Linköpings universitet
Linköping University | Department of Biomedical Engineering
Master thesis, 30 ECTS | Medicinsk Teknik
2017 | LiTH-IMT/BIT30-A-EX--17/544--SE
Identifying patterns in
physi-ological parameters of expert
and novice marksmen in
sim-ulation environment related
to performance outcomes
Johanna Karlsson
Supervisor : Tuan Pham Examiner : Göran Salerud
Upphovsrätt
Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under 25 år från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår. Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och admin-istrativ art. Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sam-manhang som är kränkande för upphovsmannenslitterära eller konstnärliga anseende eller egenart. För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/.
Copyright
The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circum-stances. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the con-sent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement. For additional information about the Linköping Uni-versity Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.
c
Abstract
The goal of this thesis is to investigate if it is possible to use measurements of physio-logical parameters to accelerate learning of target shooting for novice marksmen in Saab’s Ground combat indoor trainer (GC-IDT). This was done through a literature study that identified brain activity, eye movements, heart activity, muscle activity and breathing as related to shooting technique. The sensors types Electroencephalography (EEG), Electrooc-culography (EOG), Electrocardiogram (ECG), Electromyography (EMG) and impedance pneumography (IP) were found to be suitable for measuring the respective parameters in the GC-IDT.
The literature study also showed that previous studies had found differences in the physiological parameters in the seconds leading up to the shot when comparing experts and novices. The studies further showed that it was possible to accelerate learning by giv-ing feedback to the novices about their physiological parameters allowgiv-ing them to mimic the behavior of the experts.
An experiment was performed in the GC-IDT by measuring EOG, ECG, EMG and IP on expert and novice marksmen to investigate if similar results as seen in previous stud-ies were to be found. The experiment showed correlation between eye movements and shooting score, which was in line with what previous studies had shown. The respiration measurement did not show any correlation to the shooting scores in this experiment, it was however possible to see a slight difference between expert and novices. The other measure-ments did not show any correlation to the shooting score in this experiment. In the future, further experiments needs to be made as not all parameters could be explored in depth in this experiment. Possible improvements to such experiments are i.e. increasing the num-ber of participants and/or the numnum-ber of shots as well as marking shots automatically in the data and increasing the time between shots.
Acknowledgments
Thank you to Mats Bengtsson, Henrik Lundgren and Jerker Andersson at Saab Training and Simulation for giving me the opportunity to do this project and providing equipment to make it possible. A special thanks go to my supervisor Moa Hellberg for the encouragement and support in the project, as well as Amanda Hansson who were a great help and sounding board throughout the experiment. Another person who deserve a thank you is my examiner, Göran Salerud, who has let help me shape this project as well as let me change the shape when the project took a new direction. I also thank everyone that took their time to participate in the experiment.
Acronyms
BPM
Beats per minute
EEG
Electroencephalography
ECG
Electrocardiography
EMG
Electromyography
EOG
Electroocculography
FIR
Finite impulse response
fMRI
Functional magnetic resonance imaging
GC-IDT
Ground combat indoor trainer
HRV
Heart rate variability
IIR
Infinite impulse response
IP
Impedance pneumography
MAP
Multi-active plan
MEG
Magnetoencephalography
PCA
Principle component analysis
PPG
Photoplethysmography
Contents
Abstract iii
Acknowledgments iv
Acronyms iv
Contents vi
List of Figures viii
List of Tables ix 1 Introduction 1 1.1 Aim . . . 2 1.2 Research questions . . . 2 1.3 Delimitations . . . 2 2 Theory 3 2.1 Shooting technique and training . . . 3
2.2 The brain . . . 5
2.3 The eyes . . . 8
2.4 The heart . . . 10
