


Multi-taper coherency,cross-spectrum and individual spectra computed by segmenting two univariate time series into chunks - continuous process
Usage:
[C,phi,S12,S1,S2,f,confC,phistd,Cerr]=coherencysegc(data1,data2,win,params)
Input:
Note units have to be consistent. See chronux.m for more information.
data1 (column vector) -- required
data2 (column vector) -- required
win (length of segments) - required
params: structure with fields tapers, pad, Fs, fpass, err
- optional
tapers (precalculated tapers from dpss, or in the form [NW K] e.g [3 5]) -- optional. If not
specified, use [NW K]=[3 5]
pad (padding factor for the FFT) - optional (can take values -1,0,1,2...).
-1 corresponds to no padding, 0 corresponds to padding
to the next highest power of 2 etc.
e.g. For N = 500, if PAD = -1, we do not pad; if PAD = 0, we pad the FFT
to 512 points, if pad=1, we pad to 1024 points etc.
Defaults to 0.
Fs (sampling frequency) - optional. Default 1.
fpass (frequency band to be used in the calculation in the form
[fmin fmax])- optional.
Default all frequencies between 0 and Fs/2
err (error calculation [1 p] - Theoretical error bars; [2 p] - Jackknife error bars
[0 p] or 0 - no error bars) - optional. Default 0.
Output:
C (magnitude of coherency - frequencies x segments if segave=0; dimension frequencies if segave=1)
phi (phase of coherency - frequencies x segments if segave=0; dimension frequencies if segave=1)
S12 (cross spectrum - frequencies x segments if segave=0; dimension frequencies if segave=1)
S1 (spectrum 1 - frequencies x segments if segave=0; dimension frequencies if segave=1)
S2 (spectrum 2 - frequencies x segments if segave=0; dimension frequencies if segave=1)
f (frequencies)
confC (confidence level for C at 1-p %) - only for err(1)>=1
phistd - theoretical/jackknife (depending on err(1)=1/err(1)=2) standard deviation for phi - Note that
phi + 2 phistd and phi - 2 phistd will give 95% confidence
bands for phi - only for err(1)>=1
Cerr (Jackknife error bars for C - use only for Jackknife - err(1)=2)

0001 function [C,phi,S12,S1,S2,f,confC,phistd,Cerr]=coherencysegc(data1,data2,win,params) 0002 % Multi-taper coherency,cross-spectrum and individual spectra computed by segmenting two univariate time series into chunks - continuous process 0003 % 0004 % Usage: 0005 % [C,phi,S12,S1,S2,f,confC,phistd,Cerr]=coherencysegc(data1,data2,win,params) 0006 % Input: 0007 % Note units have to be consistent. See chronux.m for more information. 0008 % data1 (column vector) -- required 0009 % data2 (column vector) -- required 0010 % win (length of segments) - required 0011 % params: structure with fields tapers, pad, Fs, fpass, err 0012 % - optional 0013 % tapers (precalculated tapers from dpss, or in the form [NW K] e.g [3 5]) -- optional. If not 0014 % specified, use [NW K]=[3 5] 0015 % pad (padding factor for the FFT) - optional (can take values -1,0,1,2...). 0016 % -1 corresponds to no padding, 0 corresponds to padding 0017 % to the next highest power of 2 etc. 0018 % e.g. For N = 500, if PAD = -1, we do not pad; if PAD = 0, we pad the FFT 0019 % to 512 points, if pad=1, we pad to 1024 points etc. 0020 % Defaults to 0. 0021 % Fs (sampling frequency) - optional. Default 1. 0022 % fpass (frequency band to be used in the calculation in the form 0023 % [fmin fmax])- optional. 0024 % Default all frequencies between 0 and Fs/2 0025 % err (error calculation [1 p] - Theoretical error bars; [2 p] - Jackknife error bars 0026 % [0 p] or 0 - no error bars) - optional. Default 0. 0027 % Output: 0028 % C (magnitude of coherency - frequencies x segments if segave=0; dimension frequencies if segave=1) 0029 % phi (phase of coherency - frequencies x segments if segave=0; dimension frequencies if segave=1) 0030 % S12 (cross spectrum - frequencies x segments if segave=0; dimension frequencies if segave=1) 0031 % S1 (spectrum 1 - frequencies x segments if segave=0; dimension frequencies if segave=1) 0032 % S2 (spectrum 2 - frequencies x segments if segave=0; dimension frequencies if segave=1) 0033 % f (frequencies) 0034 % confC (confidence level for C at 1-p %) - only for err(1)>=1 0035 % phistd - theoretical/jackknife (depending on err(1)=1/err(1)=2) standard deviation for phi - Note that 0036 % phi + 2 phistd and phi - 2 phistd will give 95% confidence 0037 % bands for phi - only for err(1)>=1 0038 % Cerr (Jackknife error bars for C - use only for Jackknife - err(1)=2) 0039 0040 if nargin < 3; error('Need data1 and data2 and size of segment'); end; 0041 if nargin < 4; params=[]; end; 0042 [tapers,pad,Fs,fpass,err,trialave,params]=getparams(params); 0043 clear tapers pad fpass trialave 0044 if nargout > 8 && err(1)~=2; 0045 error('Cerr computed only for Jackknife. Correct inputs and run again'); 0046 end; 0047 if nargout > 6 && err(1)==0; 0048 % Errors computed only if err(1) is nonzero. Need to change params and run again. 0049 error('When errors are desired, err(1) has to be non-zero.'); 0050 end; 0051 if size(data1,2)~=1 || size(data2,2)~=1; error('works for only univariate time series'); end; 0052 0053 N=check_consistency(data1,data2); 0054 0055 dt=1/Fs; % sampling interval 0056 T=N*dt; % length of data in seconds 0057 E=0:win:T-win; % fictitious event triggers 0058 win=[0 win]; % use window length to define left and right limits of windows around triggers 0059 data1=createdatamatc(data1,E,Fs,win); % segmented data 1 0060 data2=createdatamatc(data2,E,Fs,win); % segmented data 2 0061 params.trialave=1; 0062 params.trialave=1; 0063 if err==0; 0064 [C,phi,S12,S1,S2,f]=coherencyc(data1,data2,params); % compute coherency for segmented data 0065 elseif err(1)==1; 0066 [C,phi,S12,S1,S2,f,confC,phistd]=coherencyc(data1,data2,params); % compute coherency for segmented data 0067 elseif err(1)==2; 0068 [C,phi,S12,S1,S2,f,confC,phistd,Cerr]=coherencyc(data1,data2,params); % compute coherency for segmented data 0069 end;