Biological Sciences 300/301, Smith College | Neurophysiology

Appendix: Capturing Data with PowerLab and LabChart

REVISED: March 19, 2015

This online guide to PowerLab hardware and LabChart software will help you:

1. Acquire data with LabChart

2. Analyze action potential firing rates with LabChart's Spike Histogram module.

3. Analyze bursts with LabChart's Integral function

The LabChart Help menu provides two useful resources for further information: "LabChart Help," an interactive help system, and "LabChart User Guide," an extensive instruction manual (as a PDF file).

icon for Chart application

Acquiring Data with LabChart

Launch LabChart by clicking on its icon in the dock. The software will check to see if the PowerLab input box is connected and powered up, and it will warn you if it isn't. A new chart window will appear in which data will be displayed when you record it. The figure below shows the main chart window after data has been recorded, analyzed, and saved as a file named "SPIKE demo."

Main LabChart Window

Menus and Toolbar

LabChart's menu bar (above) offers many functions, some of which are also accessible from icons in the toolbar.

Channel Settings

Before you begin recording, you must set certain parameters in the Channel Settings dialog box, which you should activate now from the Setup Menu.

1. First, set the number of channels you wish to make active. Usually this will be 1 or 2 input channels for raw data, and additional calculation channels for each input channel. Set up three channels now (the figure shows four) using the box at the bottom. You can add or remove channels later by revisiting the Channel Settings dialog in the Setup menu.

2. Click the "On" box in the first column to turn on Channel 1. From the drop-down menu (second column), select the sampling rate, the number of points per second that will be digitized for each channel. 40K is a good choice, and 20K will be satisfactory if you are concerned about the length of files.

"40K" means that 40,000 samples will be taken every second, or 40 every millisecond. This will give very good images of individual spikes when you stretch the time scale to see them.

3. Check "Same Sampling Rate on All Channels" (bottom right) to make the sampling rate apply to all the channels.

4. Select a calculation from the drop-down menu (last column) for any channels that will display raw input data (this can also be done later). This example shows Digital Filtering, which brings up the dialog box shown at the right. Select the source, filter type [high-pass], and cut-off frequency as shown to remove low-frequency components of Channel 1's signal. These settings will help eliminate slow baseline drift.

Input Amplifier Settings

5. Activate the Input Amplifier dialog box, shown below, by double-clicking on the "Input Settings" column for each channel that receives raw data (you must set each one separately). This is the easiest way to set the vertical scale (the "Range"), but it can be done only after you have neural activity to examine.

The Input Amplifier window will show you the raw data coming in on that channel. Select an appropriate voltage range from the drop-down menu to enlarge or reduce the vertical amplitude of the signal so it fits in the display box. This window is also where you should choose "Single-ended" (ie, one signal wire vs. ground for each channel), AC coupling if you are looking at extracellular spikes, and "Mains Filter" to filter out 60-cycle power line interference.

When everything is set for this channel, click "OK" and continue with the input settings for any other channels that receive raw data.

Recording Data and Saving Files

After you have established the required settings, you are ready to record data. Clicking the "Start" button at the bottom right corner of the main chart window, shown in the figure below, will begin digitizing data, which appears on the screen like a moving chart. The button becomes a "Stop" button after data acquisition begins. Each start and stop of recorded data establishes a data block in the window.

If you are merely examining the signal but do not wish to keep any of the data, clicking the button to the left of the Start button will cause a red X to appear, indicating that data are not being saved in the computer's memory. (Only the graphics screen is being written to as samples are taken and transiently displayed.) Clicking the button again (the red X vanishes) causes samples to be saved in the computer's memory, recording the data for you to look at after you stop the sampling. (Note that the data are not yet in a file; you need to Save current data (File menu) if you wish to be able to go back to that data on a future occasion.)

Zoom Window

The Zoom Window will enlarge any selected region of data from the main chart window and display it in a new window. The magnifying glass in the toolbar opens the zoom window after you have selected data to enlarge.

  • If you drag across data on one channel to create a highlight area (yellow region above), only that area will be displayed (in both height and width) in the zoom window.
  • If instead you drag across the time scale at the bottom of the chart window, you will highlight data on all the channels, each at full height.
  • You can highlight a portion of the zoom window to instantly enlarge this new selection. The "Selection History" arrows at the bottom right allow you to go forward and back through your zoom selections.
  • The buttons at the upper right determine whether data from multiple channels is overlayed on top of each other, or displayed as separate traces. Usually, you will prefer separate traces.

Analyzing firing rates with the Spike Histogram module

The Spike Histogram module allows you to select individual spike types in a multiunit recording based on their height and width. The selected spikes are displayed below the raw data on a new channel. The contents of that channel become the input to a third channel that shows the rate, interspike interval, or frequency of firing.

Spike Train Setup

In the Spike Histogram menu, select "Spike Train Setup," which will open a small window, as shown on the left below. You may create multiple examples of spike trains to display and analyze, each with a different name, based on the parameters you will set later. Click the New item to bring up a second window where you will click radio buttons to set the name ("General"), the Data Source (in this example. the raw data in Channel 1), and the threshold below which any activity is discarded as noise. (Unfortunately, you must enter the threshold numerically after examining the raw data's vertical scale.)