2.5 The muscles . . . 12
2.6 The lungs . . . 13
3 Material and method 14 3.1 Literature study . . . 14 3.2 Experiment setup . . . 15 3.3 Data analysis . . . 19 4 Results 21 4.1 Literature study . . . 21 4.2 Experiment results . . . 31 5 Discussion 46 5.1 Literature study . . . 46 5.2 Experiment . . . 48
5.3 Material and method - Literature study . . . 49
5.4 Material and method - Experimental setup . . . 49
5.5 The work in a wider context . . . 51
5.6 Ethical aspects . . . 51
6 Conclusion 53 6.1 Future work . . . 54
Bibliography 55 A Appendix A 59 B Appendix B 61 C Appendix C 63 D Appendix D 65 E Appendix E 68 E.1 Task B . . . 68 E.2 Task C . . . 71
List of Figures
2.1 The MAP-model . . . 4
2.2 An example of the 10-20 system for EEG electrode placement . . . 7
2.3 The human eye . . . 8
2.4 Principle look of a heart beat in a ECG signal . . . 11
3.1 Placement of ECG and respiration measuring electrodes . . . 17
3.2 Placement of EMG measuring electrodes . . . 17
3.3 Placement of EOG measuring electrodes . . . 17
4.1 PCA for eye movements. . . 23
4.2 Example of missed heart beats . . . 31
4.3 Eye movements for each individual in task A. . . 32
4.4 Eye movements for each individual in task D. . . 33
4.5 Eye movements for each individual in task E. . . 33
4.6 PCA for eye movements. . . 34
4.7 Standard deviation of EOG gradient . . . 35
4.8 Respiration for each individual in task A. . . 36
4.9 Respiration for each individual in task D. . . 36
4.10 Respiration for each individual in task E. . . 37
4.11 PCA for respiration. . . 38
4.12 Standard deviation for respiration measurement . . . 38
4.13 Normalized muscle activity for each individual in task A. . . 39
4.14 Normalized muscle activity for each individual in task D. . . 40
4.15 Normalized muscle activity for each individual in task E. . . 40
4.16 PCA for muscle activity . . . 41
4.17 Standard deviation of muscle tension gradient . . . 42
4.18 Beat to beat heart rate for each individual in task A. . . 43
4.19 Beat to beat heart rate for each individual in task D. . . 43
4.20 Beat to beat heart rate for each individual in task E. . . 44
4.21 RR-intervals for different scores . . . 45
E.1 Eye movements for each individual in task B. . . 68
E.2 Respiration for each individual in task B. . . 69
E.3 Normalized muscle activity for each individual in task B. . . 69
E.4 Beat to beat heart rate for each individual in task B. . . 70
E.5 Eye movements for each individual in task C. . . 71
E.6 Respiration for each individual in task C. . . 71
E.7 Normalized muscle activity for each individual in task C. . . 72
List of Tables
3.1 Description of tasks . . . 18
4.1 Summary of brain activity studies . . . 24
4.2 Prices for different EEG equipment. . . 27
C.1 Shootings scores . . . 63
1
Introduction
When a person shall learn and improve how to shoot, the training consists of giving the person a weapon and let them shoot on a target over and over until they learn reach a certain level. The trainee get general instruction from instructor on the general technique on how to hold the weapon, how to breath, where to look and so on. Then it is up to the trainee to process all this information and conceive what to use and how to alter their technique to make it work for them. This type of training take a lot of time and the trainees might not be aware of errors they do and thus not know how to change them. One way to aid the training processes is to monitor aspects of the technique that are known to affect the performance, using a suitable sensor. The data from such sensors can then be used by the trainee and make him/her aware of the errors in his/her technique. By making him/her aware of errors the trainee can change his/her behavior faster and more efficiently than if they have to identify the problems him- or herself.
Today the division Saab Training and Simulation, have some integrated sensors in their weapon simulators. These sensors are attached to the weapons used in the simulator and can measure the trigger pull, pressure against the shoulder and tilt of the weapon [29]. The result of the sensors can be seen either by a coach during the training or by the trainee after the training session has completed. This give the trainee useful information about what might be the problem if they do not perform optimally. By making them aware of the problems they can start working on it and hopefully learn the proper technique quicker.
