function handles = dsPlotFR2(data,varargin)
%PLOTFR2 - plot spike rates in various ways depending on what data was provided.
%
% As with dsPlotFR, but has additional options for controlling output of SimStudy mode.
%
% Usage:
% dsPlotFR2(data,'option',value)
%
% Inputs:
% - data: DynaSim data structure (see dsCheckData)
% - options: (same as dsCalcFR)
% 'variable' : name of field containing data on which to calculate firing
% rates (default: *_spikes or first variable in data.labels)
% 'threshold': scalar threshold value for detecting events (default: 0)
% 'bin_size' : size of temporal window over which to calculate rate [ms or
% fraction of data set] (default: 5% of the data set)
% 'bin_shift': how much to shift the bin before calculating rate again [ms
% or fraction of data set] (default: 1% of the data set)
% 'plot_type': options for sim study mode. Options include 'heatmap',
% 'heatmap_sorted' (default), 'meanFR,' 'meanFRdens', and
% 'summary'.
% 'lock_gca' : Plots within currently active axis (gca); doesn't
% open new figures or subplots.
%
%
% Examples:
% dsPlotFR(data,'bin_size',30,'bin_shift',10);
%
% TODO: add rastergrams
%
% See also: dsCalcFR, dsSimulate, dsCheckData
data=dsCheckData(data, varargin{:});
fields=fieldnames(data);
handles=[];
options=dsCheckOptions(varargin,{...
'plot_type','heatmap_sorted',{'heatmap','heatmap_sorted','meanFR','meanFRdens','summary'},...
'variable','',[],...
'threshold',1e-5,[],... % slightly above zero in case variable is point process *_spikes {0,1}
'bin_size',.05,[],... % 30
'bin_shift',.01,[],... % 10
'exclude_data_flag',0,{0,1},...
'lock_gca',false,[true,false],...
'visible','on',[],...
},false);
lock_gca = options.lock_gca;
keyvals=dsOptions2Keyval(rmfield(options,{'plot_type'}));
% calc firing rates if not already present in data
if all(cellfun(@isempty,regexp(fields,'.*_FR$')))
data=dsCalcFR(data,keyvals{:}); % equivalent: data=dsAnalyzeStudy(data,@dsCalcFR);
fields=fieldnames(data);
end
% get list of fields with firing rate data
FR_fields=fields(~cellfun(@isempty,regexp(fields,'.*_FR$')));
FR_fields=setdiff(FR_fields,'time_FR');
% store time bins for firing rate data
time=data.time_FR;
nsets=length(FR_fields);
% plot firing rates
if (numel(data)==1 || ~isfield(data,'varied')) && ~lock_gca
% plot separate firing rates vs time (one figure per element of data)
for i=1:length(data)
plotFR_SingleSim(i);
end
else
% data contains results from a set of simulations varying something
% plot average firing rates vs whatever was varied
if strcmp(options.plot_type,'summary') && ~lock_gca
plotFR_SimStudy;
else
if numel(data) > 1 && lock_gca
error('Cannot lock gca if number of elements in data is greater than 1');
end
plotFR_SimStudy_specialized;
end
end
% NESTED FUNCTIONS
function plotFR_SingleSim(i)
% purpose: plot each data set data.(FR_fields{k})
% plots for N populations (FR data sets) from this simulation
ht=320; % height per subplot row (=per population or FR data set)
if ~lock_gca; handles(end+1)=figure('position',[250 max(50,600-(nsets-1)*ht) 1400 min(ht*nsets,750)],'visible',options.visible); end
for k=1:nsets % index of firing rate data field
dat=data(i).(FR_fields{k});
bins=0:1.05*max(dat(:));
rlims=[min(bins) max(bins)];
if rlims(1)==rlims(2), rlims=rlims(1)+[0 1e-5]; end
tlims=[min(time) max(time)];
% get population name from field (assumes: pop_*)
popname=regexp(FR_fields{k},'^([a-zA-Z0-9]+)_','tokens','once');
popname=popname{1};
ncells=size(dat,2);
if ncells==1
nc=4; % number of plots for this FR data set
% 1.0 plot firing rate trace
subplot(nsets,nc,(1:nc-1)+(k-1)*nc); % plot(t,FR)
plot(time,dat(:,1),'o-','linewidth',2); ylim(rlims); xlim(tlims);
