Daqarta for DOS Contents



Free Registration

Contact Us
Daqarta for DOS
Data AcQuisition And Real-Time Analysis
Shareware for Legacy Systems

From the Daqarta for DOS Help system:



Averaging is only beneficial if you can keep each sweep aligned with the desired signal. The best way to do this is to have the averager start each sweep in response to the same stimulus that is to be presented to the subject or system under test. There are several ways to arrange this:

  • The stimulus and the averager may both be started by some external pulse. For auditory evoked potentials, a pulse generator may be used to drive a headphone to produce a series of clicks, usually first passing them through a band-pass filter that provides some control over the frequency content, and an attenuator to control the sound level. The same pulse generator may have a second output that provides an identically-timed TTL pulse to the averager External Trigger input.
  • A closely-related method is to have the the stimulus generator also generate a TTL trigger pulse at the start of each stimulus. For example, a pulse generator may be used to trigger a device called a "tone switch" or "tone gate", which produces tone bursts by gating an external tone in a controlled manner. If the tone switch needs to delay the gate pulse slightly while waiting for the next rising zero crossing of the tone, it will delay its TTL sync output pulse to match. This sync output is what you would use as the averager External Trigger, insuring that each stimulus is identical and each sweep is exactly in sync with it. You couldn't use the initial pulse for this because of the variable delay needed to synchronize to the tone.
  • The computer itself can generate the stimulus, in which case it maintains perfect alignment. Daqarta offers two general approaches to this:
    • A Stimulus Generator module can be used to generate tone bursts or other analog stimuli via DACs if they are present, or pulses from a digital I/O port or even a commandeered printer port if that is all that is available.
    • A simple pulse that lasts the duration of the sweep can be provided from any available digital I/O port without the presence of a Stimulus Generator module. This can be used to provide a step stimulus for deriving an impulse response, or can trigger some more elaborate external stimulus as in filtered clicks.
CAUTION: Do not connect any electrical equipment to a living subject without proper signal isolation techniques. A lethal shock could result.

If the signal being studied does not result from a stimulus that you generate directly, there may still be ways of getting a trigger. If the signal includes a transient peak that rises well above the noise, you may use Internal level triggering by setting the threshold Level appropriately. This may be needed when studying spontaneous natural phenomena such as nerve discharges.

If you are using level triggering and the trigger portion of the signal fades or drops out, you want the averager to stop and wait for a valid trigger. For this reason Auto-level triggering is NOT RECOMMENDED for averaging use. In this mode the level could self-adjust so that it triggers on background noise or the wrong portion of the input signal, contaminating the average.


The real world is rarely as cooperative as we would like. If the desired response is produced "some of the time", interspersed with periods of spurious responses or "artifacts", we would like to average only the good stuff and skip the trash. A typical situation is where an evoked neural response is contaminated by occasional bursts of muscle activity when the subject fidgets.

CAUTION: Do not connect any electrical equipment to a living subject without proper signal isolation techniques. A lethal shock could result.

If the noise bursts are larger than the desired response, we can reject them based upon a simple test: We collect the data for each sweep, but before adding it into the average it is scanned to see if any part of the data goes above a positive threshold or below a negative threshold. If so, that whole sweep is thrown away and not counted toward the total number of sweeps requested. If the subject is very restless, this may greatly extend the averaging time (which will no doubt make the subject still more restless!), but there may be cases when this is the only way to get useful results.

Setting the threshold limits is a matter of judgement and experience, since there is a tradeoff between accepting too much artifact and rejecting too much signal. The best way to start is to simply observe the raw Live signal and note where the usual desired peaks reach in comparison to occasional bursts where there is obvious artifact. Then set the thresholds by eye to be just beyond the normal range.

Try a few averages at these settings, then at tighter and broader ranges. Compare not only the resulting waveforms, but also the time it took to get them. If tighter thresholds give substantially better results but at a big increase in time, you will need to decide if this is acceptable or if another setting is more appropriate. It is rare to find that the same settings work best in all subjects, but you may at least find a good compromise as a starting point. Remember too that time spent tweaking the artifact limits is time that could be used to collect data at tighter settings.


In some cases, the artifacts may not be distinguishable based only on level. If they can be detected by eye on a Live trace, if they persist over several sweeps, and if the sweep repetition rate is slow enough to allow manual intervention, then you can "ride" the GREY PLUS key to simply skip averaging during the rough spots, and resume when things look better. Tedious, perhaps, but better than no data at all. To use this method, you don't want to view the average in progress while you are watching for artifacts since they will not be very evident when averaged with a lot of good data. Instead, you must view the raw instantaneous input using the backslash (\-key).


Sometimes you may need to be near the experiment in progress to determine when conditions are suitable for averaging. If your experiment is not conveniently located next to your computer keyboard, a separate digital input line can be used as an averager remote control to replace the Grey Plus skip key. You can thus have a hand-help switch on the end of a long cable, so you can turn the averager on and off from the experiment location. For tips on constructing a manual remote control, see Remote Control Switch Construction.


If you are unfortunate enough to be plagued by artifacts that can't be detected by simple thresholds, you might still get lucky (or tricky!) and devise a scheme to detect them by some other means. If you can provide Daqarta with a TTL signal to indicate the presence of artifacts, you can use the Averager Remote Control to effectively ride the Grey Plus skip key automatically. For example, you might have some sort of motion sensor that indicates when a subject is moving around, and use that as a cue to probable muscle activity artifacts. Your cleverness in devising an appropriate scheme may make the difference between data and no data.

CAUTION: Do not connect any electrical equipment to a living subject without proper signal isolation techniques. A lethal shock could result.


Sometimes you may just want to monitor an experiment or process without having discrete averages that must be started and stopped. But if the signal has a lot of noise, you can't really see what's going on. Exponential averaging provides a continuous average which looks like the normal averager output, only "live". It is a weighted average of multiple Sweeps, with recent data having more influence on the result and very old data being ignored. In this way, you can get a better view of the desired signal, at a cost of slow response to a sudden change.

The Sweeps value in Exponential mode is thus a time constant that tells not only how much averaging is going on, but also how slowly the results can respond to changes. Low values give quick response but less averaging effect.

Notice that when you first begin an average in this mode, the "exponential" effect can't start until the selected Sweeps have arrived, so initially the display is just like True average mode. The sweeps counter will show 'nn Wave', where 'nn' will count up from 0 just as in True. But when the selected Sweeps value is reached, the counter will stop increasing and the 'Wave' will change to 'WavX' to show that Daqarta is in eXponential operation.

Since the Exponential mode doesn't stop when it reaches the selected Sweeps, you leave this mode only by toggling Avg off manually.


Questions? Comments? Contact us!

We respond to ALL inquiries, typically within 24 hrs.
25 Years of Innovative Instrumentation
© Copyright 1999 - 2006 by Interstellar Research