A “how-to” guide on combining Brain Products equipment for Sport Psychophysiology
by Eduardo Bellomo, Ph.D.
Scientific Consultant (Brain Products)
Have you ever wondered what happens in the brain and in the body of an athlete in the moments before performing their sport skills? Or how these processes might differ in experts as opposed to novices? Or how performance changes in high pressure situations?
Well, these are the questions that sport psychophysiology attempts to answer by measuring EEG and peripheral physiology while participants perform a sport skill.
However, you might be also asking yourself: “ok, fascinating … but … how do I do it? How do I get markers for my event-related analysis? How do I combine different types of signal? What equipment do I need?”
In this article, we provide you with an example on the “how-to” of acquiring and analysing sport psychophysiology data. Given the potential complexity of the topic, we decided to focus on the concrete example of golf putting (considering its popularity in the field1). However, our hope is to provide you with a general idea that you can re-shape and adapt to other sport movements that are of interest to your research question.
1 In addition to golf putting, other popular sporting skills in the literature are pistol/rifle shooting (Hatfield et al., 2013), dart throwing (Cheng et al., 2005), archery (Landers et al., 1994), and basketball free-throwing (Chuang et al., 2013).
Overview
Things to keep in mind
In order to perform a sport psychophysiology investigation, it would be best to have a hardware configuration that can measure different types of signals and combine them in the same data file. Specifically, you might need to measure EEG together with other types of physiological signal (e.g., EMG, ECG) as well as additional measures which would be used to derive event markers.
In fact, sport skills are arguably different from computer-based tasks and if the goal is to measure event-related activity, it is necessary to use a bit of creativity in order to identify an objective event marker for movement initiation. Depending on the sport / movement / motor skill under scrutiny, these objective event markers could be identified by triangulating information from additional measures such as accelerometers, photosensors, or microphones. The catch is that if different types of hardware (and software) are used to acquire these additional types of measures, the signal might not be precisely synchronized (due to different internal clocks and sampling rates).
The good news is that with Brain Products’ hardware, it is possible to have a complete solution for this purpose; thereby allowing researchers to focus more on the scientific aspect and their research question rather than the setup.
What hardware do I need?
In our example, we are interested in studying EEG, EMG, and ECG, in the moments prior to the initiation of the golf putt (i.e., backswing). In order to measure all the signals needed for our research question and our identification of clear event markers, we recommend the following configuration:
| Component name | Brain Products solution |
|---|---|
| Wireless EEG amplifier | LiveAmp (32 or 64 channels) |
| Electrode cap | actiCAP slim/snap |
| Box for additional signals | Sensor and Trigger Extension (STE) and external USB power bank |
| Microphone (2 alternatives) | – StimTrak (built-in microphone) – StimTrak, Acoustical Stimulator Adaptor for StimTrak and third-party microphone |
| Photosensor | Photo Sensor |
| Accelerometer | 3D Acceleration Sensor |
| Bipolar adaptor for additional peripheral physiology (e.g., EMG or ECG) |
BIP2AUX |
| Data analysis platform | BrainVision Analyzer |
This configuration has the advantage of being fully mobile, wireless, and wearable. In fact, as shown in the overview below, the LiveAmp, the STE, and the power bank can be comfortably worn on a common tactical padded belt. Data would be streamed to the computer running the data acquisition software via Bluetooth (~10m range indoor or ~30m range outdoor), and/or stored directly on the built-in SD card. The LiveAmp itself would measure the EEG signal, whereas the STE would measure additional physiological channels as well as the additional signals necessary for the event marking. Please note that here we describe only a sub-section of the sensors available in the Brain Products portfolio. For a full list and detailed information (device properties, practical tips, data acquisition settings, and references) check our Sensor Tutorial.
Which event marker do I use?
In sports, the actual event is less obvious and clear-cut than with computer-based paradigms. Therefore, it is crucial to first think about what your event marker should be. This process must be tailored to the sporting skill that you plan to study.
In the case of golf putting, you can try to break down the skill into several components: the performer must first address the ball, and whilst keeping the putter still, adjust stance and grip, and then plan how to execute the shot in order to get the ball in the desired location. Once everything is ready, the performer would initiate the backswing, follow-through, and finally hit the ball. Thus, arguably, the reference event in golf putting is backswing initiation. So now the question is: how can we identify backswing initiation? Well, we can use an empirical approach. This involves triangulating measures from the following devices: microphone (StimTrak and Acoustical Adaptor), Photo Sensor, and 3D Acceleration Sensor. Based on these measures, it is then possible to identify with certainty the event marker during the offline data analysis in BrainVision Analyzer, as described later in this article.