Even though these sensors give useful information they do not cover all the aspects of the shooting technique training. There are more parameters that are of interest when learn-ing to shoot and some of them are associated with the human body. Examples of such parameter are respiration, focus of eyes, muscle tension and mental state. It would therefore be of interest to integrate sensors that can measure these parameters in combination with the sensors that are already existing in the simulator.
To be able to measure these parameters it is necessary to identify which sensors that would be suitable to integrate in the simulator. This means that the sensor should be safe, easy to use and not be to sensitive to movements or external events. It is also important to identify what the result from the measured parameter actually mean, so that the trainee can get useful
1.1. Aim
information from the results and use this to improve their performance. To do this, a study to identify what has previously been done in other studies will be performed. The study will focus on if patterns in the measured parameters can be seen when comparing people of different skill level.
If such patterns can be identified in the data collected from the skilled marksmen, it can hopefully be used by novice marksmen in the same way as the results from the existing sensor in the simulator are used.
1.1
Aim
The aim of this thesis is to investigate how to increase efficiency of shooting training in a simulator using measurements of physiological parameters. The aim of this thesis thus cover, to identify, with help from literature, one or more physiological parameters that are related to the performance of a marksman and sensors that are suitable to measure these parameters in the simulator environment. The aim is also to investigate what has been previously done in the area by conducting a literature study, as well as performing an experiment to investigate however the sensors found suitable to use in the simulator environment is indeed that.
1.2
Research questions
This project aim to answer the following research questions:
1. What physiological parameters, such as brain activity, eye movements, heart rate, res-piration and muscle activity, are related to learning and improvement of shooting tech-nique?
a) What type of sensors might be suitable to measure these parameters? 2. What kind of studies have been done previously in this area?
3. Is it possible to distinguish between different levels of performances, done by a marks-man, just by looking at the result from one of these sensors?
a) Are the results generalizable over several marksmen?
4. What alternatives are there to use the results to increase the training efficiency?
1.3
Delimitations
This is a project directed to investigate learning in marksmen, the results are intended to be used for this group and will not make any investigations into other types of training or target groups. The project has limited the focus of the experiment to only include the standing shooting position.
2
Theory
2.1
Shooting technique and training
When a person is learning a new task there are three stages [8]. Firstly, there is the stage of learning. This stage require a lot of mental processing, for the trainee. He/she has to coor-dinate all aspects of the technique and combine it with information given by coaches as well as process environmental information [8]. The second stage is associated with repetition or practice. At this stage the trainee is familiar with the general technique and parameters that affect the performance, there are still ways to improve but it is more related to refinements of the technique by repetition [8]. After some time of training the trainee get to the third and final stage of learning. This stage is called the automatic stage [8]. In this stage the trainee performs the task more or less automatically and can repeat good results over and over. There are some general things associated with learning to shoot, as almost any move-ment of the body or the rifle affect the performance, it is important to learn a technique that is consistent over several shots [8]. To keep the body as still as possible the general guidelines is to make sure that the balance is good by activating the muscles to counteract body swaying and rifle movements [33].
Another aspect of the shooting technique regards the eyes. As eye movements can af-fect the body balance and thus the eyes should be kept fixated in when taking a shot [33]. The breathing is another body function that creates movement of the rifle and thus the technique taught is to hold the breath while taking the shot [34].
Other body functions that are not directly related to the body stability but still can affect the outcome of the shot is the focus in the brain and heart activity [33].
2.1.1
Focus of attention
Attention is when a person direct their mind to a specific task or sensation. It can for example be to listen to someone talking or looking at something. When learning a new task attention has to be directed to the task, but there are several types of attention and research show that not all give a equally good performance [40]. Generally, an external focus of attention has
2.1. Shooting technique and training
proven to be the most effective when performing a movement [40]. External focus of atten-tion mean that the attenatten-tion is on the effect of the task at hand. Internal focus of attenatten-tion is when the focus is on how the movement is performed and how it is perceived in the body. Wulf [40] mention that studies show the results are often better when the external attention is focused far away from the person rather than near, such as the bulls eye in dart throwing versus dart flight. Even though a lot of studies show that the external attention is superior when it comes to performance, Wulf [40] point out that instructions from coaches often are related to internal attention.
