Conserved post-odor dynamics in the olfactory systems of mice and locusts.

Code ↔ Paper

1 match between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 1 match

[1] § STAR★Methods › Method details › Odor presentation paradigms ↔ Script_Fig6.m, lines 114–172 · score 0.63 · ani, asa, cin, ep, ieg, mpz

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The paper is loaded by your browser, with JavaScript: read it at doi.org or on Europe PMC.

The authors' code

MATLAB · 1,148 lines · 41 KB · CC-BY-4.0 · 1 match

%% ==================== Figure 6 ====================
close all; clc; clear;
fileNameKNN = {'m129_1_2_dataset'; 'm149_3_1_dataset';};
% Fig6a, b, g, h, i: KNN Classification of mouse glom dataset and average over sessions
% a/b => fileIter = 2;
% h => fileIter = 1;
KNN = 10;
numTrial = 10;
for fileIter = 1:length(fileNameKNN)
clearvars -except fileNameKNN fileIter KNN numTrial
load([fileNameKNN{fileIter},'\filtData2.mat']);
if fileIter == 1, numOdor = length(fieldnames(filtData)) - 1;
else, numOdor = length(fieldnames(filtData));
end
for OnOffIter = 1:2
if OnOffIter == 1
range = preDuration*fps+1:(preDuration+onDuration)*fps;
else
range = (preDuration+onDuration)*fps+1:(afterDuration+preDuration+onDuration)*fps;
end
numBinAvg = 1 + 3*(OnOffIter==2);
% ====== Restructure the dataset ======
newBlockData = cell(numOdor,1);
clear tempFiltData2
names = fieldnames(filtData);
for i = 1:numOdor
eval(strcat('tempData = filtData.', names{i+(fileIter==1)},';'));
tempFiltData = tempData(1:numTrial,range,:);
iter = ceil(length(range)/numBinAvg);
tempFiltData2 = zeros(numTrial, iter, size(tempFiltData,3));
for ii = 1:numBinAvg:length(range)
try
tempFiltData2(:,(ii-1)/numBinAvg+1,:) = mean(tempFiltData(:,ii:ii+numBinAvg-1,:),2);
catch
tempFiltData2(:,(ii-1)/numBinAvg+1,:) = mean(tempFiltData(:,ii:end,:),2);
end
end
newBlockData{i} = tempFiltData2;
end
timeLength = size(tempFiltData2,2);
% ====== Reshape to units x gloms ======
allUnits = [];
clusterAssignment = [];
for i = 1:numOdor
unitVector = newBlockData{i};
tempUnits = reshape(permute(unitVector,[2 1 3]),[],size(unitVector,3));
allUnits = [allUnits; tempUnits];
clusterAssignment = [clusterAssignment; i*ones(size(tempUnits,1),1)];
end
numUnits = size(allUnits,1);
% ====== Correlation distance with leave-trial-out block ======
tempDistMatrix = squareform(pdist(allUnits,'correlation'));
diagValue = ones(1,numUnits)*10;
fullDistMatrix = tempDistMatrix + diag(diagValue);
trialBlock = ones(timeLength)*10;
trialBlockMatrix = kron(eye(numOdor*numTrial), trialBlock);
fullDistMatrix = fullDistMatrix + trialBlockMatrix*10;
% ====== kNN ======
[~, idx] = sort(fullDistMatrix,2,"ascend");
colIndices = idx(:,1:KNN);
clusterTemp = zeros(size(colIndices));
for nn = 1:KNN
clusterTemp(:,nn) = clusterAssignment(colIndices(:,nn),1);
end
clusterAssignment(:,2) = mode(clusterTemp(:,1:KNN),2);
clusterAssignment(:,3) = clusterAssignment(:,2) == clusterAssignment(:,1);
perCorrectCluster = sum(clusterAssignment(:,3)) ./ size(clusterAssignment(:,1),1);
disp(['Percentage correctly clustered = ', num2str(perCorrectCluster)]);
% ====== Rearrange + percent ======
clear resultMatrixTrialPerRow resultMatrix percentMatrix
resultMatrix = zeros(numOdor,timeLength*numTrial);
resultMatrixTrialPerRow = zeros(numOdor*numTrial,timeLength);
percentMatrixAllTrial = zeros(numOdor,numTrial,timeLength);
percentMatrix = zeros(numOdor+1,timeLength);
for i = 1:numOdor
blockIndex = find(clusterAssignment(:,1)==i);
resultMatrix(i,1:length(blockIndex)) = clusterAssignment(blockIndex,2);
resultMatrixTrialPerRow((i-1)*numTrial+1:i*numTrial,:) = reshape(resultMatrix(i,:),timeLength,numTrial)';
tempPercent = clusterAssignment(blockIndex,3);
percentMatrixTrialPerRow = reshape(tempPercent,timeLength,numTrial)';
percentMatrixAllTrial(i,:,:) = percentMatrixTrialPerRow;
percentMatrix(i,:) = mean(percentMatrixTrialPerRow,1);
end
percentMatrix(end,:) = mean(percentMatrix(1:numOdor,:),1);
if fileIter == 1
smooth = 3 + (OnOffIter==2);
else
smooth = 2 + 2*(OnOffIter==2);
end
percentMatrix(1:end-1,:) = filtfilt(ones(1,smooth)/smooth,1,percentMatrix(1:end-1,:)')';
percentMatrix(end,:) = filtfilt(ones(1,smooth)/smooth,1,percentMatrix(end,:));
% ====== Plot Matrix ======
figure; color = ColorScheme(numOdor);
try
imagesc(resultMatrixTrialPerRow,[0.5 length(odorNames)+0.5]); colormap(color); cbar = colorbar;
catch
imagesc(resultMatrixTrialPerRow,[0.5 length(odorList)-0.5]); colormap(color); cbar = colorbar;
end
xlim([0.5 timeLength]); xticks(linspace(0.5,timeLength,9));
if OnOffIter == 1
xticklabels(linspace(0,onDuration,9));
xlabel 'Mouse ON, time after odor onset(s)'; title 'ON Prediction';
else
xticklabels(linspace(0,afterDuration,9));
xlabel 'Mouse ITI, time after odor offset (s)'; title 'ITI Prediction';
end
yticks(5:10:numOdor*numTrial);
cbar.Ticks = 1:numOdor;
if fileIter == 1
yticklabels({'ace';'eug';'cin';'mpz';'2ep';'ieg';'tcm';'als';'met';'ani';'asa';});
cbar.TickLabels={'ace';'eug';'cin';'mpz';'2ep';'ieg';'tcm';'als';'met';'ani';'asa';};
else
yticklabels({'ace';'msc';'eug';'als'})
cbar.TickLabels={'ace';'msc';'eug';'als'};
end
cbar.Direction = "reverse";
ylabel('True Label'); axis square;
% ====== Plot Percentage ======
if OnOffIter == 1
timeAxis = preDuration+1/fps : 1/fps : (preDuration+onDuration);
else
timeAxis = (preDuration+onDuration)+1/fps : 1/fps : (preDuration+onDuration+afterDuration);
end
timeAxis = timeAxis - timeAxis(1);
timeAxis = timeAxis(1:numBinAvg:end);
figure; color = ColorScheme(numOdor,0.2); hold on;
for i = 1:size(percentMatrix,1)-1
plot(timeAxis(1:size(percentMatrix,2)), percentMatrix(i,:)*100, 'LineStyle','-','LineWidth',2,'Color',color(i,:));
end
plot(timeAxis(1:size(percentMatrix,2)), percentMatrix(end,:)*100, 'Color','k','LineWidth',5);
line([0 (range(end)-range(1))/fps], [1/numOdor*100 1/numOdor*100],'LineStyle',':');
xlim([0 (range(end)-range(1))/fps]);