title([popname ': firing rate over time']);
xlabel('time (ms)'); ylabel([popname ': firing rate (Hz)']);
% 2.0 plot firing rate density
subplot(nsets,nc,nc+(k-1)*nc); % hist(FR)
if numel(bins)>1
H=hist(dat(:,1),bins)/length(dat(:,1));
h=bar(bins,H); hold on
set(get(h,'children'),'FaceAlpha',0,'EdgeAlpha',.4,'linewidth',2);
rn=ksr(bins,H,.75,length(bins));
plot(bins,rn.f,'linewidth',2); title([popname ': firing rate density']);
xlabel([popname ': firing rate (Hz)']); ylabel('fraction');
xlim(rlims); ylim([0 1]);
end
else
nc=4; % number of plots for this FR data set
% 1.0 plot firing rate heat map
subplot(nsets,nc,1+(k-1)*nc); % imagesc(t,cells,FR)
imagesc(time,1:ncells,dat'); axis xy
title([popname ': firing rates (Hz)']);
xlabel('time (ms)'); ylabel([popname ' cell index']);
caxis(rlims); xlim(tlims);
if ncells<=10
ytick=1:ncells;
else
ytick=round(linspace(1,ncells,5));
if ~ismember(ncells,ytick)
ytick=[ytick ncells];
end
end
set(gca,'ytick',ytick,'yticklabel',ytick);
% 2.0 plot sorted firing rate heat map
subplot(nsets,nc,2+(k-1)*nc); % imagesc(t,cells_sorted,FR)
tmp=sum(dat,1);
[tmp,inds]=sort(tmp);
imagesc(time,1:ncells,dat(:,inds)'); axis xy
caxis(rlims); xlim(tlims);
title([popname ': firing rates (Hz)']);
xlabel('time (ms)'); ylabel([popname ' cell index (sorted by FR)']);
if ncells<=10
ytick=1:ncells;
yticklabel=ytick(inds);
else
ytick=[];
yticklabel=[];
end
set(gca,'ytick',ytick,'yticklabel',yticklabel);
% 3.0 plot population-average firing rate trace
subplot(nsets,nc,3+(k-1)*nc); % plot(t,<FR|pop>)
plot(time,mean(dat,2),'o-','linewidth',2);
title([popname ': population average firing rate']);
xlabel('time (ms)'); ylabel([popname ': avg firing rate (Hz)']);
ylim(rlims); xlim(tlims);
% 4.0 plot firing rate density
if numel(bins)>1
subplot(nsets,nc,4+(k-1)*nc); % hist(FR)
H=hist(dat(:),bins)/numel(dat);
h=bar(bins,H); hold on
set(get(h,'children'),'FaceAlpha',0,'EdgeAlpha',.4,'linewidth',2);
rn=ksr(bins,H,.75,length(bins));
plot(bins,rn.f,'linewidth',2); title([popname ': firing rate density']);
xlabel([popname ': firing rate (Hz)']); ylabel('fraction');
xlim(rlims); ylim([0 1]);
end
end
end
end
function plotFR_SimStudy
% calculate average firing rates across population and time
varied=data(1).varied;
% eliminate parameters that had the same value in all sims or were non-numeric
keep=zeros(size(varied));
for v=1:length(varied)
if isnumeric(data(1).(varied{v})) && length(unique([data.(varied{v})]))>1
keep(v)=1;
end
end
varied=varied(keep==1);
nvaried=length(varied); % number of model components varied across simulations
nsims=length(data);
FRmu=zeros(nsims,nsets);
for i=1:nsims % simulations
for k=1:nsets % populations
FRmu(i,k)=mean(data(i).(FR_fields{k})(:));
end
end
% collect info on parameters varied
params=zeros(nsims,nvaried);
for j=1:nvaried
if isnumeric(data(1).(varied{j}))
params(:,j)=[data.(varied{j})];
else
% todo: handle sims varying non-numeric model components
% (eg, mechanisms)
end
end
% plots for N populations and M varied elements
% plot how avg firing rate for each pop varies with each parameter
ht=320; % height per subplot row (=per population or FR data set)
wt=500;
if ~lock_gca; handles(end+1)=figure('position',[250 max(50,600-(nsets-1)*ht) min(1500,500+(nvaried-1)*wt) min(ht*nsets,750)],'visible',options.visible); end
cnt=0;
for k=1:nsets % populations
popname=regexp(FR_fields{k},'^([a-zA-Z0-9]+)_','tokens','once');
popname=popname{1};
ncells=size(data(1).(FR_fields{k}),2);
for j=1:nvaried % varied parameters
% calculate mean avg FRs for each value of this varied parameter
pvals=unique(params(:,j)); % unique values for this parameter j
