Description |
%function k=FindOptKern(InSig,OutSig,MaxnLags)
%
% Given InSig, OutSig, and the MaxnLags, this will find the optimal causal
% kernal of length MaxnLags that when convolved with InSig gives the
% optimal approximation to OutSig in a least squares sense.
%
% Note that this code produces the optimal causal kernel. the basic
% technique does not have to produce a causal kernel, but this code does.
%
% Input
% InSig: An N x 1 vector of the input signal (i.e. Trial type in the
% trial history example)
% OutSig: An N x 1 vector of the output signal (i.e. RT in the trial
% history example)
% MaxnLags: A scalar for the maximum number of lags to calculate up to.
% Must be less than the length of the signal.
%
% k: A MaxnLags x 1 vector of the optimal kernel.
%
% For an explanation of this, See Corrado, Sugrue, Seung and Newsome (2004)
% pp. 591-592
% Also see: http://hebb.mit.edu/courses/9.29/2004/lectures/lecture03.pdf
%
% by MJN on 090318 |