xticks(linspace(0,(range(end)-range(1))/fps,9));
if OnOffIter == 1
xticklabels(linspace(0,onDuration,9));
xlabel 'Mouse ON, time after odor onset (s)'; title 'ON Prediction';
else
xticklabels(linspace(0,afterDuration,9));
xlabel 'Mouse ITI, time after odor offset (s)'; title 'ITI Prediction';
end
ylabel 'Classification rate (%)'; axis square;
end
end
% -------- Average Across Sessions for A & B--------------
fileNameKNN = {
'm149_1_2_dataset';
'm148_2_1_dataset';
'm149_3_1_dataset';
'm141_8_1_dataset';
'm141_5_1_dataset';
'm141_1_1_dataset';
};
KNN = 10;
numTrial = 10;
totalPercentMatrixON = nan(length(fileNameKNN), 200);
totalPercentMatrixOFF = nan(length(fileNameKNN), 500);
commonTimeON = linspace(0, 5, 200);
commonTimeOFF = linspace(0, 50, 500);
for fileIter = 1:length(fileNameKNN)
clearvars -except fileNameKNN fileIter KNN numTrial totalPercentMatrixON totalPercentMatrixOFF commonTimeON commonTimeOFF
load([fileNameKNN{fileIter},'\filtData2.mat']);
numOdor = length(fieldnames(filtData));
names = fieldnames(filtData);
for OnOffIter = 1:2
if OnOffIter == 1
range = preDuration*fps+1:(preDuration+onDuration)*fps;
else
range = (preDuration+onDuration)*fps+1:(afterDuration+preDuration+onDuration)*fps;
end
numBinAvg = 1 + 3*(OnOffIter==2);
newBlockData = cell(numOdor,1);
clear tempFiltData2
for i = 1:numOdor
eval(strcat('tempData = filtData.', names{i},';'));
tempFiltData = tempData(1:numTrial,range,:);
iter = ceil(length(range)/numBinAvg);
tempFiltData2 = zeros(numTrial, iter, size(tempFiltData,3));
for ii = 1:numBinAvg:length(range)
try
tempFiltData2(:,(ii-1)/numBinAvg+1,:) = mean(tempFiltData(:,ii:ii+numBinAvg-1,:),2);
catch
tempFiltData2(:,(ii-1)/numBinAvg+1,:) = mean(tempFiltData(:,ii:end,:),2);
end
end
newBlockData{i} = tempFiltData2;
end
timeLength = size(tempFiltData2,2);
allUnits = [];
clusterAssignment = [];
for i = 1:numOdor
unitVector = newBlockData{i};
tempUnits = reshape(permute(unitVector,[2 1 3]),[],size(unitVector,3));
allUnits = [allUnits; tempUnits];
clusterAssignment = [clusterAssignment; i*ones(size(tempUnits,1),1)];
end
numUnits = size(allUnits,1);
tempDistMatrix = squareform(pdist(allUnits,'correlation'));
diagValue = ones(1,numUnits)*10;
fullDistMatrix = tempDistMatrix + diag(diagValue);
trialBlock = ones(timeLength)*10;
trialBlockMatrix = kron(eye(numOdor*numTrial), trialBlock);
fullDistMatrix = fullDistMatrix + trialBlockMatrix*10;
[~, idx] = sort(fullDistMatrix,2,"ascend");
colIndices = idx(:,1:KNN);
clusterTemp = zeros(size(colIndices));
for nn = 1:KNN
clusterTemp(:,nn) = clusterAssignment(colIndices(:,nn),1);
end
clusterAssignment(:,2) = mode(clusterTemp(:,1:KNN),2);
clusterAssignment(:,3) = clusterAssignment(:,2) == clusterAssignment(:,1);
perCorrectCluster = sum(clusterAssignment(:,3)) ./ size(clusterAssignment(:,1),1);
disp(['Percentage correctly clustered = ', num2str(perCorrectCluster)]);
clear resultMatrixTrialPerRow resultMatrix percentMatrix
resultMatrix = zeros(numOdor,timeLength*numTrial);
resultMatrixTrialPerRow = zeros(numOdor*numTrial,timeLength);
percentMatrixAllTrial = zeros(numOdor,numTrial,timeLength);
percentMatrix = zeros(numOdor+1,timeLength);
for i = 1:numOdor
blockIndex = find(clusterAssignment(:,1)==i);
resultMatrix(i,1:length(blockIndex)) = clusterAssignment(blockIndex,2);
resultMatrixTrialPerRow((i-1)*numTrial+1:i*numTrial,:) = reshape(resultMatrix(i,:),timeLength,numTrial)';
tempPercent = clusterAssignment(blockIndex,3);
percentMatrixTrialPerRow = reshape(tempPercent,timeLength,numTrial)';
percentMatrixAllTrial(i,:,:) = percentMatrixTrialPerRow;
percentMatrix(i,:) = mean(percentMatrixTrialPerRow,1);
end
percentMatrix(end,:) = mean(percentMatrix(1:numOdor,:),1);
if OnOffIter == 1, smooth = 2; else, smooth = 4; end
percentMatrix(1:end-1,:) = filtfilt(ones(1,smooth)/smooth,1,percentMatrix(1:end-1,:)')';
percentMatrix(end,:) = filtfilt(ones(1,smooth)/smooth,1,percentMatrix(end,:));
if OnOffIter == 1
origTimeON = linspace(0,onDuration,length(percentMatrix(end,:)));
interpCurveON = interp1(origTimeON, percentMatrix(end,:), commonTimeON,'linear');
totalPercentMatrixON(fileIter,:) = interpCurveON;
else
origTimeOFF = linspace(0,afterDuration,length(percentMatrix(end,:)));
interpCurveOFF = interp1(origTimeOFF, percentMatrix(end,:), commonTimeOFF,'linear');
totalPercentMatrixOFF(fileIter,:) = interpCurveOFF;
end
end
end
totalPercentMatrixON(end+1,:) = mean(totalPercentMatrixON, 1,'omitnan');
totalPercentMatrixOFF(end+1,:) = mean(totalPercentMatrixOFF, 1,'omitnan');
timeAxisON = linspace(0,onDuration,size(totalPercentMatrixON,2));
figure; hold on;
for r = 1:fileIter
plot(timeAxisON, totalPercentMatrixON(r,:)*100,'Color',[0.7 0.7 0.7],'LineWidth',2);
end
plot(timeAxisON, totalPercentMatrixON(end,:)*100,'Color','k','LineWidth',4);
chanceLevel = 100/numOdor; yline(chanceLevel,':','Chance','LabelHorizontalAlignment','left');
xlim([0 onDuration]); xticks(linspace(0,onDuration,9)); xticklabels(linspace(0,onDuration,9));
xlabel('Mouse ON, time after odor onset (s)'); ylabel('Classification rate (%)'); title('ON Prediction (n = 6)'); axis square;
timeAxisOFF = linspace(0,afterDuration,size(totalPercentMatrixOFF,2));
figure; hold on;
for r = 1:fileIter
plot(timeAxisOFF, totalPercentMatrixOFF(r,:)*100,'Color',[0.7 0.7 0.7],'LineWidth',2);
end
plot(timeAxisOFF, totalPercentMatrixOFF(end,:)*100,'Color','k','LineWidth',4);
chanceLevel = 100/numOdor; yline(chanceLevel,':','Chance','LabelHorizontalAlignment','left');
xlim([0 afterDuration]); xticks(linspace(0,afterDuration,9)); xticklabels(linspace(0,afterDuration,9));
xlabel('Mouse ITI, time after odor offset (s)'); ylabel('Classification rate (%)'); title('ITI Prediction (n = 6)'); axis square;
% ---------- Average Across Session for H -------------
fileNameKNN = {
'm129_1_2_dataset';
'm122_4_4_dataset';
'm125_1_1_dataset';
'm126_1_1_dataset';
'm129_3_1_dataset';
'm129_4_2_dataset';
'm134_6_1_dataset';
'm1716_1_1_dataset';
};
KNN = 10;
numTrial = 10;
