function y=edr(varargin)

 y = edr(data_type,signal,r_peaks,fs,pqoff, jpoff, gain_ecg, channel, show)

 ECG-derived Respiratory (EDR) signal computation from given
 single-lead ECG signal based on the signed area under the QRS complex.

 Required Parameters:

 data_type
       A 1x1 integer specifying the file data_type
       0 --> if Matlab file
       1 --> if record in MIT format

 signal
       A Nx1 integer array containing the ECG signal in mV (if data_type=0)
       OR a char string containing record name (if data_type=1)

 fs
       A 1x1 integer specifying the sampling frequency in hz (for Matlab variables only)

 r_peaks

       A Mx1 integer array containing locations of r peaks on signal in s
       OR a char string containing the extension of the annotation file
       with r peaks in samples (e.g. "qrs") (if data_type=1)

 optional parameters:

 gain_ecg
       A 1x1 integer specifying dig_max/phy_max (default=1)

 channel
       A 1x1 integer>1 (default=1) indicating ECG channel (if data_type=1)

 pqoff
       A 1x1 integer>0 specifying average distance between PQ junction and
       R peak, in samples

 jpoff
       A 1x1 integer>0 specifying average distance between R peak and
       J point, in samples

 show
       A 1x1 boolean if true, generates a plot of the estimated
       respiration signal (default = 0).


 output:

 y
       A Mx2 integer matrix containing time in seconds and edr

 This code was written by Sara Mariani at the Wyss Institute at Harvard
 based on the open-source PhysioNet code edr.c
 (http://www.physionet.org/physiotools/edr/)
 by George Moody

 Author: Sara Mariani, 2014
 Last Modified: November 17, 2014

 please report bugs/questions at sara.mariani@wyss.harvard.edu

 Example - Extract EDR signal from ECG in PhysioNet's Remote server:
 signal='fantasia/f1o02';
 r_peaks='ecg';
 data_type=1;
 channel=2;
 show=1;
 y=edr(data_type,signal,r_peaks,[],[],[],channel,show);
 wfdb2mat('f1o02')
 [~,signal,Fs,~]=rdmat('f1o02m');
 resp=signal(:,1);
 resp=resp-mean(resp);
 resp=resp*200;
 sec=length(resp)/Fs;
 xax=[.25:.25:sec];
 r=interp1(y(:,1), y(:,2), xax, 'spline');
 figure
 plot(xax,r)
 hold on
 plot([1:length(resp)]/Fs,resp,'r')
 legend('edr','respiratory signal')
 xlabel('time (s)')

 see also: ecgpuwave, gqrs