%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [significant, VAlt, Ratio] = CalcValues(avg_psd)
% CalcValues.m
% Author: Alexander Khaustov; alexander dot khaustov at gmail dot com
% Copyright (C) 2008 St.-Petersburg Institute of Cardiological Technics (Incart), www.incart.ru
% This software is released under the terms of the GNU General
% Public License (http://www.gnu.org/copyleft/gpl.html).
%
% CalcValues calculates the alternans values for Spectral Method
% INPUT:
% avg_psd (num_of_beats/2+1 x num_of_leads) is the psd averaged across the entire S-Tend segment
% OUTPUT:
% significant (1 x num_of_leads) 1 if the alternans value is statistically significant against the Param.RatioThreshold in the lead and 0 otherwise
% VAlt (1 x num_of_leads) spectral amplitude @ 0.5 cycles-per-beat -
% 'voltage of alternans'
% Ratio (1 x num_of_leads) 'alternans ratio'
global Param
NoiseStart = floor(2 * 0.4 * (length(avg_psd) - 1) + 1);
NoiseEnd = floor(2 * 0.46 * (length(avg_psd) - 1) + 1);
% Calculate the mean and std of the noise
nmean = mean(avg_psd(NoiseStart : NoiseEnd - 1));
nstd = std(avg_psd(NoiseStart : NoiseEnd - 1));
% spectral amplitude @ 0.5 cycles-per-beat
v = avg_psd(length(avg_psd));
if (v > nmean)
VAlt = sqrt(v - nmean);
Ratio = (v - nmean) / nstd;
else
VAlt = 0;
Ratio = 0;
end;
significant = Ratio >= Param.RatioThreshold;
return;