eps)
if size(f(a).output{1}(eps(e)).errorhist_bybin,1)<146
orighist = [f(a).output{1}(eps(e)).errorhist_bybin; nan(1,50)];
disp('adding a row')
else
orighist = f(a).output{1}(eps(e)).errorhist_bybin;
end
% normedhist(:,:,e) = orighist./repmat(sum(orighist,2),1,length(erange));
normedhist(:,:,e) = orighist;
if e>24 %plot the first 24
break
end
subplot(4,6,e); hold on;
imagesc(erange,[1:146],normedhist(:,:,e)); set(gca,'YDir','normal'); caxis([0 100]); colorbar;
axis tight; title(sprintf('%s%de%d',animals{a}(1:3),f(a).epochs{1}(eps(e),1),f(a).epochs{1}(eps(e),2)))
end
animmean{a} = nanmean(normedhist,3); % stack data from all trials
figure(all); subplot(2,3,a); imagesc(erange,[1:146],animmean{a});
set(gca,'YDir','normal'); title(animals{a}); axis tight; ylabel('posbin'); caxis([0 100]); colorbar; xlabel('decode leads ------ decode lags')
clear normedhist
end
%% 2. plot measurable fraction, mean error, error across eps/posbins
figure;
% 1-measurable = arm-arm; measurable-sameseg = box-arm
for a = 1:length(animals)
measurable{a} = cell2mat(arrayfun(@(x) x.measurable_bin_frac,f(a).output{1},'UniformOutput',0))';
sameseg = cell2mat(arrayfun(@(x) x.same_seg,f(a).output{1},'UniformOutput',0))';
bycat{a} = [sameseg, measurable{a}-sameseg, 1-measurable{a}];
binerr = arrayfun(@(x) abs(x.meanerror),f(a).output{1},'UniformOutput',0)';
short = find(cellfun(@(x) length(x)<140,binerr));
for s = short'
binerr{s} = [binerr{s} nan(1,140-length(binerr{s}))];
end
binerr = vertcat(binerr{:});
meanerr{a} = nanmean(binerr,2);
%if we want to exclude the worse days:
%binerr(meanerr{a}>1,:) = nan;
%meanerr{a}(meanerr{a}>1) = nan;
subplot(4,2,[3 4]); hold on; title('error over eps')
plot(repmat([1:size(binerr,1)],2,1),[nanmean(binerr,2)'+nanstd(binerr,[],2)'./sqrt(size(binerr,2));nanmean(binerr,2)'-nanstd(binerr,[],2)'./sqrt(size(binerr,2))],'Color',animcol(a,:));
plot([1:size(binerr,1)], nanmean(binerr,2),'.-','Color',animcol(a,:),'Linewidth',2);
subplot(4,2,[5 6]); hold on; title('error over posbins')
plot(repmat([1:size(binerr,2)],2,1),[nanmean(binerr)+nanstd(binerr)./sqrt(sum(~isnan(binerr(1,:))));nanmean(binerr)-nanstd(binerr)./sqrt(sum(~isnan(binerr(1,:))))],'Color',animcol(a,:));
plot([1:size(binerr,2)], nanmean(binerr),'.-','Color',animcol(a,:),'Linewidth',2);
segnum = [0, cumsum(diff(~isnan(binerr(2,:)))==1)]+1;
segnum(isnan(binerr(2,:)))=nan;
errbyseg{a} = grpstats(binerr',segnum,'mean')';
end
subplot(4,2,1); hold on; plot4a(bycat,'gnames',{'sameseg','box-arm','diffarms'}); title('by category'); ylim([0 1]); ylabel('fraction samples');
text(9,.5,sprintf('ss mean: %.02f',mean(cellfun(@(x) nanmean(x(:,1)),bycat))))
text(9,.3,['n=' num2str(cellfun(@(x) sum(~isnan(x(:,1))), bycat))])
subplot(4,2,2); hold on; plot4a(meanerr); title('mean err'); ylabel('deviation (bins)'); xlabel('over epochs'); ylim([0 2])
subplot(4,2,[7 8]); hold on; plot4a(errbyseg,'gnames',{'box','1','2','3','4','5','6','7','8'}); ylim([0 5]); ylabel('deviation (bins)');
%% 3. plot measurable fraction, mean error, CIs, numtets, numspikes
figure;
parentdir = '/mnt/stelmo/anna/';
for a = 1:length(animals)
sameseg{a} = cell2mat(arrayfun(@(x) x.same_seg,f(a).output{1},'UniformOutput',0))';
binerr = arrayfun(@(x) abs(x.meanerror),f(a).output{1},'UniformOutput',0)';
short = find(cellfun(@(x) length(x)<140,binerr));
for s = short'
binerr{s} = [binerr{s} nan(1,140-length(binerr{s}))];
end
binerr = vertcat(binerr{:});
meanerr{a} = nanmean(binerr,2);
subplot(4,1,1); hold on; title('frac same segment'); plot(sameseg{a},'.-','Color',animcol(a,:)); ylim([.5 1])