In a study made by Bortoli et al. [4], a group of elite shooter got to estimate their con-trol level over some core components. From the results Bortoli et al. created a model, called multi-active plan (MAP) that was divided into four "Types". An illustration of the MAP-model is shown in figure 2.1 below. Depending on the marksman’s estimated control level and the shooting score, the shoot was said to be one of the four types. With this they showed that there are two types of performance that are considered optimal in relation to the shoot-ing score but differ in the control level. Type 1 was defined as an optimal result and a low control level, which was describes as a "flow"-state for the shooter where they experience that everything just works without effort [4]. In Type 2 the result was still optimal but the control level over the core components were higher, this was describe as "Plan B"-state because it requires more effort than type 1 performance but it still gave optimal results. In the two other types the performance were sub-optimal and the control level were high for type 3, which was assumed to be related to a too high level of focus on the core components or that the focus were on task irrelevant information. For type 4 the control level was low and that was thought to be because the focus was just not there.
Figure 2.1: The MAP-model. There are four different types of performance in the MAP-model, type 1 which means that the performance is optimal while the control level is low, type 2 which means that the performance is optimal but the control level is high. Type 3 means that the performance is sub-optimal and the control level is high while type 4 means that the performance is sub-optimal and the control level is low1.
With the MAP-model as a guide a study was made to see if the model could be used as a guiding tool for non-elite athletes to help them perform optimally. This study was made by Bertollo et al. [3] and involved giving a group of cyclist directions on what they should focus on. They gave the instructions to the participants for the different types they wanted to achieve. For type 1 they asked the participants to focus on a metronome that reproduced their preferred pedaling rate (PPR). To achieve type 2 they asked the participants to focus their attention on the PPR and to achieve type 3 performance they asked the participants to
2.2. The brain
focus on errors and feelings of fatigue. Whit this study Bertollo et al. [3] showed that the cyclist that where asked to focus their attention on either external components of the task, i.e the metronome or internal task-relevant components, i.e their PPR, were able to perform for a longer time before exhaustion than when they were asked to focus on feelings and task-irrelevant components.
2.1.2
Ground combat indoor trainer (GC-IDT)
Saab has a simulator for training of shooting techniques and tactical aspects of marksman-ship. This simulator is called Ground combat indoor trainer (GC-IDT). The GC-IDT simulator enable marksmen to train in a safe environment. In the simulator the weapons used are made to be as close replica as possible to make the training as realistic as possible [29]. The weapons do not however use real ammunition, instead the weapons "fire" invisible laser light on a projected target and a camera register the result [29].
On the weapon there are sensors that measure trigger pulling, canting and pressure against the shoulder [29]. The information from these sensors are at the present used to give verbal feedback to the trainees after they finished a training session.
2.2
The brain
The brain is extremely complex and this section does not intend to cover everything about the brain, rather just a small description of what is relevant to know when reading this thesis. The human brain consist of several billions of neurons [38]. Neurons are cells that can send information form one part of the brain, or in the body, to another by generating an electrical pulse that move along a part of the neuron that is called the axon. The neurons are connected to each other with dendrites and synapses [5]. The dendrite on one neuron receive signals from the synapse of another neuron and pass it along the axon to its synapses. Each neuron can be connected to several hundred other neurons by their synapses. This creates a network of connected neurons that together form the brain and the nervous system.
The information is as mentioned sent by electrical pluses along the axon of the neuron [38]. This process is called an action potential. To generate an action potential there has to be a stimulus from another nerve cell that change the potential over the cell membrane on the neuron [5]. The stimulus comes from the synapse in the form of neurotransmitters which is molecules that are released from the synapse into the synaptic cleft and then attach to ion channels on the dendrite of another cell [5]. This make ions flow in or out of the cell which change the electrical potential between the outside and inside of the cell [5]. If the potential change is above a certain threshold an action potential start to move along the axon. The strength of the action potential is always the same and does not depend on the size of the incoming stimulus.