totalPercentMatrixON = nan(length(fileNameKNN), 25);
totalPercentMatrixOFF = nan(length(fileNameKNN), 600);
commonTimeON = linspace(0, 1, 25);
commonTimeOFF = linspace(0, 17, 600);
for fileIter = 1:length(fileNameKNN)
clearvars -except fileNameKNN fileIter KNN numTrial commonTimeOFF commonTimeON totalPercentMatrixON totalPercentMatrixOFF
load([fileNameKNN{fileIter},'\filtData2.mat']);
numOdor = length(fieldnames(filtData)) - 1;
names = fieldnames(filtData);
for OnOffIter = 1:2
if OnOffIter == 1
range = preDuration*fps+1:(preDuration+onDuration)*fps;
else
range = (preDuration+onDuration)*fps+1:(afterDuration+preDuration+onDuration)*fps;
end
numBinAvg = 1 + 3*(OnOffIter==2);
newBlockData = cell(numOdor,1);
clear tempFiltData2
for i = 1:numOdor
eval(strcat('tempData = filtData.', names{i+1},';'));
tempFiltData = tempData(1:numTrial,range,:);
iter = ceil(length(range)/numBinAvg);
tempFiltData2 = zeros(numTrial, iter, size(tempFiltData,3));
for ii = 1:numBinAvg:length(range)
try
tempFiltData2(:,(ii-1)/numBinAvg+1,:) = mean(tempFiltData(:,ii:ii+numBinAvg-1,:),2);
catch
tempFiltData2(:,(ii-1)/numBinAvg+1,:) = mean(tempFiltData(:,ii:end,:),2);
end
end
newBlockData{i} = tempFiltData2;
end
timeLength = size(tempFiltData2,2);
allUnits = [];
clusterAssignment = [];
for i = 1:numOdor
unitVector = newBlockData{i};
tempUnits = reshape(permute(unitVector,[2 1 3]),[],size(unitVector,3));
allUnits = [allUnits; tempUnits];
clusterAssignment = [clusterAssignment; i*ones(size(tempUnits,1),1)];
end
numUnits = size(allUnits,1);
tempDistMatrix = squareform(pdist(allUnits,'correlation'));
diagValue = ones(1,numUnits)*10;
fullDistMatrix = tempDistMatrix + diag(diagValue);
trialBlock = ones(timeLength)*10;
trialBlockMatrix = kron(eye(numOdor*numTrial), trialBlock);
fullDistMatrix = fullDistMatrix + trialBlockMatrix*10;
[~, idx] = sort(fullDistMatrix,2,"ascend");
colIndices = idx(:,1:KNN);
clusterTemp = zeros(size(colIndices));
for nn = 1:KNN
clusterTemp(:,nn) = clusterAssignment(colIndices(:,nn),1);
end
clusterAssignment(:,2) = mode(clusterTemp(:,1:KNN),2);
clusterAssignment(:,3) = clusterAssignment(:,2) == clusterAssignment(:,1);
perCorrectCluster = sum(clusterAssignment(:,3)) ./ size(clusterAssignment(:,1),1);
disp(['Percentage correctly clustered = ', num2str(perCorrectCluster)]);
clear resultMatrixTrialPerRow resultMatrix percentMatrix
resultMatrix = zeros(numOdor,timeLength*numTrial);
resultMatrixTrialPerRow = zeros(numOdor*numTrial,timeLength);
percentMatrixAllTrial = zeros(numOdor,numTrial,timeLength);
percentMatrix = zeros(numOdor+1,timeLength);
for i = 1:numOdor
blockIndex = find(clusterAssignment(:,1)==i);
resultMatrix(i,1:length(blockIndex)) = clusterAssignment(blockIndex,2);
resultMatrixTrialPerRow((i-1)*numTrial+1:i*numTrial,:) = reshape(resultMatrix(i,:),timeLength,numTrial)';
tempPercent = clusterAssignment(blockIndex,3);
percentMatrixTrialPerRow = reshape(tempPercent,timeLength,numTrial)';
percentMatrixAllTrial(i,:,:) = percentMatrixTrialPerRow;
percentMatrix(i,:) = mean(percentMatrixTrialPerRow,1);
end
percentMatrix(end,:) = mean(percentMatrix(1:numOdor,:),1);
if OnOffIter == 1, smooth = 3; else, smooth = 4; end
percentMatrix(1:end-1,:) = filtfilt(ones(1,smooth)/smooth,1,percentMatrix(1:end-1,:)')';
percentMatrix(end,:) = filtfilt(ones(1,smooth)/smooth,1,percentMatrix(end,:));
if OnOffIter == 1
origTimeON = linspace(0,onDuration,length(percentMatrix(end,:)));
interpCurveON = interp1(origTimeON, percentMatrix(end,:), commonTimeON,'linear');
totalPercentMatrixON(fileIter,:) = interpCurveON;
else
origTimeOFF = linspace(0,afterDuration,length(percentMatrix(end,:)));
interpCurveOFF = interp1(origTimeOFF, percentMatrix(end,:), commonTimeOFF,'linear');
totalPercentMatrixOFF(fileIter,:) = interpCurveOFF;
end
end
end
totalPercentMatrixON(end+1,:) = mean(totalPercentMatrixON, 1,'omitnan');
totalPercentMatrixOFF(end+1,:) = mean(totalPercentMatrixOFF, 1,'omitnan');
timeAxisON = linspace(0,onDuration,size(totalPercentMatrixON,2));
figure; hold on;
for r = 1:fileIter
plot(timeAxisON, totalPercentMatrixON(r,:)*100,'Color',[0.7 0.7 0.7],'LineWidth',2);
end
plot(timeAxisON, totalPercentMatrixON(end,:)*100,'Color','k','LineWidth',4);
chanceLevel = 100/numOdor; yline(chanceLevel,':','Chance','LabelHorizontalAlignment','left');
xlim([0 onDuration]); xticks(linspace(0,onDuration,5)); xticklabels(linspace(0,onDuration,5));
xlabel('Mouse ON, time after odor onset (s)'); ylabel('Classification rate (%)'); title('ON Prediction (n = 8)'); axis square;
timeAxisOFF = linspace(0,afterDuration,size(totalPercentMatrixOFF,2));
figure; hold on;
for r = 1:fileIter
plot(timeAxisOFF, totalPercentMatrixOFF(r,:)*100,'Color',[0.7 0.7 0.7],'LineWidth',2);
end
plot(timeAxisOFF, totalPercentMatrixOFF(end,:)*100,'Color','k','LineWidth',4);
chanceLevel = 100/numOdor; yline(chanceLevel,':','Chance','LabelHorizontalAlignment','left');
xlim([0 afterDuration]); xticks(linspace(0,afterDuration,9)); xticklabels(linspace(0,afterDuration,9));
xlabel('Mouse ITI, time after odor offset (s)'); ylabel('Classification rate (%)'); title('ITI Prediction (n = 8)'); axis square;
%% Fig6c: PCA analysis and ITI classification
close all; clc; clear
fileName = 'm149_3_1_dataset\';
load(strcat(fileName, '\filtData3.mat'));
odorAbbrev = {'met', 'ace', 'als', 'eug'};
odorOrder = [2 1 4 3];
solBlocks = 4:7;
nOdors = 4;
nTrials = 10;
nGlom = size(blockData.block4,3);
onDuration = 5; preDuration = 2;fps = 16;
trials = 1:nTrials;
nbins = nTrials;
pcDim = 1:3;
offDuration = 50;
timeBins = (onDuration+preDuration)*fps+1:fps*(onDuration+preDuration+offDuration);
[colorMap] = ColorScheme(nOdors);
medRsp = [];
for oo = 1:nOdors
tempName = strcat('filtData.block', num2str(oo));
currentOdor = eval(tempName);
temp = squeeze(mean(currentOdor(trials, timeBins,:),2)); %single trial
medRsp = [medRsp; temp];
end
allData = [medRsp];
[~,data_pca,latent] = pca(allData);
[v] = latent(pcDim)/sum(latent);