nv=length(pvals); % number of values used for this parameter j
rvals=zeros(nv,1);
for v=1:nv
idx=(params(:,j)==pvals(v)); % sims with param j = value v
% average across all sims with param j set to value v
rvals(v)=mean(FRmu(idx,k)); % avg pop k FR given this value
end
npoints=length(find(idx)); % number of points in this average
% plot avg FRs
cnt=cnt+1; % subplot index
subplot(nsets,nvaried,cnt);
plot(pvals,rvals,'o-','linewidth',2);
% todo: add confidence intervals / error bars
axis tight; %ylim([0 max(FRmu(:))]);
xlabel(strrep(varied{j},'_','\_'));
ylabel([popname ': avg firing rate (Hz)']);
title([popname ': firing rate averaged over pop, time, sims']);
% add text: ncells, length(time), npoints
xpos=min(xlim)+.7*diff(xlim);%.05*diff(xlim);
ypos=min(ylim)+.2*diff(ylim);%.85*diff(ylim);
text(xpos,ypos,sprintf('per point:\n #cells=%g\n #bins=%g\n #sims=%g',ncells,length(time),npoints));
end
end
if length(varied)==2
% plots for N populations and 2 varied elements
% organize and imagesc FRmu(param 1, param 2)
if ~lock_gca; handles(end+1)=figure('position',[1150 max(50,600-(nsets-1)*ht) 500 min(ht*nsets,750)],'visible',options.visible);end
pvals1=unique(params(:,1)); nv1=length(pvals1);
pvals2=unique(params(:,2)); nv2=length(pvals2);
for k=1:nsets
popname=regexp(FR_fields{k},'^([a-zA-Z0-9]+)_','tokens','once');
popname=popname{1};
FRmu2=zeros(nv1,nv2);
for i=1:nv1
for j=1:nv2
idx=(params(:,1)==pvals1(i))&(params(:,2)==pvals2(j));
FRmu2(i,j)=mean(FRmu(idx,k));
end
end
% plot imagesc(param1,param2,FRmu)
subplot(nsets,1,k);
imagesc(pvals1,pvals2,FRmu2'); axis xy
xlabel(strrep(varied{1},'_','\_')); ylabel(strrep(varied{2},'_','\_'));
title([popname ': avg firing rate (Hz)']);
caxis([0 max(FRmu(:))]); colorbar
end
end
end
function plotFR_SimStudy_specialized % Specifc plot for all sims, based on options.plot_type
%%
% New code (imported from dsPlot)
num_sims=length(data); % number of simulations
% make subplot adjustments for varied parameters
if num_sims>1 && isfield(data,'varied')
% collect info on parameters varied
varied=data(1).varied;
num_varied=length(varied); % number of model components varied across simulations
num_sims=length(data); % number of data sets (one per simulation)
% collect info on parameters varied
param_mat=zeros(num_sims,num_varied); % values for each simulation
param_cell=cell(1,num_varied); % unique values for each parameter
% loop over varied components and collect values
for j=1:num_varied
if isnumeric(data(1).(varied{j}))
param_mat(:,j)=[data.(varied{j})]; % values for each simulation
param_cell{j}=unique([data.(varied{j})]); % unique values for each parameter
else
% todo: handle sims varying non-numeric model components
% (eg, mechanisms) (also in dsPlotFR and dsSelect)
end
end
param_size=cellfun(@length,param_cell); % number of unique values for each parameter
% varied parameter with most elements goes along the rows (everything else goes along columns)
row_param_index=find(param_size==max(param_size),1,'first');
row_param_name=varied{row_param_index};
row_param_values=param_cell{row_param_index};
num_rows=length(row_param_values);
%num_cols=num_sims/num_rows;
num_figs=1;
% collect sims for each value of the row parameter
indices={};
for row=1:num_rows
indices{row}=find(param_mat(:,row_param_index)==row_param_values(row));
end
num_per_row=cellfun(@length,indices);
num_cols=max(num_per_row);
sim_indices=nan(num_cols,num_rows);
% arrange sim indices for each row in a matrix
for row=1:num_rows
sim_indices(1:num_per_row(row),row)=indices{row};
end
% sim_indices=[];
% for row=1:num_rows
% sim_indices=[sim_indices find(param_mat(:,row_param_index)==row_param_values(row))];
% end
else
num_rows = 1;
