In this hyperscanning series, we first introduced our gold standard solution based on BrainAmps, then a wireless solution using CGX Quick headsets and the StimTrigger, as well as the extremely flexible LSL solution. In this fourth part, we demonstrate another hardware-based option: using the trigger mirror mode which can be used with actiCHamp Plus and LiveAmp together with the Sensor & Trigger Extension.
Looking at all of the preceding articles, you may be left wondering: “What if I just want to combine amplifiers via trigger cables?”, and there is indeed a great option available to do this: the trigger mirror mode. This mode is available for actiCHamp Plus and LiveAmp, given that you have a Sensor & Trigger Extension available. That said, we will focus on the actiCHamp Plus, as it is particularly well suited for this type of hyperscanning, especially if you want to combine more than two amplifiers.
Connecting a series of amplifiers with forwarded triggers
If you look at the back of the actiCHamp Plus and the front of the Sensor & Trigger Extension for LiveAmp, you will see two 9-pin D-SUB connectors, one labeled Trigger In, one labeled Trigger Out. While the Trigger In port is frequently used, the Trigger Out port serves some interesting purposes, too: For actiCHamp Plus, you can choose between Trigger Mirror Mode and the Trigger Dialogbox. For LiveAmp (connected to a Sensor & Trigger Extension), you can choose between Mirror Trigger In and Sync Out, where triggers are sent with an adjustable frequency. You can activate these features in BrainVision Recorder, under Amplifier > Digital Port Settings. See Figure 1 and Figure 2 for actiCHamp Plus and LiveAmp, respectively.
Figure 1: For the actiCHamp Plus, you can enable Trigger Mirror Mode as one use case of the Trigger Out Port.
Figure 2: Enabling Mirror Trigger In for LiveAmp with Sensor & Trigger Extension. The Trigger Out Port on the Sensor & Trigger Extension can be used to either forward incoming triggers or to send a synchronization signal with adjustable frequencies.
These options give you some useful means for testing, for example as a way to generate triggers with the Dialogbox, or to send reoccurring markers via Sync Out that can be used for later synchronization with other data sets. However, the Trigger Mirror Mode (or Mirror Trigger In) is particularly useful when you want to use the same trigger input for subsequent amplifiers. In theory, these amplifiers can even be different devices, as long as amplifiers prior to the last in the chain have this trigger forwarding functionality.
After Trigger Mirror Mode/Mirror Trigger In has been selected in BrainVision Recorder, simply take a 9-pin D-SUB male-to-female cable and connect one end to the Trigger Out port of the first amplifier and the second end to the Trigger In of the subsequent amplifier. Repeat, if necessary, by adding amplifiers to the end of the chain with additional cables.
Daisy chaining multiple actiCHamp Plus amplifiers with minimum trigger delays
We already mentioned that we want to focus on the actiCHamp Plus, and there is a good reason: triggers are forwarded to the next amplifier within very short time frames. The actiCHamp Plus offers exceptionally high base sampling rates. Even when you save data with a nominal sampling rate of 500 Hz or 1 000 Hz, the base sampling rate in the background is much higher. Specifically: it is 10 000 Hz for all the possible sampling rates below and equal to this value (100, 200, 250, 500, 1 000, 2 500, 5 000, and 10 000 Hz). For higher sampling rates, the base sampling rates even increase up to 50 000 Hz or 100 000 Hz1. These high base sampling rates play an important role in the delay you can expect until a trigger arriving at the Trigger In is forwarded on the Trigger Out port. In fact, this delay is only dependent on the base sampling rate, not the nominal sampling rate. You can expect a maximum delay of one sample, which is 100 µs for 10 000 Hz, 20 µs for 50 000 Hz, and 10 µs for 100 000 Hz. This means that you do not have to worry about noticeable delays of triggers that increase with the length of the chain. It is unlikely that there will even be one sample of delay if you use standard sampling rates of around 1 000 Hz or 5 000 Hz.
1 Please note that the maximum sampling rate depends on the number of channels, i.e., 100 000 Hz for up to 32 EEG channels, 50 000 Hz for up to 64 EEG channels, and 25 000 Hz for more than 64 EEG channels.
Combining actiCHamp Plus with other devices
Hyperscanning setups do not necessarily have to consist of only the same amplifiers. Maybe the experiment is not symmetrical, in as much as not everyone is doing the exact same thing under the exact same conditions. For a two-amplifier setup, it may be okay that the first amplifier is an actiCHamp Plus and the second one is a LiveAmp, for example. While using identical hardware is generally recommended, using different hardware across participants is possible. You just have to make sure that the second amplifier can receive triggers. In theory, this can be expanded infinitely, as long as all amplifiers from the first to the penultimate are actiCHamp Plus devices.