v1 = round(v(1)*10000)/100;
v2 = round(v(2)*10000)/100;
v3 = round(v(3)*10000)/100;
st1 = strcat('PC', string(pcDim(1)), '(', num2str(v1),'%)');
st2 = strcat('PC', string(pcDim(2)), '(', num2str(v2),'%)');
st3 = strcat('PC', string(pcDim(3)), '(', num2str(v3),'%)');
fff = figure(1);
set(fff, 'units','normalized','outerposition',[0 0 0.4 0.5])
hold on; grid on;
xlabel(st1,'fontsize', 12);
ylabel(st2,'fontsize', 12);
zlabel(st3,'fontsize', 12);
for t = 1:nOdors
data2 = data_pca(((t-1)*nTrials+1):(nTrials*t), pcDim);
for tt = trials
if tt == nTrials
plot3(data2(tt,1), data2(tt,2), data2(tt,3),...
'o', 'Color', colorMap(t,:),...
'MarkerFaceColor', colorMap(t,:),...
'LineWidth', 2, 'DisplayName',strcat('ITI: ',odorAbbrev{t}),...
'LineStyle', 'none', 'MarkerSize',8);
else
plot3(data2(tt,1), data2(tt,2), data2(tt,3),...
'o', 'Color', colorMap(t,:),...
'MarkerFaceColor', colorMap(t,:),...
'LineWidth', 2, 'HandleVisibility','off',...
'LineStyle', 'none', 'MarkerSize',8);
end
hold on;
end
end
legend('Location','eastoutside');
title({'Filled marker = 55s ITI'});
axis square;
%% Fig6d kNN leave-one-out classification for ITI spont activity
clc; close all;
k = 20;
labels = zeros(1, nOdors * nTrials);
nTrials = 10;
nPts = zeros(1, nOdors);
labeltemp = cell(1, nOdors);
for oo = 1:nOdors
labels((oo-1)*nTrials+1:oo*nTrials) = oo;
nPts(oo) = nTrials;
labeltemp{oo} = strcat('ITI:', odorAbbrev{oo});
end
cosDist = squareform(pdist(data_pca(:,1:3),'cosine'));
nPtsTot = size(data_pca,1);
classAccur = [];
pointAccur = [];
classifClass = zeros(nOdors,nOdors);
for dd = 1:nPtsTot
[~, indd] = sort(cosDist(dd,:));
closePt = indd(2:k+1); %k-number closest points
closeLabel(dd,1:k) = labels(closePt); %label of closest points
predictedLabel(dd) = mode(closeLabel(dd,1:k));
queryLabel = labels(dd); %label of query pt
compPt = (queryLabel == closeLabel(dd,:)); %accuracy
pointAccur(dd) = sum(compPt)/length(compPt); %normalized accuracy
classifClass(queryLabel,predictedLabel(dd)) = classifClass(queryLabel,predictedLabel(dd))+1;
end
normclassifClass = classifClass./nPts; % normalized classification
for dd = 1:nOdors
classAccur(dd) = mean(pointAccur((oo-1)*nPts+1:oo*nPts));
end
classRate = diag(normclassifClass)*100;
classRateITI = mean(classRate(1:nOdors));
% Confusion matrix for each increment
figure;
imagesc(normclassifClass); hold on;
cc = colorbar; cc.Label.String = 'P(Classification)';
% colormap(flipud(gray))
caxis([0 1])
cc.Ticks = 0:0.5:1;
set(gca, 'XTick',1:4,'xticklabel',labeltemp,'YTick',1:4,'yticklabel',labeltemp)
ylabel('Actual'); xlabel('Predicted')
axis square
title({strcat('k-NN Classification: k=',num2str(k)),...
strcat('ITI Classification Rate:',num2str(classRateITI),'%')})
%% Fig6d_control kNN leave-one-out classification with shuffled data label
clc; close all;
repeat = 1000;
repeatNormClassifClass = zeros(nOdors,nOdors);
repeatClassAccur = zeros(repeat,1);
repeatClassRate = 0;
for ri = 1 : repeat
closeLabel_shuf = zeros(nPtsTot, k);
predictedLabel_shuf = zeros(1, nPtsTot);
pointAccur_shuf = zeros(1, nPtsTot);
classifClass_shuf = zeros(nOdors, nOdors);
shufLabels = labels(randperm(numel(labels))); % shuffled training labels
for dd = 1:nPtsTot
[~, indd] = sort(cosDist(dd,:));
closePt = indd(2:k+1);
% neighbors use SHUFFLED labels
closeLabel_shuf(dd,1:k) = shufLabels(closePt);
predictedLabel_shuf(dd) = mode(closeLabel_shuf(dd,1:k));
queryLabel = labels(dd); % still true label
compPt = (queryLabel == closeLabel_shuf(dd,:));
pointAccur_shuf(dd) = mean(compPt);
classifClass_shuf(queryLabel, predictedLabel_shuf(dd)) = ...
classifClass_shuf(queryLabel, predictedLabel_shuf(dd)) + 1;
end
normclassifClass_shuf = classifClass_shuf ./ nTrials;
repeatNormClassifClass = repeatNormClassifClass + normclassifClass_shuf;
classRate_shuf = diag(normclassifClass_shuf) * 100;
classRateITI_shuf = mean(classRate_shuf);
repeatClassAccur(ri) = classRateITI_shuf;
end
repeatNormClassifClass = repeatNormClassifClass ./ repeat;
repeatClassRate = mean(repeatClassAccur);
repeatClassStd = std(repeatClassAccur);
fprintf('Repeat %d times with Mean = %.2f%% and STD = %.2f \n', repeat, repeatClassRate , repeatClassStd);
figure;
imagesc(normclassifClass_shuf); hold on;
cc = colorbar; cc.Label.String = 'P(Classification)';
caxis([0 1])
cc.Ticks = 0:0.5:1;
set(gca, 'XTick',1:nOdors,'xticklabel',labeltemp, ...
'YTick',1:nOdors,'yticklabel',labeltemp);
ylabel('Actual'); xlabel('Predicted');
axis square
title({sprintf('k-NN Shuffled Classification: k = %d', k), ...
sprintf('ITI Shuffled Classification Rate (REAL): %.2f %%', classRateITI_shuf)});
figure; imagesc(repeatNormClassifClass); hold on;
cc = colorbar; cc.Label.String = 'P(Classification)';
caxis([0 1])
cc.Ticks = 0:0.5:1;
set(gca, 'XTick',1:nOdors,'xticklabel',labeltemp,'YTick',1:nOdors,'yticklabel',labeltemp)
ylabel('Actual'); xlabel('Predicted')
axis square
title({'Fig 6d-control', sprintf('Mean classification rate: %.2f%% (n = %d)', repeatClassRate, repeat)});
%% ============= Fig6 a/b/h: accuracy and stats comparing datasets ===========================
clear;
% SETTINGS
% datasetDir = 'm129_1_2_dataset'; % Uncomment for Fig6h
datasetDir = 'm149_3_1_dataset'; % Uncomment for Fig6ab
KNN = 10;
numTrial = 10;
nPerm = 1000;
qFDR = 0.05;
useFDR = true;
doMaxT = true;
% Load data
S = load(fullfile(datasetDir,'filtData2.mat'));
filtData = S.filtData;
fps = S.fps; preDuration = S.preDuration; onDuration = S.onDuration; afterDuration = S.afterDuration;
names = fieldnames(filtData);
odorFields = names;
numOdor = numel(odorFields);
stats = struct([]);
for OnOffIter = 1:2
% Time window + binning
if OnOffIter == 1
range = preDuration*fps+1 : (preDuration+onDuration)*fps;
numBinAvg = 1; smoothN = 2; durSec = onDuration;
plotTitle = sprintf('ON Prediction | %s', datasetDir);
xlab = 'Mouse ON, time after odor onset (s)';
else
range = (preDuration+onDuration)*fps+1 : (afterDuration+preDuration+onDuration)*fps;