sim_indices=ones(1,num_rows); % index into data array
num_cols=1;
end
ht=320; % height per subplot row (=per population or FR data set)
if ~lock_gca; handles(1) = figure('units','normalized','position',[0,1-min(.33*num_rows,1),min(.25*num_cols,1) min(.33*num_rows,1)],'visible',options.visible); end
if ~lock_gca; hsp = subplot_grid(num_rows,num_cols); end
axis_counter = 0;
for row=1:num_rows
for col=1:num_cols
sim_index=sim_indices(col,row); % index into data array for this subplot
axis_counter=axis_counter+1; % number subplot axis we're on
if isnan(sim_index)
continue;
end
if ~lock_gca; hsp.set_gca(axis_counter); end
num_pops = 1;
if isfield(data,'varied')
if num_sims>1
% list the parameter varied along the rows first
str=[row_param_name '=' num2str(row_param_values(row)) ': '];
for k=1:num_varied
fld=data(sim_index).varied{k};
if ~strcmp(fld,row_param_name)
val=data(sim_index).(fld);
str=[str fld '=' num2str(val) ', '];
end
end
if num_pops>1
legend_strings=cellfun(@(x)[x ' (mean)'],pop_names,'uni',0);
end
else
str='';
for k=1:length(data.varied)
fld=data(sim_index).varied{k};
str=[str fld '=' num2str(data(sim_index).(fld)) ', '];
end
end
text_string{row,col}=['(' strrep(str(1:end-2),'_','\_') ')'];
end
k=1;
i=sim_index;
dat=data(i).(FR_fields{k});
bins=0:1.05*max(dat(:));
rlims=[min(bins) max(bins)];
if rlims(1)==rlims(2), rlims=rlims(1)+[0 1e-5]; end
tlims=[min(time) max(time)];
% get population name from field (assumes: pop_*)
popname=regexp(FR_fields{k},'^([a-zA-Z0-9]+)_','tokens','once');
popname=popname{1};
ncells=size(dat,2);
nc=4; % number of plots for this FR data set
% 1.0 plot firing rate heat map
if strcmp(options.plot_type,'heatmap')
%subplot(nsets,nc,1+(k-1)*nc); % imagesc(t,cells,FR)
imagesc(time,1:ncells,dat'); axis xy
if ~lock_gca; hsp.figtitle([popname ': firing rates (Hz) ']); title(text_string{row,col}); end
if row == num_rows; xlabel('time (ms)'); end; ylabel([popname ' cell index']);
caxis(rlims); xlim(tlims);
if ncells<=10
ytick=1:ncells;
else
ytick=round(linspace(1,ncells,5));
if ~ismember(ncells,ytick)
ytick=[ytick ncells];
end
end
set(gca,'ytick',ytick,'yticklabel',ytick);
end
% 2.0 plot sorted firing rate heat map
if strcmp(options.plot_type,'heatmap_sorted')
%subplot(nsets,nc,2+(k-1)*nc); % imagesc(t,cells_sorted,FR)
tmp=sum(dat,1);
[tmp,inds]=sort(tmp);
imagesc(time,1:ncells,dat(:,inds)'); axis xy
caxis(rlims); xlim(tlims);
if ~lock_gca; hsp.figtitle([popname ': firing rates (Hz) ']); title(text_string{row,col}); end;
if row == num_rows; xlabel('time (ms)'); end; ylabel([popname ' cell index (sorted by FR)']);
if ncells<=10
ytick=1:ncells;
yticklabel=ytick(inds);
else
ytick=[];
yticklabel=[];
end
set(gca,'ytick',ytick,'yticklabel',yticklabel);
end
% 3.0 plot population-average firing rate trace
if strcmp(options.plot_type,'meanFR')
%subplot(nsets,nc,3+(k-1)*nc); % plot(t,<FR|pop>)
plot(time,mean(dat,2),'o-','linewidth',2);
if ~lock_gca; hsp.figtitle([popname ': pop. avg FR']); title(text_string{row,col}); end
if row == num_rows; xlabel('time (ms)'); end; ylabel([popname ': avg firing rate (Hz)']);
ylim(rlims); xlim(tlims);
end
% 4.0 plot firing rate density
if strcmp(options.plot_type,'meanFRdens')
if numel(bins)>1
%subplot(nsets,nc,4+(k-1)*nc); % hist(FR)
H=hist(dat(:),bins)/numel(dat);
h=bar(bins,H); hold on
set(get(h,'children'),'FaceAlpha',0,'EdgeAlpha',.4,'linewidth',2);
rn=ksr(bins,H,.75,length(bins));
plot(bins,rn.f,'linewidth',2); if ~lock_gca; hsp.figtitle([popname ': FR density']); end; title(text_string{row,col});
if row == num_rows; xlabel([popname ': firing rate (Hz)']); end; ylabel('fraction');
xlim(rlims); ylim([0 1]);
end
end
end
end
end
end