Bio 301 - Averaging Sweeps

Biological Sciences 300/301, Smith College | Neurophysiology

Appendix: Averaging Oscilloscope Sweeps

Averaging Sweeps on Rigol Oscilloscopes

If the signal you are observing is triggered and repetitive (such as the ERG), but contaminated with electrical noise, you can "clean up" the signal and remove random noise by using signal averaging. The oscilloscope provides this feature in the Acquire menu.

To set up signal averaging:

1. Press the Acquire button in the MENU area (top right) to launch the Acquire menu.

2. Press the top gray menu button to launch the Acquisition sub-menu.

3. Select Average instead of the usual Normal setting by turning the selection knob and pressing it.

4. The Averages choice will appear in the Acquire menu. Turn and press the selection knob to specify 8 or 16 sweeps.

The more sweeps in the average, the more the signal will be cleaned of random noise, but also the longer you will have to wait after making any changes in the display or the experiment. Eight sweeps is a good choice to start with.

5. Dismiss the Acquire menu.

To end signal averaging, repeat steps 1 to 3, but select Normal in the Acquisition sub-menu.

How Averaging Works

The figure on the left is a heavily amplified ERG response to a dim light flash. Random electronic noise obscures the shape of the response. On the right, 16 responses to the dim flash have been averaged. Most of the electronic noise has been removed, making it easier to observe and measure the basic form of the ERG.

The oscilloscope averages sweeps by adding the most recent n sweeps together and dividing by n. For example, if you select 8 sweeps to average, the displayed signal will gradually improve as the first eight sweeps are acquired and averaged. After eight sweeps, the next average will drop the first sweep but include the ninth; the average after that will drop the second sweep but include the tenth, etc. The average always represents a moving group of the most recent n sweeps. This is sometimes called "boxcar averaging."

If you make any change in the oscilloscope's controls (such as adjusting the vertical position of the trace), the average starts over and you will have to wait for n sweeps before a fully averaged signal is displayed. This includes the Run/Stop button, which interrupts averaging.

If you change any aspect of the experiment (such as changing the light intensity for the ERG), you will have to wait for n sweeps before a signal representative of the new condition is displayed. Before then, the signal will be a mixture of the old response and the new one.

Any signal that can be triggered to remain in one place on the screen can be averaged. The trigger can be from any source related to the signal of interest, such as the stimulus flash for an ERG, the synchronizing pulse from a stimulator, or even the response of interest itself (such as a large spike accompanied by unrelated smaller spikes in an extracellular recording). Averaging improves the ability to see the triggered response, while random components of the signal that are not related to the triggered event are diminished.

	Biological Sciences 300/301, Smith College \| Neurophysiology Appendix: Averaging Oscilloscope Sweeps

	Averaging Sweeps on Rigol Oscilloscopes If the signal you are observing is triggered and repetitive (such as the ERG), but contaminated with electrical noise, you can "clean up" the signal and remove random noise by using signal averaging. The oscilloscope provides this feature in the Acquire menu.
	To set up signal averaging:

	1. Press the Acquire button in the MENU area (top right) to launch the Acquire menu. 2. Press the top gray menu button to launch the Acquisition sub-menu. 3. Select Average instead of the usual Normal setting by turning the selection knob and pressing it. 4. The Averages choice will appear in the Acquire menu. Turn and press the selection knob to specify 8 or 16 sweeps. The more sweeps in the average, the more the signal will be cleaned of random noise, but also the longer you will have to wait after making any changes in the display or the experiment. Eight sweeps is a good choice to start with. 5. Dismiss the Acquire menu. To end signal averaging, repeat steps 1 to 3, but select Normal in the Acquisition sub-menu.
	How Averaging Works The figure on the left is a heavily amplified ERG response to a dim light flash. Random electronic noise obscures the shape of the response. On the right, 16 responses to the dim flash have been averaged. Most of the electronic noise has been removed, making it easier to observe and measure the basic form of the ERG. The oscilloscope averages sweeps by adding the most recent n sweeps together and dividing by n. For example, if you select 8 sweeps to average, the displayed signal will gradually improve as the first eight sweeps are acquired and averaged. After eight sweeps, the next average will drop the first sweep but include the ninth; the average after that will drop the second sweep but include the tenth, etc. The average always represents a moving group of the most recent n sweeps. This is sometimes called "boxcar averaging." If you make any change in the oscilloscope's controls (such as adjusting the vertical position of the trace), the average starts over and you will have to wait for n sweeps before a fully averaged signal is displayed. This includes the Run/Stop button, which interrupts averaging. If you change any aspect of the experiment (such as changing the light intensity for the ERG), you will have to wait for n sweeps before a signal representative of the new condition is displayed. Before then, the signal will be a mixture of the old response and the new one. Any signal that can be triggered to remain in one place on the screen can be averaged. The trigger can be from any source related to the signal of interest, such as the stimulus flash for an ERG, the synchronizing pulse from a stimulator, or even the response of interest itself (such as a large spike accompanied by unrelated smaller spikes in an extracellular recording). Averaging improves the ability to see the triggered response, while random components of the signal that are not related to the triggered event are diminished.