function [m]=my_filter_individualneurons(filename)
coeffthreshold=10;
time_conversion=20000;
f=10000; %nyquist frequency
totallength=0;
spikes_sparse=[];

%flags
plot_unfilteredfiltered_comparison=0;
plot_filteredspike_comparison=0;
flag_plotconvolutions=0;

for numfile=1:1%numfiles
    
    load(sprintf('%s.mat',filename));

    %discards intial data from each trial b/c this intial data shows a prominent transient behavior
    data=[];
    for i=1:length(dataStruct.data)
        temp=dataStruct.data(i);
        temp=temp{:,:};
        data=[data; temp(1001:end,:)];
    end
    
    
    spikes=[];
    numpoints=length(data);
    
    data_fft=zeros(f,1);
    for i=1:64
        
        %%%%%%%%%%%%%%%%%%%%%%%%%
        %filter               %%%
        %%%%%%%%%%%%%%%%%%%%%%%%%
        clear temp;
        temp=detrend(data(:,i));
        data_fft=fft(temp, f);
        data_fft2=data_fft;
        %zero window around frequencies with coefficients above threshold
        for j=1:f
            if abs(data_fft(j,1)) > coeffthreshold
                if j-2<1
                    low=1;
                else low=j-2;
                end
                if j+2>f
                    high=f;
                else high=j+2;
                end
                for k=low:high
                    data_fft(k,1)=0;
                end
            end
            
        end
        
        %zero out frequencies above 5000Hz
        data_fft(round(5000*f/time_conversion+1):f-round(5000*f/time_conversion-1))=0;
        
        %zero out frequencies below 500Hz
        data_fft(1:round(500*f/time_conversion))=0;
        data_fft(f-round(500*f/time_conversion-1):f)=0;
        
        %compute inverse transform
        temp=real(ifft(data_fft,numpoints));
        temp=temp-mean(temp);
        
        
        if plot_unfilteredfiltered_comparison==1
            subplot(8,8,i);
            plot(dataStruct.timeVec,dataStruct.data(:,i), 'r', dataStruct.timeVec,temp, 'g');
            xlabel('Time (ms)');
            ylabel('Voltage');
            title('Comparison of Filtered and Unfiltered Signals');
        end
        
        
        %%%%%%%%%%%%%%%%%%%%%%%%%
        %extract spikes       %%%
        %%%%%%%%%%%%%%%%%%%%%%%%%
        
        load template_spikes
        
        numspikes=zeros(10,1);
        template_shape2=zeros(10,length(temp));
        for neuronline=1:10
            %convolve spike shape for a given neuronal cell line with filtered data
            shapei=template_spikes(neuronline);
            shapei=shapei{:,1};
            temp_shape=conv(shapei',temp');
            
            %sets limits on part of convolution that is retained
            conv_min=round((length(shapei)-1)/2);
            conv_max=conv_min+length(temp)-1;
            
            %each row of temp_shape2 is the result of convolving the template shape with the filtered signal, the ends of the convolution are discarded.
            temp_shape2(neuronline,:)=temp_shape(conv_min:conv_max);
            
            %find spikes for convolved data
            %set threshold
            spikefactor=.1;
            spikethresh=shapei'*shapei*spikefactor*std(temp)/max(abs(shapei));
                
            %set maximum spike time
            spikemax=2e-7;
            
            %holds spikes contained in convolved data
            holder=logical(zeros(1,length(temp)));
            
            %puts spike at maximum of signal
            for j=1:numpoints
                if (temp_shape2(neuronline,j) > spikethresh)|(temp_shape2(neuronline,j) < (-1*spikethresh))
                    maxpoint=j;
                    for k=j:j+round(spikemax*time_conversion)
                        if temp_shape2(neuronline,k) > temp_shape2(neuronline,j)
                            maxpoint=k;
                        end
                    end
                    holder(1,maxpoint)=1;
                    j=k+1;
                end
            end
            
            temp_shape2(neuronline,:)=sparse(holder);
            numspikes(neuronline,1)=double(holder)*double(holder)';
        end
        
        
        if flag_plotconvolutions==1
            subplot(11,1,1);
            plot([1:length(temp)]'/20000,temp);
            xlabel('Time (s)','fontsize',8);
            ylabel('Voltage (mv)','fontsize',8);
            title(sprintf('Filtered Signal of Electrode %d, Threshold Factor=%s',i,spikefactor), 'fontweight','bold');
            for neuronline=1:10
                subplot(11,1,neuronline+1);
                plot([1:length(temp)]'/20000,temp_shape2(neuronline,:)');
                title(sprintf('Convolution of Template Spike %d and the Filtered Signal',neuronline),'fontsize',8,'fontweight','bold');
            end
        end
        
        
        %find neurons at a given electrode that have at least a threshold number of spikes
        spike_threshhold=100;
        for nueronline=1:10
            if numspikes(neuronline,1) > spikethreshold
                spikes=sparse([spikes'; neuronline i temp_shape2(neuronline,:)]');
            end
        end
        
        if plot_filteredspike_comparison==1
            subplot(8,8,i)
            plot(dataStruct.timeVec,temp,'g', dataStruct.timeVec, .0005*spikes(1:numpoints,i),'b');
            xlabel('Time (ms)');
            ylabel('Voltage / Spike');
            title('Comparison of Filtered Signal and Extracted Spikes');
        end
    end
end

% save extracted spikes as sparse matrix
save(sprintf('%s_spikesindivneurons.mat',filename),'spikes');

[m]=spikes;