numBinAvg = 4; smoothN = 4; durSec = afterDuration;
plotTitle = sprintf('ITI Prediction | %s', datasetDir);
xlab = 'Mouse ITI, time after odor offset (s)';
end
% Build newBlockData (odor -> trials x timeBins x features)
newBlockData = cell(numOdor,1);
for i = 1:numOdor
tempData = filtData.(odorFields{i});
tempFiltData = tempData(1:numTrial, range, :);
iter = ceil(length(range)/numBinAvg);
tempFiltData2 = zeros(numTrial, iter, size(tempFiltData,3));
for ii = 1:numBinAvg:length(range)
b = (ii-1)/numBinAvg + 1;
try
tempFiltData2(:,b,:) = mean(tempFiltData(:,ii:ii+numBinAvg-1,:),2);
catch
tempFiltData2(:,b,:) = mean(tempFiltData(:,ii:end,:),2);
end
end
newBlockData{i} = tempFiltData2;
end
timeLength = size(newBlockData{1},2);
% Reshape into allUnits + trueLabels
allUnits = [];
trueLabels = [];
for i = 1:numOdor
unitVector = newBlockData{i};
tempUnits = reshape(permute(unitVector,[2 1 3]), [], size(unitVector,3));
allUnits = [allUnits; tempUnits];
trueLabels = [trueLabels; i*ones(size(tempUnits,1),1)];
end
numUnits = size(allUnits,1);
numGlom = size(allUnits,2)
rowStd = std(allUnits,0,2);
zidx = (rowStd == 0);
if any(zidx)
allUnits(zidx,:) = allUnits(zidx,:) + 1e-12*randn(sum(zidx), size(allUnits,2));
end
% Distance matrix + leave-trial-out blocks
tempDistMatrix = squareform(pdist(allUnits,'correlation'));
tempDistMatrix(isnan(tempDistMatrix)) = 1;
diagValue = ones(1,numUnits)*10;
fullDistMatrix = tempDistMatrix + diag(diagValue);
trialBlock = ones(timeLength)*10;
trialBlockMatrix = kron(eye(numOdor*numTrial), trialBlock);
fullDistMatrix = fullDistMatrix + trialBlockMatrix*10;
[~, idx] = sort(fullDistMatrix,2,'ascend');
colIndices = idx(:,1:KNN);
% Real decoding
[real_black, real_perOdor] = compute_percent_per_odor( ...
trueLabels, colIndices, trueLabels, numOdor, timeLength, numTrial, smoothN);
% Shuffle null distribution
shuf_black = zeros(nPerm, timeLength);
for pi = 1:nPerm
shufLabels = trueLabels(randperm(numUnits));
shuf_black(pi,:) = compute_percent_per_odor( ...
shufLabels, colIndices, trueLabels, numOdor, timeLength, numTrial, smoothN);
end
% p-values per time bin + FDR / maxT
p_point = (sum(shuf_black >= real_black,1) + 1) / (nPerm + 1);
if useFDR
sig_fdr = mafdr(p_point,'BHFDR',true) < qFDR;
else
sig_fdr = p_point < 0.05/numel(p_point);
end
p_maxT = nan(size(p_point));
sig_maxT = false(size(p_point));
if doMaxT
shuf_mu = mean(shuf_black,1);
shuf_sd = std(shuf_black,0,1) + eps;
real_z = (real_black - shuf_mu) ./ shuf_sd;
maxZ = zeros(nPerm,1);
for pi = 1:nPerm
z_perm = (shuf_black(pi,:) - shuf_mu) ./ shuf_sd;
maxZ(pi) = max(z_perm);
end
p_maxT = (sum(maxZ >= real_z,1) + 1) / (nPerm + 1);
sig_maxT = p_maxT < 0.05;
end
% Plot
timeAxis = linspace(0, durSec, timeLength);
figure; hold on;
lo = prctile(shuf_black,2.5,1);
hi = prctile(shuf_black,97.5,1);
fill([timeAxis fliplr(timeAxis)], [lo*100 fliplr(hi*100)], ...
[0.7 0.7 0.7], 'FaceAlpha', 0.25, 'EdgeColor', 'none');
plot(timeAxis, mean(shuf_black,1)*100, 'Color', [0.5 0.5 0.5], 'LineWidth', 2);
odorColors = ColorScheme(numOdor,0.25);
for oo = 1:numOdor
plot(timeAxis, real_perOdor(oo,:)*100, 'LineWidth', 2, 'Color', odorColors(oo,:));
end
plot(timeAxis, real_black*100, 'k', 'LineWidth', 5);
yline(100/numOdor, ':', 'Chance', 'LabelHorizontalAlignment','left');
xlabel(xlab); ylabel('Classification rate (%)');
title(sprintf('%s | nPerm=%d', plotTitle, nPerm));
axis square; box on; xlim([0 durSec]);
p_for_bar = p_point;
p_for_bar(~sig_fdr) = 1;
yl = ylim;
yBar = yl(1) + 0.05*(yl(2)-yl(1));
plot_sig_bars(timeAxis, p_for_bar, yBar, 8);
fprintf('\n=== %s ===\n', plotTitle);
fprintf('Min pointwise p: %.4g\n', min(p_point));
fprintf('FDR(q=%.3f) sig bins: %d/%d\n', qFDR, sum(sig_fdr), numel(sig_fdr));
if doMaxT
fprintf('MaxT(FWER) sig bins: %d/%d | min p_maxT=%.4g\n', ...
sum(sig_maxT), numel(sig_maxT), min(p_maxT));
end
% store
stats(OnOffIter).timeAxis = timeAxis;
stats(OnOffIter).real_perOdor = real_perOdor;
stats(OnOffIter).real_black = real_black;
stats(OnOffIter).shuf_black = shuf_black;
stats(OnOffIter).p_point = p_point;
stats(OnOffIter).sig_fdr = sig_fdr;
stats(OnOffIter).p_maxT = p_maxT;
stats(OnOffIter).sig_maxT = sig_maxT;
stats(OnOffIter).odorFields = odorFields;
end
% ============= Fig6 e/f/i: accuracy and stats comparing datasets ===========================
clear;
% % Fig6ef list:
% fileNameKNN = {
% 'm149_1_2_dataset';
% 'm148_2_1_dataset';
% 'm141_8_1_dataset';
% 'm141_1_1_dataset';
% 'm149_3_1_dataset';
% };
% Fig6i list:
fileNameKNN = {
'm122_4_4_dataset';
'm125_1_1_dataset';
'm126_1_1_dataset';
'm134_6_1_dataset';
'm129_1_2_dataset';
'm129_4_2_dataset';
'm129_3_1_dataset';
'm1716_1_1_dataset';
};
OnOffIter = 2; % 1=ON, 2=ITI/OFF, 3=ITI/OFF random structure
KNN = 10;
numTrial = 10;
binSizePts = 5;
qFDR = 0.05;
sessColor = [0.55 0.80 0.82];
semColor = [0.70 0.70 0.70];
meanColor = [0 0 0];
chanceColor = [0.55 0.55 0.55];
pThresh = [0.05, 0.01, 0.001];
pColors = {[0 0 0]; [0.85 0 0]; [0 0 0.85]};
% Time for interpolation
if OnOffIter == 1
commonTime = linspace(0,5,200);
xlab = 'Time (s)';
elseif OnOffIter == 2
commonTime = linspace(0,50,500);
xlab = 'Time after odor offset (s)';
else
commonTime = linspace(0,17,500);
xlab = 'Time after odor offset (s)';
end
% means across odors by session
nSess = numel(fileNameKNN);
nTime = numel(commonTime);
accSess = nan(nSess, nTime);
chanceFixed = [];
for s = 1:nSess
datasetDir = fileNameKNN{s};
fprintf('Session %d/%d: %s\n', s, nSess, datasetDir);
% Load + parse
S = load(fullfile(datasetDir,'filtData2.mat'));
filtData = S.filtData;
fps = S.fps; preDuration = S.preDuration; onDuration = S.onDuration; afterDuration = S.afterDuration;
names = fieldnames(filtData);
odorFields = names;
numOdor = numel(odorFields);
if isempty(chanceFixed), chanceFixed = 1/numOdor; end
% Window + binning
if OnOffIter == 1
range = preDuration*fps+1 : (preDuration+onDuration)*fps;
numBinAvg = 1; smoothN = 2; durSec = onDuration;