In the brain there are different parts that are specialized on handling different things. The cerebellum is for example mainly focused on coordinating fine movements and maintaining balance of the body, while the brain stem among other things connect brain with the spinal cord and control vital body functions such as the heart rate [38]. The largest part of the brain is the cerebrum which processes most sensory information, control movements, make decisions and other types of highly complex processes [38]. In the cerebrum there are white matter and gray matter. The white matter is mainly located in the middle of the brain and consists of axons and myelin, which is a form of isolation that is present on the axons [38]. The gray matter which is neuron cell bodies are located in the outer part of the brain that is called the cerebral cortex [38].
2.2. The brain
In the cerebral cortex there are areas, which are specialized in processing different types of information. In the back of the brain there are areas that process sensory information, such as visual, auditory, olfactory and somatosensory [38]. Moving forward in the brain the areas that process motor information are located. The primary motor area generate contractions of muscles while the premotor area contains memories of movements that require muscles to move in a certain way [38]. The frontal parts of the brain deal with more abstract processes such as decision making, intellect and making plans [38]. In the areas that process informa-tion about sensainforma-tions from and movements of the body have a specific area that is dedicated to specific body parts [38]. The size of the area is related to the complexity of the movement or the sensations that come form the area, for example the fingers have a larger area in the brain devoted to them than the upper arms [38].
Worth noting is that the brain is divided into a left and a right hemisphere and the left brain hemisphere control the right part of the body and vice versa [38].
2.2.1
Measure brain activity
There are different ways of measuring the brain activity. One way is to use a magnetic resonance imaging (MRI) equipment and a technique called functional magnetic rensonace imaging (fMRI) to look at changes in oxygen consumption in the brain due to some task or activity performed by the test subject [17]. This is a large and expensive device that is not possible to move around, nor can the person that is measured on move around in the equipment [17]. However it is very good if one is interested in locating the exact brain area that become activated when preforming a task [17]. Another large and expensive device that can measure brain activity is a magnetoencephalograph (MEG), this device measure the magnetic fields that are created by brain activity [17]. With MEG it is possible to monitor change in brain activity with both high spatial and temporal resolution, however it is sensi-tive to movements of the test subject [17].
Neither fMRI nor MEG are suitable for this project and will therefore not be described any further. A technique that is better suited is electroencephalography (EEG), which mea-sure electrical activity of the brain.
Electroencephalography (EEG) is a method to measure the electrical activity of the brain [14]. The brain activity is measured by placing electrodes on the scalp of a person and monitor the electrical activity. The number of electrodes can be varying from 1 to several hundreds depending on how exact measurements that are needed [14]. The sampling rate of the EEG is very high compared to other techniques used to measure brain activity. EEG often has a sampling rate of 100 Hz or more [14] while techniques like functional magnetic resonance imaging (fMRI) have a sample rate of around 1 Hz [14]. This means that the tem-poral resolution of EEG is superior compared to fMRI but the spatial resolution is worse [14]. The reason for the low spatial resolution is that the measurements are made from outside the scalp, which means that activity deep in the brain cannot be distinguished but instead the activity from a larger area in the brain are summed up and the resulting activity correspond to the whole area [14].
Electrodes are placed on the head according to the 10-20 system, meaning that the dis-tance between the electrodes are either 10 % or 20 % of the total head circumference. To keep track of where each electrode is on the head the electrodes are given a letter and a number (or a z) which correspond to its position. The numbers are either odd, meaning that the electrode is on the left side of the brain or even which means it is on the right side of the brain. The different letters are F, T, C, P and O, corresponding to the different brain areas of
2.2. The brain
the brain, and the letters stands for; frontal, temporal, central, parietal and occipital. If there is a z instead of a number it means that it is on the middle line between left and right side of the brain. [14] An example of electrode placements using the 10-20 system is shown in figure 2.2.
Figure 2.2: An example of the 10-20 system for EEG electrode placement2.
