RemoveCorrSpikes(d,​Fs,sptm,varargin)

removes false-positive neural spikes from a multi channel data matrix via inter-chan correlations

% --------[sptm,spam] = RemoveCorrSpikes(d,Fs,sptm,varargin)-------------
% Removes spikes from multichannel data that are highy correlated with
% other channels...usually on order of .85 or so. This technique helps eliminate
% false postivie spikes due to movement artifacts common in all channels.
%
% This code is an interpretation of:
% Paralikar et. al. 2010
%
% >>> INPUTS >>>
% Required:
% d = data matrix. note, if vector, function will end without doing
% anything. Needs to be in COLUMN format
% Fs = sampling rate
% sptm = matrix of spike times in samples. also in column format
% Optional:
% corr_thresh = correlation coefficient threshold for removing spikes.
% Default = 0.8
% spamp = matrix of spike amplitudes...doesn't contribute to the
% function, but if included indexes removed from sptm will also be
% removed from spamp.
% spsnip = matrix of spike snips...doesn't contribute to the function but if
% included indexes removed from sptm will also be removed from spsnip
%
% <<< OUTPUTS <<<
% sptm = spiketimes after eliminating bad spikes...bad spikes will have
% value of 0.
% spamp = spike amps after eliminating bad spikes (OPTIONAL)
% spsnip = matrix of spikesnips after eliminating bad spikes (optional)
%
% By JMS, 11/12/2015
%---------------------------------------------------------------