subplot(4,1,2); hold on; title('mean err during movement')
plot(repmat([1:size(binerr,1)],2,1),[nanmean(binerr,2)'+nanstd(binerr,[],2)'./sqrt(size(binerr,2));nanmean(binerr,2)'-nanstd(binerr,[],2)'./sqrt(size(binerr,2))],'Color',animcol(a,:));
plot([1:size(binerr,1)], nanmean(binerr,2),'.-','Color',animcol(a,:));
subplot(4,1,3); hold on; title('CI')
movemeanCI = cell2mat(arrayfun(@(x) x.movemeanCI,f(a).output{1},'UniformOutput',0))';
plot(movemeanCI,'.-','Color',animcol(a,:))
immomeanCI = cell2mat(arrayfun(@(x) x.immomeanCI_norips,f(a).output{1},'UniformOutput',0))';
plot(immomeanCI,'.:','Color',animcol(a,:));
ripsmeanCI = cell2mat(arrayfun(@(x) x.immomeanCI_rips,f(a).output{1},'UniformOutput',0))';
plot(ripsmeanCI,'.--','Color',animcol(a,:)); ylim([0 75]);
subplot(4,1,4); hold on; title('tetnums')
tetinfo = loaddatastruct([parentdir animals{a} '/filterframework/'],animals{a},'tetinfo');
tetlist = evaluatefilter(tetinfo,'isequal($area,''ca1'')');
tetnum = table2array(rowfun(@(x) sum(tetlist(:,1)==x(1) & tetlist(:,2)==x(2)),table(f(a).epochs{1})));
plot(tetnum,'.-','Color',animcol(a,:))
end
%% 4. save mean error into taskstruct for each epoch for future filtering
for a = 1:length(animals)
destdir = sprintf('/mnt/stelmo/anna/%s/filterframework/',animals{a});
dayep = f(a).epochs{1};
binerr = arrayfun(@(x) nanmean(abs(x.meanerror)),f(a).output{1},'UniformOutput',0)';
sameseg = arrayfun(@(x) x.same_seg,f(a).output{1},'UniformOutput',0)';
for e = 1:size(dayep,1)
AG_addtaskinfo(destdir,animals{a},dayep(e,1),dayep(e,2),'decode_error',binerr{e});
if sameseg{e}<.75
AG_addtaskinfo(destdir,animals{a},dayep(e,1),dayep(e,2),'gooddecode',0);
else
AG_addtaskinfo(destdir,animals{a},dayep(e,1),dayep(e,2),'gooddecode',1);
end
end
end
%% 5. plot mean decoding error per day
figure; set(gcf,'Position',[369 473 862 408]);
animcols = get(0,'DefaultAxesColorOrder');
for a = 1:length(animals)
eps = find(arrayfun(@(x) ~isempty(x.meanerror),f(a).output{1}));
meanerr = (arrayfun(@(x) x.meanerror,f(a).output{1}(eps),'UniformOutput',0));
wronglength = cellfun(@(x) size(x,2),meanerr)~= median(cellfun(@(x) size(x,2),meanerr));
disp(sprintf('%s %d eps with abnormal length',animals{a},sum(wronglength)));
meanerr = vertcat(meanerr{~wronglength});
subplot(2,1,1); hold on; plot(eps(~wronglength),nanmean(meanerr,2),'-o','Color',animcols(a,:),'MarkerSize',3); title('mean error across track by epoch')
subplot(2,1,2); hold on; plot(nanmean(meanerr,1),'-o','Color',animcols(a,:),'MarkerSize',3); title('mean error across days by posbin')
end
%% 6. analyze decoding error over the course of a trial block
clearvars -except f animals animcol
bybin = figure(); set(gcf,'Position',[675 1 1202 973]);
slopes = figure(); set(gcf,'Position',[675 1 1202 973]);
for a = 1:length(animals)
eps = find(arrayfun(@(x) ~isempty(x.trialwise),f(a).output{1}));
for e = 1:length(eps)
tphase = f(a).output{1}(eps(e)).trialwise.taskphase;
tphase(mod(tphase,1)>.85) = tphase(mod(tphase,1)>.85)+.1; % .9 trials should be considere rewarded visits in this case
tnum = [1:length(tphase)]';
rewarded = tphase>0 & mod(tphase,1)==0;
outers = f(a).output{1}(eps(e)).trialwise.outerarm;
diffs = f(a).output{1}(eps(e)).trialwise.diffbybin;
repstarts = [find(tphase==1); length(outers)+1];
for c = 1:length(repstarts)-1;
goalarm = outers(repstarts(c));
if any(outers==goalarm & tnum<repstarts(c))
pre{c} = diffs(outers==goalarm & tnum<repstarts(c),:);
else
pre{c} = nan(1,13);
end
dur{c} = nan(16,13);