Historically the electrodes that have been used is so called wet electrodes, then a gel is placed between the electrode and the skin so that the conductivity increases [14]. Nowadays there are semi-dry and dry electrodes as well, which do not require any conductive gel. Semi-dry electrodes have a saline solution that is released from the electrode when it is placed on the skin [20]. Dry electrodes, as the name suggest, uses no liquids or gel, instead it is coated with a conductive ink [20]. Wet electrodes are still superior to both semi-dry and dry electrodes but the semi-dry electrode is much better than the dry electrode [20].
As mentioned the EEG measure the electrical activity of the brain and this electrical ac-tivity look like waves and are divided into different categories depending on the frequency of the waves. The categories are delta (δ, 0.5-3.5 Hz), theta (θ, 3.5-7.5 Hz), alpha (α, 7.5-12.5 Hz), beta (β, 12.5-30 Hz) and gamma (γ, 30-60 Hz) [14]. Depending on where in the brain and the sleep state of the person the different brain waves are either normal or indicate a pathological state of the brain [14]. An example of this is that delta and theta waves are normal when an adult is a sleep but can indicate a brain tumour or epilepsy when the person is awake [14].
One major problem with EEG is that it is quite sensitive to artifacts, since the amplitude of the EEG is very low [14]. Artifacts can come from the body, for example eye movements, blinking, muscle tension, heart activity and body movements [16]. Other things that might cause artifacts in the EEG recordings are the attachment of the electrodes, power line noise and more. Some of these artifacts can be reduced by placing the electrodes correctly and ensuring that the conductivity is as high as possible, while other require signal filtering [16]. Some artifacts can be removed by low and/or high pass filters, while others require that the source of the artifact is monitored separately, e.g. measure the heart activity with an ECG and remove it [16].
2.3. The eyes
2.3
The eyes
The eyes are located in the head and is the part of the visual system that take in light from the surroundings, transform it from light to electrical signals and send the electrical signal to the brain [12]. The eyes are made up of different parts, figure 2.3 describe the principal look of the eye.
Figure 2.3: The human eye3.
The outermost part of the eye is the cornea, which protects the eye from the outside environment to minimize infections and physical wear and tear of the eye [12]. To let light into the eye the cornea is transparent [12]. Sclera is the tissue that gives the eye a global shape [12]. In the eye there is a pupil which lets light into the eye [12]. The pupil is located in the iris and depending on how much light there is outside the pupil can become bigger, to let more light in, or smaller, to let less light in [12].
Behind the pupil is the lens of the eye, which has the function to focus light onto the retina in the back of the eye [12]. The retina consists of nerve cells that translate the light from the environment to nerve impulses that can be sent to the brain for processing [12]. There are two types of specialized parts of the visual nerves, these are photoreceptors which transform the light. The first type that is responsible for brightness, motion and contrast vision is the rods [12]. The second, called cones, are more specialized in disguising colours and resolution [12]. The density of the cones and rods varies over the retina and although they are present over the whole retina there is a special area, called foveal area, that has the highest density and thus also have the possibility to give the highest resolution of the light [13]. When the rods and cones have processed the light and pass it on to the brain it is transported through the optical nerve, which leave the eye in the optical disk [12].
The eyes are moved around with six muscles that surround each eye [12]. These mus-cles work together and pull the eyes in the direction in which should be moved [13]. The muscles of the left and right eye always move together and in synchrony so that the eyes move equally much either in the same direction, conjugate movements, or in opposite direc-tion, dysjunctive movements, to each other [13]. There are different ways in which the eyes can move and these depend on what the person is doing with the eyes [13].
2.3. The eyes
Saccades are one type of eye movement that are associated with for example reading [13]. Saccades of the eyes mean that they move in the same direction but change position very rapidly [13]. The velocity of the movements can be between 800˝/s to 1000˝/s [13].
The eyes can also focus on a specific point, called fixation [13]. Fixations mean that the light from area of interest is focused in the foveal areas of the eyes so that the object or area is processed with the highest possible resolution [13]. However the eyes are not completely still as the name may suggest instead they move in microsaccades, which are similar to saccades but much smaller and more associated with fixation of gaze [13].