else
range = (preDuration+onDuration)*fps+1 : (afterDuration+preDuration+onDuration)*fps;
numBinAvg = 4; smoothN = 4; durSec = afterDuration;
end
[timeAxis_s, meanTrace_s] = decode_timecurve( ...
filtData, odorFields, range, fps, numTrial, numBinAvg, KNN, smoothN, durSec);
accSess(s,:) = interp1(timeAxis_s, meanTrace_s, commonTime, 'linear', 'extrap');
end
% STATS: one-sided t-test vs chance + BH-FDR
[xStat, accStat] = bin_timecourses(commonTime, accSess, binSizePts);
nBins = numel(xStat);
p_raw = nan(nBins,1);
for b = 1:nBins
y = accStat(:,b);
y = y(~isnan(y));
if numel(y) >= 2
[~, p_raw(b)] = ttest(y, chanceFixed, 'Tail','right');
end
end
sig_fdr = bh_fdr_mask(p_raw, qFDR);
% Plot
figure; hold on;
for s = 1:nSess
plot(commonTime, accSess(s,:)*100, 'Color', sessColor, 'LineWidth', 2);
end
meanAcc = mean(accSess, 1, 'omitnan');
semAcc = std(accSess, 0, 1, 'omitnan') ./ sqrt(sum(~isnan(accSess),1));
fill([commonTime fliplr(commonTime)], ...
[(meanAcc-semAcc)*100 fliplr((meanAcc+semAcc)*100)], ...
semColor, 'FaceAlpha', 0.6, 'EdgeColor','none');
plot(commonTime, meanAcc*100, 'Color', meanColor, 'LineWidth', 3.5);
yline(chanceFixed*100, '--', 'Color', chanceColor, 'LineWidth', 1.3);
xlabel(xlab);
ylabel('Classification rate (%)');
ylim([0 100]);
xlim([commonTime(1) commonTime(end)]);
box off;
yl = ylim;
yBar = yl(1) + 0.03*diff(yl);
plot_sig_dots(xStat, p_raw, sig_fdr, yBar, pThresh, pColors, 7.5);
% Functions for Figure 6 A, B, H
function [blackLine, perOdor] = compute_percent_per_odor(labelVec, colIndices, trueLabels, ...
numOdor, timeLength, numTrial, smoothN)
numUnits = numel(trueLabels);
KNN = size(colIndices,2);
clusterTemp = zeros(numUnits, KNN);
for nn = 1:KNN
clusterTemp(:,nn) = labelVec(colIndices(:,nn));
end
predLabel = mode(clusterTemp,2);
correct = (predLabel == trueLabels);
perOdor = zeros(numOdor, timeLength);
for oo = 1:numOdor
idx = find(trueLabels == oo);
tmpMat = reshape(correct(idx), timeLength, numTrial)';
perOdor(oo,:) = mean(tmpMat,1);
end
if smoothN > 1
for oo = 1:numOdor
perOdor(oo,:) = filtfilt(ones(1,smoothN)/smoothN, 1, perOdor(oo,:));
end
end
blackLine = mean(perOdor,1);
end
function plot_sig_bars(timeAxis, pvals, y, lw)
if nargin < 4, lw = 6; end
thr = [0.05, 0.01, 0.001];
cols = {[0 0 0], [1 0 0], [0 0 1]};
for k = 1:numel(thr)
sig = pvals < thr(k);
d = diff([false sig false]);
starts = find(d==1);
ends = find(d==-1) - 1;
for s = 1:numel(starts)
i1 = starts(s); i2 = ends(s);
line([timeAxis(i1) timeAxis(i2)], [y y], ...
'Color', cols{k}, 'LineWidth', lw, 'Clipping', 'on');
end
end
end
% Functions for Figure 6 E, F, I
function [timeAxis, meanTrace] = decode_timecurve( ...
filtData, odorFields, range, fps, numTrial, numBinAvg, KNN, smoothN, durSec)
numOdor = numel(odorFields);
% bin within range for each odor: trials x timeBins x features
blockData = cell(numOdor,1);
for i = 1:numOdor
tempData = filtData.(odorFields{i});
temp = tempData(1:numTrial, range, :);
iter = ceil(length(range)/numBinAvg);
tempBinned = zeros(numTrial, iter, size(temp,3));
for ii = 1:numBinAvg:length(range)
b = (ii-1)/numBinAvg + 1;
try
tempBinned(:,b,:) = mean(temp(:,ii:ii+numBinAvg-1,:),2);
catch
tempBinned(:,b,:) = mean(temp(:,ii:end,:),2);
end
end
blockData{i} = tempBinned;
end
timeLength = size(blockData{1},2);
% units + labels
allUnits = [];
trueLabels = [];
for i = 1:numOdor
X = blockData{i};
U = reshape(permute(X,[2 1 3]), [], size(X,3));
allUnits = [allUnits; U];
trueLabels = [trueLabels; i*ones(size(U,1),1)];
end
numUnits = size(allUnits,1);
rowStd = std(allUnits,0,2);
zidx = (rowStd==0);
if any(zidx)
allUnits(zidx,:) = allUnits(zidx,:) + 1e-12*randn(sum(zidx), size(allUnits,2));
end
% distance matrix
distMat = squareform(pdist(allUnits,'correlation'));
distMat(isnan(distMat)) = 1;
distMat = distMat + diag(ones(numUnits,1)*10);
% exclude within-odor same-trial neighbors
trialBlock = ones(timeLength)*10;
penalty = kron(eye(numOdor*numTrial), trialBlock);
distMat = distMat + penalty*10;
[~, idx] = sort(distMat,2,'ascend');
nnIdx = idx(:,1:KNN);
[meanTrace, ~] = compute_percent_per_odor_2( ...
trueLabels, nnIdx, trueLabels, numOdor, timeLength, numTrial, smoothN);
timeAxis = linspace(0, durSec, timeLength);
end
function [meanTrace, perOdor] = compute_percent_per_odor_2(labelVec, nnIdx, trueLabels, ...
numOdor, timeLength, numTrial, smoothN)
numUnits = numel(trueLabels);
KNN = size(nnIdx,2);
neigh = zeros(numUnits, KNN);
for nn = 1:KNN
neigh(:,nn) = labelVec(nnIdx(:,nn));
end
predLabel = mode(neigh,2);
correct = (predLabel == trueLabels);
perOdor = zeros(numOdor, timeLength);
for oo = 1:numOdor
idx = find(trueLabels == oo);
tmpMat = reshape(correct(idx), timeLength, numTrial)';
perOdor(oo,:) = mean(tmpMat,1);
end
if smoothN > 1
for oo = 1:numOdor
perOdor(oo,:) = filtfilt(ones(1,smoothN)/smoothN, 1, perOdor(oo,:));
end
end
meanTrace = mean(perOdor,1);
end
function [xBin, accBin] = bin_timecourses(x, accSess, binSizePts)
[nSess, nTime] = size(accSess);
nBins = floor(nTime/binSizePts);
xT = x(1:nBins*binSizePts);
accT = accSess(:,1:nBins*binSizePts);
accR = reshape(accT, nSess, binSizePts, nBins);
accBin = squeeze(mean(accR,2,'omitnan'));
xR = reshape(xT, binSizePts, nBins);
xBin = mean(xR,1,'omitnan')';
end
function sig = bh_fdr_mask(p, q)
sig = false(size(p));
v = ~isnan(p);
pv = p(v);
if isempty(pv), return; end
[ps, ord] = sort(pv);
m = numel(ps);
th = (1:m)'/m*q;
pass = ps <= th;
if any(pass)
kmax = find(pass,1,'last');
idxValid = find(v);
sig(idxValid(ord(1:kmax))) = true;
end
end
function plot_sig_dots(x, p_raw, sig_fdr, yBar, pThresh, pColors, markerSize)
elig = sig_fdr & ~isnan(p_raw);
for i = 1:numel(x)
if ~elig(i), continue; end
if p_raw(i) < pThresh(3)
c = pColors{3};
elseif p_raw(i) < pThresh(2)
c = pColors{2};
elseif p_raw(i) < pThresh(1)
c = pColors{1};
else
continue
end
plot(x(i), yBar, 'o', ...
'MarkerFaceColor', c, ...
'MarkerEdgeColor', c, ...
'MarkerSize', markerSize);
end
end

