Example of Plotting Simulation Vector Data

See the gvDsPowerExample.m inside the GV examples folder.

%% params
eqns='IB';
vary={
  'IB','Iapp',1:2;
  'IB','gNa',1:2;
  };

study_dir = fullfile(pwd, 'gvDsPowerExample');

%% simulate
dsSimulate(eqns, 'time_limits',[0 250], 'vary',vary, 'study_dir',study_dir, 'analysis_functions',{@gvCalcPower});

%% move to power_results
resultsDir = fullfile(study_dir, 'results');

powerResultsDir = fullfile(study_dir, 'power_results');
mkdir(powerResultsDir);

powerResultFiles = lscell(fullfile(resultsDir, '*_gvCalcPower.mat'));

for thisFile = powerResultFiles(:)'
  src = fullfile(resultsDir, thisFile{1});
  dest = fullfile(powerResultsDir, thisFile{1});
  
  movefile(src, dest);
end

%% run ds
cd(study_dir);
gvImportDsPower();
gvRun();

Use this function to import power result data and save the 'gvArrayData.mat' file. Once that file is created, future calls to GV can use the standard commands (i.e., gvr, gvRun, gv.Run). Importantly, the power results should be moved to a separate folder. Be sure to edit this function under the edit section to write the power function called by DynaSim and the path to the separated power results directory.

function gvObj = gvImportDsPower
%% gvImportDsPower
% Purpose: This function imports power data from a DynaSim simulation to a gv
% hypercube, enabling each simulation point to hold the vector data of
% frequencies, instead of just a scalar, for each simulation. Importantly, one
% should keep the power data in a separate folder, so it is not imported with
% the rest of the data initially. After creation of the gvArrayData.mat object,
% one can run GV using the normal commands (i.e., gvr, gvRun, gv.Run).
%
% Implementation:
%   1) Add a new axis to the gvArray hypercube for the vector data, in this case  
%      the frequencies for the fft power.
%   2) Import the power data and store in a separate gvArray hypercube.
%   3) Merge the gvArray hypercube objects
%
% Author: Erik Roberts, 2018

%% ---- Edit ----
powerFnName = @gvCalcPower; % the function used to calculate the power
powerResultsPath = fullfile(pwd, 'power_results'); % the path to the directory storing the power results
maxFreq = 100; % hz
%% --------------

% make gv object
gvObj = gv();

% import ds data
gvObj.importDsData(pwd, 'saveBool', 0);

% grab the axis metadata which would be discarded
axismeta = gvObj.model.data.dsData.axis(1).axismeta;

%% power data
% import ds power data separately
powerResults = dsImportResults(powerResultsPath, 'import_scope','custom', 'func',powerFnName, 'as_cell',1);

% trim to maxFreq
powerResults = cellfun(@trim2maxFreq, powerResults, 'uni',0);

nFreqs = length(powerResults{1});
freqs = powerResults{1}(:,2);

% remove freqs
powerResults = cellfun(@removeFreqCol, powerResults, 'uni',0);

% take log power
powerResults = cellfun(@log, powerResults, 'uni',0);

%% power freq axis
% Add freq axis to original ds data in gv hypercube. All the scalar simulation
% data will be stored in the first value of the frequency axis. The rest of the
% frequencies will be empty for other data types besides power.

% add new axes for power
gvObj.model.data.dsData.axis(end+1) = gvArrayAxis;

% add axis info
gvObj.model.data.dsData.axis(end).name = 'freq';
gvObj.model.data.dsData.axis(end).values = freqs(1);

%% sim IDS
analysisFnAxInd = contains(gvObj.model.data.dsData.axisNames, 'analysisFn');
simIDind = contains(gvObj.model.data.dsData.axis(1).values, 'simID');
simIDs = sliceAnyDim(gvObj.model.data.dsData.data, simIDind, analysisFnAxInd);

simIDs = cell2mat(simIDs);

simIDsSize = size(simIDs);
simIDsSize(end+1) = nFreqs; % for later reshape

simIDs = simIDs(:); % reshape to col vector

%% get axis names and vals
axis_names = gvObj.model.data.dsData.exportAxisNames;
axis_vals = gvObj.model.data.dsData.exportAxisVals;
axis_vals{1} = {'power'};
axis_vals{end} = freqs;

%% make new data of correct dims
data = zeros(length(simIDs), nFreqs); % col vector

nSims = length(powerResults);
for simID = 1:nSims
  if ~isempty(powerResults{simID})
    data(simIDs == simID, :) = powerResults{simID};
  else
    data(simIDs == simID, :) = nan;
  end
end

% reshape the data
data = reshape(data, simIDsSize);

%% add data to new gvArray obj
tempObj = gvArray;
tempObj = tempObj.importData(data, axis_vals, axis_names);

%% merge objects
gvObj.model.data.dsData.merge(tempObj);

%% add axis metadata
% this was dropped in merge

axismeta.dataType{end+1} = 'numeric';
gvObj.model.data.dsData.axis(1).axismeta = axismeta;

%% save gvArrayData
dynasimData = gvObj.model.data.dsData;
filePath = fullfile(pwd, 'gvArrayData.mat');
save(filePath, 'dynasimData') % save gvArray obj

%% Nested Fn
  function cellOut = removeFreqCol(cellIn)
    % just take first sxx col, discarding second freq col
    cellOut = cellIn(:,1);
  end

  function cellOut = trim2maxFreq(cellIn)
    % remove freqs above maxFreq
    selectedFreqs = cellIn(:,2) <= maxFreq;
    
    cellOut = cellIn(selectedFreqs,:);
  end

end

function cellOut = gvCalcPower(data,varargin)
%% gvCalcPower
%  Purpose: This wrapper function for dsCalcPower is used as a DynaSim analysis
%           function for calculating the power spectrum vector of a
%           simulation. It should be used in tandem with gvImportDsPower.
%
%  Inputs:
%   unit_type: power calculated for 'SUA' or 'MUA' (default).
%
% See also: dsCalcPower, gvImportDsPower

%% Check inputs
options=dsCheckOptions(varargin,{...
  'variable',[],[],...
  'output_suffix','',[],...
  'unit_type', 'MUA', {'SUA', 'MUA'},...
  },false);

% select variable
var = dsSelectVariables(data(1),options.variable, varargin{:});
var = var{1};

% calculate the power
data = dsCalcPower(data, varargin{:});

% pull out power and freqs
switch options.unit_type
  case 'SUA'
    sxx = data.([var '_Power_SUA' options.output_suffix]).Pxx;
    freq = data.([var '_Power_SUA' options.output_suffix]).frequency;
  case 'MUA'
    sxx = data.([var '_Power_MUA' options.output_suffix]).Pxx;
    freq = data.([var '_Power_MUA' options.output_suffix]).frequency;
end

% store power and freqs in a matrix
cellOut = {[sxx(:), freq(:)]};

end % gvCalcPower