Another type of eye movement is called smooth pursuit, then the eyes follow a moving target continuously [13]. This movement is slower than saccades and the velocity is between 30˝/s to 50˝/s [13].
2.3.1
Measure eye movements
Measuring eye movements can be done in various ways. It is often measured to determine different patterns of eye movements to decide what a person has looked at and also how they have scanned the environment [11]. The different equipment that can be used to track the eye movements in slightly different ways but all have the common goal to measure the visual attention of a person [11]. Although visual attention is dependent on both the eyes behavior as well as the brain, it is assumed that visual attentions is strongly related to the point in the environment that is focused in the foveal area [11].
One type of eye tracker uses video monitoring of the eye [11]. This eye tracker monitor the relative distance between the pupil center and a reflection in the cornea that appears when an applied light, often infra-red light, is shone at the eye [11]. These devices can determine the gaze direction of the person but have generally a rather small sampling rate [11].
Another eye tracker which also was one of the first ones developed is the scleral contact lens, also called search coil [11]. This device is placed directly on the cornea of the eye and is able to make more precise measurements than any other eye tracker [11]. The problem is however that it is invasive and can possibly harm the eye if not placed properly [11]. Neither of these devices will be used in this project, instead an eye tracking device called electroocculography (EOG) will be used.
Electroocculogram (EOG) measure the electrical potential difference between the cornea and retina of the eye [11]. This potential difference is due to that the retina is more negatively charged compared to the cornea, i.e. the eye can be seen as a dipole [22]. Movements of the eye thus change the position of the dipole and this change can be measured using surface electrodes [22]. The change in potential is approximately proportional to the sine of the angle of the movement and the rotations can be calculated using this property [22].
Electrodes are placed around the eye, and the electrodes can then pick up the potential changes when the eye move [11]. As the measurements are done in relation to the head, EOG has to be used in combination with a head tracker to be able to determine the point of gaze [11]. The electrodes are placed above and below the eye to detect vertical movements, there are also electrodes to the left and right of the eye to detect horizontal movements [11]. There are some artifacts affecting the EOG signal, e.g. the potential difference in the eye is slightly varying which means that the change in potential due to movement might vary
2.4. The heart
and thus the EOG equipment need to be calibrated often [22]. There are also muscle artifacts that affect the signal [22].
2.4
The heart
The heart is responsible for pumping blood around in the body and ensuring that the blood flow past the lungs to fill up with oxygen that the body uses to function. To do this the heart is divided into two sides, the right side receive blood from the body and pump it to the lungs and the left side receive oxygenated blood from the lungs and pump it out to the body [38]. Both the right and left side of the heart consists of what is called an atrium and a ventricle. The blood first enter the atrium and then the atrium contracts which pushes the blood into the ventricle [38]. Then the ventricle contracts and the blood is pushed out of the heart into circulation in the blood vessels [38].
The contraction of the heart, also called a heart beat, is controlled with electrical impulses in the walls of the heart. As the blood needs to move through the heart when the heart contracts it is important that atrium and ventricles do not contract at the same time. There is, because of this, a very specific activation path that control the heart contraction. A heart contraction start with an electrical impulse in a point of the heart called the sinoatrial node (SA node) which spreads through the walls of the atrium and make them contract [35]. The impulse then move to a point called the atrioventricular node (AV node) where it is paused for a short while to maximize the blood flow into the ventricles [35]. Thereafter the impulse continue to a point called bundle of His and spread out in the walls of the ventricles in what is called Purkinje fibers which causes the ventricles to contract and the blood to be pumped out of the heart [35].
The contractions of the heart are in the normal case controlled by SA node and the auto-nomic nervous system, which is the part of the nervous system responsible for all automatic involuntary functions in the body [38]. The autonomic nervous system can either increase the rate of the heart contractions, i.e increase the heart rate, if the sympatic part of the system is increased or decrease the heart rate when the parasympathic part of the system increase [35]. The SA node has a natural pace to trigger a heart beat at around 100 beat per minute [38].