Script_Fig6.m, under CC-BY-4.0 · at the source

Overview

Authors: Elizabeth H. Moss¹, Doris Ling², Cameron L. Smith³, Feiyang Deng², Ryan Kroeger³, Jacob Reimer³, Baranidharan Raman², Benjamin R. Arenkiel¹

ORCID iDs: Elizabeth H. Moss

Institutions: Baylor College of Medicine (United States); Washington University in St. Louis (United States)

Journal: iScience, volume 29, issue 4, article 115373

Dates: received 28 July 2025; accepted 12 March 2026; published online 16 March 2026

Type: Research article · Language: English

License: CC BY

Identifiers: DOI 10.1016/j.isci.2026.115373 · PMID 41971993 · PMCID PMC13066794 · OpenAlex W7136786262

Open access: gold, a free copy (OpenAlex)

Status: code verified

Categories: mouse (organism), other (organism)

Methods: Spectral & time-frequency, Statistics, Smoothing, state filtering, decompositions, Machine learning, fMRI & imaging, Single-unit activity, calcium imaging

Keywords: biological sciences, neuroscience, sensory neuroscience

Topic: Olfactory and Sensory Function Studies (Sensory Systems, Neuroscience), according to OpenAlex

Funding: NSF (1724218, 2021795); ONR (N00014-19-1-2049, N00014-21-1-2343); NIH (R00DC019505, UF1NS111692); McNair Medical Institute; NINDS (R01NS078294); NIDDK (R01DK109934); DOD (PR180451-PRMP)

Citations: not cited yet (Europe PMC); 63 references in the paper

Research resources: MATLAB RRID:SCR_001622, ScanImage RRID:SCR_014307, LabView RRID:SCR_014325