Using trigger forwarding with LiveAmps
For LiveAmp, however, the delay may become noticeable. As there are no such high base sampling rates behind the LiveAmp’s nominal sampling rates, the delays of forwarded triggers are limited by one sample of the nominal sampling rate, i.e., either 250, 500, or 1 000 Hz. This translates to 4, 2, or 1 ms, respectively. While it is still possible to use a number of daisy chained LiveAmps, it may be better to consider other hyperscanning solutions, such as one based on LSL alone, unless a potential delay of several milliseconds is inconsequential for your particular setup.
Why not just use a y-cable?
You could also directly split the cable coming from the trigger source and link it to all the receiving amplifiers. However, the trigger mirror mode is more flexible, as you can simply add more and more amplifiers as you like and reduce the number again as easily. For a y-cable, you would always have to add more and more branches which is difficult to undo when the number of receivers is reduced again. Since the amplitudes of the trigger signals are attenuated with every additional branch, the number of amplifiers you can connect to one source is also limited.
Combine the convenience of LSL with the precision of hardware triggers
If you use this setup with BrainVision Recorder, you will end up with as many recordings as participants (or daisy chained amplifiers). All will have the hardware triggers included and you can use those to synchronize the individual signals offline. However, you can also use LSL connectors and LabRecorder instead. Please check our bci.plus blog post to learn more about using LSL with Brain Products amplifiers.
The LSL connectors on our GitHub page have the option to activate mirror mode similar to the settings in BrainVision Recorder. Figure 3 and Figure 4 demonstrate how to select the mirror mode for actiCHamp Plus and LiveAmp.
Figure 3 (left): The actiCHamp LSL connector (minimum version 1.15). Select Mirror Input under Hardware Trigger Output Mode to activate the forwarding of triggers for your actiCHamp Plus. Figure 4 (right): For LiveAmp, triggers can be forwarded only if a Sensor & Trigger Extension is connected. In this case, you can select Mirror under STE Out Mode to activate the feature.
There are two more advantages you can benefit from when using LSL connectors
– at least for actiCHamp Plus:
Instead of separate recordings, you have all the data combined in a single XDF file. Again, check our blog post or our article Hyperscanning series part 3 article for more information on how to deal with XDF files and stay tuned for the next issue of our newsletter to learn more about how to merge signals from multiple sources in BrainVision Analyzer 2. In addition to the convenience of having all your signals in one file, you can still rely on the hardware triggers.
In the LSL connector for actiCHamp, you can freely set the Base Sampling Rate (Figure 3). Usually, this rate cannot be directly set, but in the LSL connector, you can set it to higher than usual values and use a higher Sub Sample Divisor to arrive at the same Nominal Sampling Rate that you would normally use. This way, you may even further reduce the possible delay of 100 µs while still using sampling rates lower than 10 000 Hz.
All of these considerations are visualized in Figure 5. We demonstrate such a setup by using a series of actiCHamp Plus amplifiers, as well as an arbitrary trigger-receiving amplifier at the end of the chain that can either be used with BrainVision Recorder or LSL connectors and LabRecorder.
Figure 5: Demonstration of the flexible setup. In theory, the number of amplifiers (n) is not limited, but we recommend that all but the last of those is an actiCHamp Plus, represented by aC1 and aCn-1, (n-1 is only relevant for more than 2 amplifiers). The final trigger receiver can be any amplifier that has a trigger input port. Using LiveAmps with Sensor & Trigger Extensions is possible, but here you can expect higher delays based on the nominal sampling rate. Amplifiers can be controlled directly in BrainVision Recorder or with their respective LSL connectors to send LSL streams to the network which can be recorded in LabRecorder. The result is either multiple recordings in the BrainVision data format or a single XDF file comprising all the individual LSL streams. Either way, you will be able to make use of hardware triggers.
Hyperscanning based on trigger forwarding presents a flexible solution where you can easily change the number of involved amplifiers and where you can choose what recording software you want to use. The maximum delay of forwarded triggers is below the time between single samples, which is especially useful for the actiCHamp Plus which operates with an internal sampling rate of at least 10 000 Hz.
As always, please don’t hesitate to reach out to us if you have any questions.