Online Signal Averaging Example

Import strobe store gizmo data into Matlab using SynapseLive during the
experiment
Plot the average waveform
Good for Evoked Potential visualization

Show Code Download

Contents

Housekeeping

Clear workspace and close existing figures. Add SDK directories to Matlab path.

close all; clc;
[MAINEXAMPLEPATH,name,ext] = fileparts(cd); % \TDTMatlabSDK\Examples
[SDKPATH,name,ext] = fileparts(MAINEXAMPLEPATH); % \TDTMatlabSDK
addpath(genpath(SDKPATH));

Variable Setup

Set up the variables for the data you want to extract. We will extract a single channel from a fixed duration strobed storage gizmo.

EVENT = 'StS1';
CHANNEL = 1;

show the last N waveforms in the plot.

N = 5;

Set KEEPALL to 0 to only show the running average of the last N waveforms. Otherwise, all waveforms in the block are included in the average.

KEEPALL = 0;

Setup SynapseLive

t = SynapseLive('MODE', 'Preview', 'EXPERIMENT', 'OnlineAveragingDemo'); % we will default to 'Preview' mode
t.TYPE = {'snips'}; % we only care about these types of events
t.VERBOSE = false;
first_pass = true;

Connected to TANK: F:\Tanks\OnlineAveragingDemo-190523-094854, BLOCK: TempBlk
Waiting for initial data..................done

The Main Loop

prevWaves = cell(1,N);
nsweeps = 0;
while 1

    % slow it down a little
    pause(1)

    % get the most recent data, exit loop if the block has stopped
    if isempty(t.update)
        break
    end

    % read the snippet events
    r = t.get_data(EVENT);
    if isstruct(r)
        if ~isnan(r.data)
            % get our channel of data
            chan_data = r.data(r.chan == CHANNEL,:);
            nsize = size(chan_data,1);

            % cache the waveforms in our circular buffer
            prevWaves = circshift(prevWaves, -nsize);
            for i = 1:(min(nsize, N))
                prevWaves{i} = chan_data(end-(i-1),:);
            end

            % find average signal
            cache_ind = ~cellfun('isempty', prevWaves);
            if KEEPALL == 0
                % if we are only keeping the previous N, do average on just those
                avg_data = mean(cell2mat(prevWaves(cache_ind)'), 1);
            else
                if first_pass
                    first_pass = false;
                    nsweeps = nsize;
                    avg_data = new_mean;
                else
                    new_mean = mean(chan_data, 1);
                    % add new average into the old average
                    avg_data = (avg_data .* nsweeps + new_mean * nsize) / (nsweeps + nsize);
                end
            end

            nsweeps = nsweeps + nsize;

            % plot the preview N waves in gray
            t_ms = 1000*(1:numel(avg_data)) / r.fs;
            plot(t_ms, cell2mat(prevWaves(cache_ind)')','Color', [.85 .85 .85]); hold on;

            % plot the average signal in thick blue
            plot(t_ms, avg_data, 'b', 'LineWidth', 3); hold off;

            % finish up plot
            title(sprintf('nsweeps = %d, last %d shown', nsweeps, N));
            xlabel('Time, ms','FontSize',12)
            ylabel('V', 'FontSize', 12)
            temp_axis = axis;
            temp_axis(1) = t_ms(1);
            temp_axis(2) = t_ms(end);
            axis(temp_axis);

            % force the plots to update
            try
                snapnow
            catch
                drawnow
            end

            % for publishing, end early
            if exist('quitEarly','var') && nsweeps > 30
                break
            end
        end
    end
end

Runtime Output