Note that the spike discriminator accepts only positive-going spikes. If your data has predominantly negative-going spikes, you must click the "Invert Data" box in the Data Source window to make them appear positive. In that same window, you may choose to analyze an entire block of data, or only a highlighted selection from within the block.


Once you have set the spike train criteria, LabChart quickly identifies all events that are potential spikes. In a multiunit recording, these will invariably be spikes from more than one axon. In this example, the program has found 1559 events that exceed the noise threshold.

The next step is to isolate spikes belonging to one axon, based on their height and width. From the drop down menu belonging to Train 1, select "Open" and then "Discriminator" from the submenu, as shown in the adjoining figure.

This will open a new data window, "SH: Discriminator," shown below. In this example, the raw data had a variety of spike-heights, and the discriminator has found six clear clusters of dots (with perhaps two more clusters near the baseline). Each dot represents a single spike in the raw data, plotted according to the spike's height and width. Click on a dot to display the shape of its corresponding spike on the right side of the discriminator window. Clicking several dots in a cluster will let you judge whether they represent the same spike.

In this example, the boundaries of the isolation rectangle have been dragged to enclose a cluster of dots representing the second biggest spike in the sample. Defining the isolation rectangle is the key step in isolating a single axon's spike from all the others in a multiunit recording.

If the spikes are relatively sparse, choosing a bigger dot-size (square box at bottom left of the discriminator window) will make them easier to see.

Displaying the Discriminator Output on a Channel

Display the output of the spike discriminator on Channel 2, as shown in the middle trace (blue). Select "Discriminator" from the menu invoked by the arrow next to the words "Channel 2" at the right of the main data window. A dialog box will appear (shown at the left). Select the spike train to display.

The identified spikes will appear as vertical lines of uniform size on Channel 2, allowing you to see if they correspond to a single spike type in the raw data channel. In this example, the two blue markers on Channel 2 correspond to the same medium-sized spike on Channel 1, indicating that the discriminator has successfully isolated a spike. You may need to expand the time scale and scroll through the data to make this comparison.

If you change the selection criteria in the SH:Discriminator window, the displayed spikes on Channel 2 will instantly reflect the new criteria. This allows you to fine-tune your selection criteria.

Displaying the Firing Rate

The last step is to choose a method for displaying the firing rate of the isolated spikes. From the drop-down menu belonging to Channel 3 (far right of the main data window, selected by the arrow next to the words "Channel 3," choose "Cyclic Measurements."

A new dialog box will open, shown below. In this example, Frequency has been selected as the Measurement. The Source Channel was selected as Channel 2, the clean train of markers identified by the SH:Discriminator window. The scale for the Frequency was set to an appropriate range using Auto Scale. (The Set Scale dialog box, shown at the bottom left of the figure below, will let you refine the scale later if necessary.) From the Detection Settings drop-down Preset menu, "General - Simple threshold" was chosen to pick off the Discriminator markers, and the Threshold was set to 2 V.

"Frequency" plots instantaneous frequency. The software measures the time between each pair of spike markers on Channel 2, and calculates the firing rate (impulses/sec) as if the neuron were firing steadily at that rate. Thus the plot on Channel 3 rises and falls with each new spike, reflecting the interval prior to that spike. You can estimate the average firing rate by eyeballing a "best-fit" line through the instantaneous frequency plot.


Analyze bursts with the Integral Function

The integral module adds up (integrates) the moment-to-moment values of the raw data signal to show the overall shape and timing of a burst of spikes. Integration is used when it would be difficult to isolate a single unit among many similar units, or where multi-unit activity is of central interest.

The record below shows two data channels (1 and 2) and their integrals (3 and 4, color-coded to match the original data).

Using the Integral Function

To add integrals to the LabChart record:

  • From the Setup menu, select Channel Settings. In the dialog box, add an additional blank channel for each Input Amplifier channel.
  • In the drop-down menu for each new channel (green circle in the figure above), select Integral (near the bottom of the list). This will bring up a dialog box like the one shown for Channel 4.
  • In the dialog box, select the channel that will be the Source of the integrated data. In this example, the Source is Channel 2.
  • Select the Integral type (Positive Only or Absolute Value are appropriate choices). This specifies the components of the signal that will be added up.
  • Select the Reset type as "Time Constant Decay." Decay prevents the running sum (the integral) from becoming bigger and bigger indefinitely, bringing the integral back to the baseline.
  • Select the value for the Time Constant Decay time as described below.

Choosing the Time Constant Decay

Set the Time Constant Decay empirically so that bursts are smoothed an appropriate amount, neither too little nor too much. The decay time in the example above is 75 msec (0.075 seconds). Try various values to see which setting works best with your data.

In the left example below, the decay time is 5 msec (0.005 seconds). This is too short, since individual spikes appear with little smoothing. The example on the right (500 msec, or 0.5 seconds) is too long. It smooths the bursts well, but with very long rise- and fall-times that misrepresent the timing of the burst.

The final example (50 msec) is more useful: individual spikes are smoothed, but the burst rises and decays in a reasonable time.

In practice, you will need to try different values to see which one represents your data well.


Text © 2012, 2015 by Richard F. Olivo