dur{c}(1:sum(rewarded & tnum>=repstarts(c) & tnum<repstarts(c+1)),:) = diffs(rewarded & tnum>=repstarts(c) & tnum<repstarts(c+1),:);
errs{c} = diffs(~rewarded & tphase>0 & tnum>=repstarts(c) & tnum<repstarts(c+1),:);
last{c} = dur{c}(sum(rewarded & tnum>=repstarts(c) & tnum<repstarts(c+1)),:);
if any(outers==goalarm & tnum>repstarts(c+1)) % visits to the previous goal arm later in the ep
post{c} = diffs(outers==goalarm & tnum>repstarts(c+1),:); % does not include visits to goal in subs search phase
else
post{c} = nan(1,13);
end
end
ep_pre{e} = vertcat(pre{:});
ep_dur{e} = cat(3,dur{:});
ep_errs{e} = vertcat(errs{:});
ep_last{e} = vertcat(last{:});
ep_post{e} = vertcat(post{:});
clear pre post dur errs last
end
% full arm
figure(bybin)
allpre = vertcat(ep_pre{:}); allpost = vertcat(ep_post{:}); alllast = vertcat(ep_last{:}); alldur = cat(3,ep_dur{:});
subplot(2,4,a); hold on
%boxplot(nanmean(allpre,2),'Positions',1,'Symbol','','Widths',.8,'Colors',animcol(a,:))
boxplot(squeeze(nanmean(alldur(1,:,:),2)),'Positions',2,'Symbol','','Widths',.8,'Colors',animcol(a,:));
boxplot(nanmean(alllast,2),'Positions',3,'Symbol','','Widths',.8,'Colors',animcol(a,:));
%boxplot(nanmean(allpost,2),'Positions',4,'Symbol','','Widths',.8,'Colors',animcol(a,:));
xlim([0 5]); set(gca,'xtick',[1:4],'xticklabel',{'pre','first','last','post'}); title('full arm'); ylim([0 3]);
%text(1,2.5,sprintf('pre-post %.05f',ranksum(nanmean(allpre,2),nanmean(allpost,2))))
text(1,2,sprintf('first-last %.05f',ranksum(squeeze(nanmean(alldur(1,:,:),2)),nanmean(alllast,2))))
subplot(2,4,a+4); hold on
%boxplot(nanmean(allpre(:,9:13),2),'Positions',1,'Symbol','','Widths',.8,'Colors',animcol(a,:))
boxplot(squeeze(nanmean(alldur(1,9:13,:),2)),'Positions',2,'Symbol','','Widths',.8,'Colors',animcol(a,:));
boxplot(nanmean(alllast(:,9:13),2),'Positions',3,'Symbol','','Widths',.8,'Colors',animcol(a,:));
%boxplot(nanmean(allpost(:,9:13),2),'Positions',4,'Symbol','','Widths',.8,'Colors',animcol(a,:));
xlim([0 5]); set(gca,'xtick',[1:4],'xticklabel',{'pre','first','last','post'}); title('last 25cm'); ylim([0 5]); ylabel('deviation')
%text(1,3.5,sprintf('pre-post %.05f',ranksum(nanmean(allpre(:,9:13),2),nanmean(allpost(:,9:13),2))))
text(1,3,sprintf('first-last %.05f',ranksum(squeeze(nanmean(alldur(1,9:13,:),2)),nanmean(alllast(:,9:13),2))))
text(-1,3.5,sprintf('n=%d',sum(~isnan(nanmean(alldur(1,9:13,:),2)))));
% calculate slope for each contingency
for cont = 1:size(alldur,3)
tris = nanmean(alldur(:,:,cont),2);
if sum(~isnan(tris))>=4 %don't calculate slope if there are less than 4 trials
slope(cont,:) = polyfit(find(~isnan(tris)),tris(~isnan(tris)),1);
tris = nanmean(alldur(:,9:13,cont),2);
slopelast(cont,:) = polyfit(find(~isnan(tris)),tris(~isnan(tris)),1);
else
slope(cont,:) = [nan nan];
slopelast(cont,:) = [nan nan];
end
end
figure(slopes)
subplot(2,4,a); hold on; histogram(slope(:,1),[-1:.05:1],'Normalization','probability','FaceColor',animcol(a,:));
title([animals{a} ' fullarm']); ylim([0 .6]); ylabel('frac blocks'); xlabel('slope'); xlim([-1 1])
text(-1,.4,sprintf('n=%d',sum(~isnan(slope(:,1)))));
subplot(2,4,a+4); hold on; histogram(slopelast(:,1),[-1:.05:1],'Normalization','probability','FaceColor',animcol(a,:));
title([animals{a} ' last 25cm']); ylim([0 .6]); ylabel('frac blocks'); xlabel('slope'); xlim([-1 1])
text(-1,.4,sprintf('n=%d',sum(~isnan(slopelast(:,1)))));
% errorbar([1:4]',[nanmean(nanmean(allpre(:,1:4),2)); nanmean(nanmean(alldur(1,1:4,:),2),3); nanmean(nanmean(alllast(:,1:4),2)); nanmean(nanmean(allpost(:,1:4),2))], ...