Please note that signals from different sensors might not be measured in the same units as EEG, and therefore, it might be necessary to change the workspace settings (for 3D Acceleration Sensor – check the box “Diff. Unit”, write “g” as “Unit”, 1450 as “Gradient” and 0 as “Offset” in the channel table) or enable the Scientific View in BrainVision Recorder (to visualize them with the correct scaling).
As shown in the setup overview image above, the 3D Acceleration Sensor and the Photo Sensor are used to mark the transition from stillness (when addressing the ball) to movement (backswing initiation). The 3D Acceleration Sensor is mounted on the back side of the putter and tracks the XYZ kinematics of the putter; the Photo Sensor is placed vertically to be aligned with the putter during the addressing phase; the microphone would be used to mark the ball hit (see the blue dashed boxes in the “Overview of the setup” image). Depending on the environment and the skill, it would be either possible to use the built-in microphone of the StimTrak or a 3rd party professional microphone combined with the Acoustical Stimulator Adaptor and the StimTrak.
You might be asking yourself, why do we need all these sensors? Well in an ideal world, the participant would line up the putter to the Photo Sensor while addressing the ball, and since he/she would keep a relatively still stance, the accelerometer signal would stabilise. Then, at backswing initiation, the putter would move backwards starting a chain reaction: the Photo Sensor would capture the change in light triggered by the club head’s movement, the accelerometer would track the backward movement (very evident on the Z-axis) as well as the subsequent follow-through (this movement has a very typical footprint as shown in Figure 1), and the microphone would capture the sound of the putter hitting the ball.
This is what we expect in most of the cases; however, in some trials, something might go somewhat differently. For example, the participant might forget to address the ball in a way that the putter is truly in line with the Photo Sensor; or when adjusting the stance, might accidentally tilt the putter and make the Photo Sensor go on and off a few times consecutively; or after initiating the backswing, might decide to stop and go back to the addressing position. Therefore, in these specific cases, it is really advantageous to have all the three sources of information (putter-acceleration, photosensor, and microphone):
Figure 1. The image shows the footprint of a putt in terms of 3D acceleration (specifically, the Z-axis represents the back [negative values] and forth [positive values] movements), microphone (impact of the club with the ball), and photosensor (putter moved away from addressing position). For each of these channels, a marker was generated with dedicated functions in Analyzer, as described in the main text.
1. Microphone marker
The microphone information can be used in order to pre-select the portions of data where the sound of a ball-hit is detected. This can be done by identifying the time point at which the sound signal goes beyond a certain threshold. In BrainVision Analyzer, you could rectify the Microphone signal in order to make this thresholding operation easier (Transformations > Dataset Preprocessing > Rectify). Then, use the Level Trigger (Transformations > Dataset Preprocessing > Level Trigger) transformation to detect when the signal goes beyond a specific amplitude threshold and then set a “microphone” marker (ball-hit). Since ball-hit happens a few milliseconds after the backswing initiation, perform a first segmentation (e.g., -5000 to +2000) to isolate the data points that include the data of interest (preparation, backswing initiation, and ball hit). The segments should correspond to the number of putts (trials) you anticipated in your paradigm. Since the microphone might pick up additional sounds on top of the ball impact, it would be better to go through each segment and look for the putt footprint of the microphone, photosensor, and Z-axis acceleration (as shown in Figure 1). For this operation, you can use the Artifact Rejection function (Artifact Rejection/Reduction > Artifact Rejection > Manual) and mark the invalid trials, so that Analyzer will ignore them in the future steps.
2. Photosensor markers
The next step would be to apply a Baseline Correction (Segment Analysis Funtions > Baseline Correction > -4000 to -1000) to adjust the voltage offset of each trial and then apply the MinMax Markers (Solutions > Markers > MinMax Markers) solution on the photosensor channel around ball-hit (e.g., -1500 to +500), in order to work out when the putter was first moved backwards and set MinMax photosensor markers.
3. Acceleration sensor maker
At this stage, we would have Microphone and MinMax photosensor markers and are ready to look at the putter-acceleration data in order to identify our backswing marker. As shown in Figure 1, the Z-axis shows the most prominent change in direction around the data points where we find our Microphone and MinMax photosensor markers. In order to identify this last marker, we recommend a semi-automatic procedure. First, apply the MinMax Markers solution on the Z-axis and on the data points preceding the ball hit (e.g., -1000 to -300), in order to detect the maximum value before an acceleration in the backwards direction (negative deflection). You can focus on the Max marker and you can call it Max-Z. Then use the transformation Edit Markers (Transformations > Dataset Preprocessing > Edit Markers) and select the Graphical option, which will enable you, eventually, to drag and drop it in your desired position.