Abstract

Interpreting chemical information and translating it into ethologically relevant output is a shared challenge of olfactory systems across species, but are olfactory computations conserved across species to overcome these common challenges? To investigate this, we compared neural activity in the locust antennal lobe (AL) and mouse olfactory bulb (OB) both during and after odor presentations. We found that odors activated nearly mutually exclusive neural ensembles during odor presentations (“ON response”) and after the termination of the odor stimulus (“OFF response”). ON and OFF responses evoked by a single odor were anticorrelated with each other. Inverted OFF responses persisted long after odor termination in both AL and OB, and enhanced contrast between odors that were experienced close together in time. Together our results show how post-odor neural activity, relative to odor-evoked activity, is similar across two distinct species, revealing a conserved mechanism for enhancing contrast between odors at the neural level.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above, with 1 match between paragraphs and lines of code.

figshare 31435543

License: CC-BY-4.0

State: the link answers, verified on 30 September 2026

Evidence: files inventoried

Languages: MATLAB (13)

Size: 77 files, 13 scripts

Software Heritage: not checked

Found in: “Data and code availability”

Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation

Tools: Signal Processing Toolbox (6 files), Statistics and Machine Learning Toolbox (5 files)

Availability: 1 check, the latest on 30 September 2026: the link answers (HTTP 200)

30 September 2026: the link answers (HTTP 200)

8 files

ColorScheme.m, MATLAB, 90 lines
ColorScheme2.m, MATLAB, 90 lines
Script_Fig1.m, MATLAB, 90 lines
Script_Fig2.m, MATLAB, 501 lines
Script_Fig3.m, MATLAB, 427 lines
Script_Fig4.m, MATLAB, 460 lines
Script_Fig5.m, MATLAB, 670 lines
Script_Fig6.m, MATLAB, 1,148 lines, 1 match

At the source: doi.org/10.6084/m9.figshare.31435543

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
8 scripts, each with its path and the digest of its content;
1 match between paragraphs of the paper and lines of the code (method lexical-v1);
neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data and code availability

• Locust electrophysiology and mouse calcium imaging data is publicly available at Figshare: https://doi.org/10.6084/m9.figshare.31435543. DOI is listed in the key resources table. • Original code used to analyze data and generate figures is publicly available at Figshare: https://doi.org/10.6084/m9.figshare.31435543. DOI is listed in the key resources table. • Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 30 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 8 authors, 3 keywords, 7 funders, 61 references, 3 RRIDs.

Cite

This paper

Moss, E. H., Ling, D., Smith, C. L., Deng, F., Kroeger, R., Reimer, J., Raman, B., & Arenkiel, B. R. (2026). Conserved post-odor dynamics in the olfactory systems of mice and locusts. iScience, 29(4), 115373. https://doi.org/10.1016/j.isci.2026.115373

BibTeX

@article{moss2026conserved, author = {Moss, Elizabeth H. and Ling, Doris and Smith, Cameron L. and Deng, Feiyang and Kroeger, Ryan and Reimer, Jacob and Raman, Baranidharan and Arenkiel, Benjamin R.}, title = {{Conserved post-odor dynamics in the olfactory systems of mice and locusts}}, journal = {iScience}, year = {2026}, month = mar, volume = {29}, number = {4}, pages = {115373}, publisher = {Elsevier}, issn = {2589-0042}, doi = {10.1016/j.isci.2026.115373}, url = {https://doi.org/10.1016/j.isci.2026.115373}, pmid = {41971993}, pmcid = {PMC13066794} }

RIS

TY - JOUR AU - Moss, Elizabeth H. AU - Ling, Doris AU - Smith, Cameron L. AU - Deng, Feiyang AU - Kroeger, Ryan AU - Reimer, Jacob AU - Raman, Baranidharan AU - Arenkiel, Benjamin R. TI - Conserved post-odor dynamics in the olfactory systems of mice and locusts T2 - iScience J2 - iScience PY - 2026 DA - 2026/03/16 VL - 29 IS - 4 SP - 115373 SN - 2589-0042 PB - Elsevier DO - 10.1016/j.isci.2026.115373 UR - https://doi.org/10.1016/j.isci.2026.115373 LA - en ER -

CSL-JSON

{ "id": "10.1016/j.isci.2026.115373", "type": "article-journal", "title": "Conserved post-odor dynamics in the olfactory systems of mice and locusts", "container-title": "iScience", "author": [ { "family": "Moss", "given": "Elizabeth H." }, { "family": "Ling", "given": "Doris" }, { "family": "Smith", "given": "Cameron L." }, { "family": "Deng", "given": "Feiyang" }, { "family": "Kroeger", "given": "Ryan" }, { "family": "Reimer", "given": "Jacob" }, { "family": "Raman", "given": "Baranidharan" }, { "family": "Arenkiel", "given": "Benjamin R." } ], "container-title-short": "iScience", "volume": "29", "issue": "4", "page": "115373", "DOI": "10.1016/j.isci.2026.115373", "PMID": "41971993", "PMCID": "PMC13066794", "ISSN": "2589-0042", "publisher": "Elsevier", "URL": "https://doi.org/10.1016/j.isci.2026.115373", "language": "en", "issued": { "date-parts": [ [ 2026, 3, 16 ] ] } }

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.7554/elife.107905 [code]: Adult-neurogenesis allows for representational stability and flexibility in early olfactory system.
Journal: eLife
In common: 7 references
[2] doi:10.1038/s41592-026-03023-y: Isotonic and minimally invasive optical clearing media for live cell imaging ex vivo and in vivo.
Journal: Nature methods
In common: mouse, 5 references
[3] doi:10.1126/sciadv.aee1002 [code]: Theta oscillations are an organizational unit of odor processing in the olfactory bulb.
Journal: Science advances
In common: Signal Processing Toolbox, Statistics and Machine Learning Toolbox, 2 references
[4] doi:10.1038/s41467-026-70356-9: A topographical organization in the primary olfactory cortex.
Journal: Nature communications
In common: mouse, 3 references
[5] doi:10.1016/j.isci.2026.115897 [code]: Experience and behavior modulate piriform cortex odor representation in freely moving mice.
Journal: iScience
In common: Signal Processing Toolbox, Statistics and Machine Learning Toolbox, mouse, 1 reference
[6] doi:10.1002/cne.70197: Anatomical Investigation Reveals a Second Olfactory System in Locusts.
Journal: The Journal of comparative neurology
In common: other, 2 references
[7] doi:10.1038/s41467-026-71742-z [code]: Hypercapnia dissociates neuronal and hemodynamic responses impairing neurovascular coupling and functional brain connectivity.
Journal: Nature communications
In common: Signal Processing Toolbox, Statistics and Machine Learning Toolbox, mouse, 1 reference
[8] doi:10.7554/elife.110685 [code]: Sensory adaptation and pupil-linked arousal support flexible evidence accumulation during perceptual decision making.
Journal: eLife
In common: Signal Processing Toolbox, Statistics and Machine Learning Toolbox, 1 reference
[9] doi:10.1038/s41467-026-71356-5 [code]: Generalization of fear learning is shaped by inhibitory sensory processing in mice.
Journal: Nature communications
In common: mouse, 2 references
[10] doi:10.1126/sciadv.aeg3535 [code]: Undoing of firing rate adaptation enables invariant population codes.
Journal: Science advances
In common: 2 references

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.