% [nanstd(nanmean(allpre(:,1:4),2)); nanstd(nanmean(alldur(1,1:4,:),2),[],3); nanstd(nanmean(alllast(:,1:4),2)); nanstd(nanmean(allpost(:,1:4),2))],'k.');
% bar([1:4]',[nanmean(nanmean(allpre(:,1:4),2)); nanmean(nanmean(alldur(1,1:4,:),2),3); nanmean(nanmean(alllast(:,1:4),2)); nanmean(nanmean(allpost(:,1:4),2))],'FaceColor',animcol(a,:));
% set(gca,'xticklabel',{'pre','first','last','post'})
% bar(18,nanmean(nanmean(allpost(:,1:4),2)),'FaceColor',cols(18,:),'Linewidth',2)
% title('arm start'); axis tight
% cols = cool(18);
% figure(bybin); subplot(3,4,a) ; hold on;
% plot(nanmean(vertcat(ep_pre{:})),':','Color',cols(1,:),'Linewidth',2);
% alldur = cat(3,ep_dur{:}); %alldur = nanstd(cat(3,ep_dur{:}),[],3);
% for t =1:16
% plot(nanmean(alldur(t,:,:),3),'Color',cols(t,:),'Linewidth',1)
% end
% plot(nanmean(vertcat(ep_post{:})),':','Color',cols(18,:),'Linewidth',2);
% xlabel('arm bins'); xlim([0 14]) % legend({'pre','1','2','3','4','5','6','7','8','9','10','11','12','13','14','15','16','post'})
%
% subplot(3,4,a+4); hold on
% allpre = vertcat(ep_pre{:}); allpost = vertcat(ep_post{:});
% errorbar([-1, 1:16,18]',[nanmean(nanmean(allpre(:,1:4),2)); nanmean(nanmean(alldur(:,1:4,:),2),3); nanmean(nanmean(allpost(:,1:4),2))],[nanstd(nanmean(allpre(:,1:4),2)); nanstd(nanmean(alldur(:,1:4,:),2),[],3); nanstd(nanmean(allpost(:,1:4),2))],'k.');
% bar(-1,nanmean(nanmean(allpre(:,1:4),2)),'FaceColor',cols(1,:),'Linewidth',2)
% for t =1:16
% if sum(~isnan(nanmean(alldur(t,1:4,:),2)))>=10
% bar(t,nanmean(nanmean(alldur(t,1:4,:),2),3),'FaceColor',cols(t+1,:),'Linewidth',2)
% end
% end
% bar(18,nanmean(nanmean(allpost(:,1:4),2)),'FaceColor',cols(18,:),'Linewidth',2)
% title('arm start'); axis tight
% subplot(3,4,a+8); hold on
% errorbar([-1, 1:16,18]',[nanmean(nanmean(allpre(:,10:13),2)); nanmean(nanmean(alldur(:,10:13,:),2),3); nanmean(nanmean(allpost(:,10:13),2))],[nanstd(nanmean(allpre(:,10:13),2)); nanstd(nanmean(alldur(:,10:13,:),2),[],3); nanstd(nanmean(allpost(:,10:13),2))],'k.');
% bar(-1,nanmean(nanmean(allpre(:,10:13),2)),'FaceColor',cols(1,:),'Linewidth',2)
% for t =1:16
% if sum(~isnan(nanmean(alldur(t,10:13,:),2)))>=10
% bar(t,nanmean(nanmean(alldur(t,10:13,:),2),3),'FaceColor',cols(t+1,:),'Linewidth',2)
% end
% end
% bar(18,nanmean(nanmean(allpost(:,10:13),2)),'FaceColor',cols(18,:),'Linewidth',2)
% title('arm end'); axis tight
end
%plot example slope for one contingency
figure
plot(1:16,nanmean(alldur(:,:,2),2),'.','MarkerSize',20); lsline; ylim([0 4]); ylabel('deviation'); xlabel('rewarded visit number')
title('single block example, monty block2')