The result of this procedure is to identify a series of Max-Z markers, which correspond to the backswing initiation marker. Now, we are ready to analyse the data (see Data analysis options paragraph).
Peripheral physiology (e.g., EMG and ECG)
Together with EEG, you might be interested in collecting additional physiological information. Again, this may be specific to the paradigm and sport. As for our example, in golf putting, it is common to additionally measure ECG signal (e.g., experts vs novices show a steeper event-related heart rate deceleration prior to swing onset; Cooke et al., 2014), and EMG from muscles of the non-dominant arm (e.g., experts vs novices show a firmer grip prior and during putt execution; Cooke et al., 2014).
Since these signals are measured with a bipolar configuration, it is possible to use the BIP2AUX, which is a small preamplifier that allows you to record bipolar signals with classic passive electrodes (Multitrodes). It plugs into the STE via the AUX port. If ECG and EMG from two muscles are to be measured, then x3 BIP2AUX are needed. When using more than one BIP2AUX, only one GND electrode is needed (since it would be internally shared in the STE).
With ECG, for example, you could place the GND electrode on the left clavicle, and the other two electrodes on the right clavicle and on the lowest left rib. For EMG, you could place the electrodes respectively on the flexor and extensor carpi radialis of the non-dominant arm. Before applying the sensors, make sure to clean the skin underneath the passive electrodes with an abrasive gel with high chloride content.
In order for the signal to be displayed and acquired in the correct unit in BrainVision Recorder, you should also make sure to change the workspace settings for each specific BIP2AUX channel (check the box “Diff. Unit”, write “µV” as “Unit”, 0.1 as “Gradient” and 0 as “Offset”).
Moreover, for a better signal visualization, we recommend applying a high-pass display filter at 20Hz for EMG and a low-pass for ECG. Additional settings, such as sampling rate and filters depend on the experimental requirements.
Data analysis options
Once all the backswing (Max-Z) markers have been identified, you are ready to proceed with the next steps of your analysis.
As for the EEG, the most common types of techniques reported in the literature are time-frequency analyses such as ERSD (event-related synchronization/desynchronization), Wavelets, or Connectivity, for which BrainVision Analyzer has dedicated transformations. As for the peripheral physiology sensors, Analyzer contains basic transformations and solutions for such purposes, which are described in the respective online articles for ECG and EMG.
As a bonus track, from the 3D Acceleration Sensor, it is also possible to derive information on the putter kinematics. Although Analyzer does not have dedicated transformations for this analysis, you could still pre-process the data as described in this online article and then export the data to a third-party platform (e.g., MATLAB®) either via the Generic Data Export module (Export > Node Export > Generic Data) or the Create MAT File (Solutions > Matlab > Create MAT File).
Conclusion
We hope that this article was helpful for you and that you now have an idea of what you need to start planning a sport psychophysiology investigation by combining different hardware from the Brain Products portfolio. As mentioned throughout this article, the actual type of hardware and the way event-markers are identified will need to be tailored to the specific sporting skill, so we hope that we have inspired you to find a configuration in order to fit your unique needs.
If you do so, do not forget to cite us in your next publication so we can see how you are using our products in sport psychophysiology research.
References
Cooke, A., Kavussanu, M., Gallicchio, G., Willoughby, A., McIntyre, D., & Ring, C. (2014).
Preparation for action: Psychophysiological activity preceding a motor skill as a function of expertise, performance outcome, and psychological pressure.
Psychophysiology, 51(4), 374-384.Chen, J., Hung, T.‐M., Lin, J. H., Lo, L. C., Kao, J. F., Hung, C.‐L., … Lai, Z.‐S. (2005).
Effects of anxiety on EEG coherence during dart throw.
In T. Morris & P. Terry (Eds.), 11th World Congress of Sport Psychology (pp. 2–5). Sydney, AUS.Chuang, L. Y., Huang, C. J., & Hung, T. M. (2013).
The differences in frontal midline theta power between successful and unsuccessful basketball free throws of elite basketball players.
International Journal of Psychophysiology, 90, 321-328. https://doi.org/10.1016/j.ijpsycho.2013.10.002Hatfield, B. D., Costanzo, M. E., Goodman, R. N., Lo, L.-C., Oh, H., Rietschel, J. C., … Haufler, A. (2013).
The influence of social evaluation on cerebral cortical activity and motor performance: a study of “Real-Life” competition.
International Journal of Psychophysiology, 90, 240–9. http://doi.org/10.1016/j.ijpsycho.2013.08.002Landers, D. M. Han, M., Salazar, W., & Petruzzello, S. J. (1994).
Effects of learning on encephalographic and electrocardiographic patterns in novice archers.
International Journal of Sport Psychology, 25, 313-330.