Heart rate variability (HRV)
Heart rate variability is a phenomenon that is related to the breathing and autonomic nervous system. The phenomenon is that the time between two consecutive heart beats varies [35]. The time depends on the breathing in the way that the time between two beats decrease when the lungs fill with air and increase when the air is exhaled [35]. Generally the HRV is larger when the breathing is slow, and the difference can be up to 25-30 beats per minute [35]. The HRV is also related to the sympatic and parasympatic activity in the nervous system [35]. This is reflected in the frequency components between 0-0.5 Hz in the HRV, sympatic activity results in more low-frequency components (0.04-0.15 Hz) while parasympatic activity result in high frequency components (0.15-0.40 Hz) [35].
This also mean that the HRV is related to stress in a person. The reason for this is that the sympatic nervous system is activated during stress while the parasympatic system is activated during periods of calm [38].
2.4. The heart
2.4.1
Measure heart rate and heart rate variability
Measuring heart rate can be done in many different ways. The most simple measurement is to place two fingers on for example the wrist and count the pulsations during a certain time and directly get how many beats per minute (BPM) that the heart beats. Heart rate can also be measured optically using light diodes [30]. As the blood is pumped by the heart there is a slight variation of blood volume in the blood vessels in the body that are related to the heart rate and this can be measured with the light [30]. When the blood volume is larger more light is absorbed by the blood compared to when the blood volume is low, this creates a variation in how much light that is able to pass through for example a finger without being absorbed [30]. By measuring the light that has passed through with a photodetector the heart rate can be calculated [30]. This technique is called photoplethysmography (PPG) [30].
It is also possible to measure the electrical activity of the heart using a technique called electrocardiogram (ECG). ECG is also the gold standard to measure HRV and it is also pos-sible to extract the respiration rate from the ECG signal [35]. Studies have shown that it is possible to measure these parameter with PPG as well, but according to a review article by Schäfer and Vagedes [30], ECG give more reliable results during movement and exercise and will therefore be used in this project. For this reason there will be no further description of PPG.
Electrocardiogram (ECG)
Electrocardiogram (ECG) is a way of measuring the electrical activity in the heart. It is done by placing electrodes on the body that pick up the electrical signal described in section 2.4. An example of what an ECG signal can look like is shown in figure 2.4. Here the P-wave correspond to the contraction of the atrias, the QRS-complex is related to the contraction on the ventricles and the T-wave is related to the relaxation of the ventricles [35]. Relaxation of the atrium exist as well, however it occurs at the same time as ventricular contraction and since it is a much weaker signal only the ventricular contraction is seen in the ECG [35].
Figure 2.4: Principle look of a heart beat in a ECG signal4.
ECG measure the voltage difference between two or more electrodes. There are two com-mon configurations, unipolar lead and bipolar lead [35]. Bipolar lead measure the difference
2.5. The muscles
between electrodes placed on different sights of the body [35]. The unipolar lead uses a reference electrode positioned at a place on the body that does not experience any change in voltage due to the heart activity. The voltage difference is then calculated between one electrode and the reference [35]. The electrodes can be placed on different sites on the body and there can be a varying number of electrodes depending on the practical aspects of the measurement [35].
ECG experience artifacts and noise just as any other signal. This means that the result-ing signal do not have the smooth look as in figure 2.4 and there are a number of common artifacts that affect the ECG signal. One of them are called baseline wandering and is a slow change in the baseline amplitude that occurs because of for example movements of the body, breathing or poor contact between electrode and skin [35]. Other artifacts are power line noise, which is caused by power line interference, and electrode motion artifacts, which is caused by impedance changes due to skin stretching [35]. These artifacts can, to some extent, be removed by filtering the ECG signal and thus remove the frequency components that comes from the noise [35]. There are also artifacts that are harder to remove, due to the fact that the frequencies of the noise is approximately the same as the ECG signal frequencies, e.g. muscle activity [35].
Another parameter that affect the ECG signal is the respiration. This is of course a problem if one is only interested in the activity of the heart since the signal components that are related to the respiration then have to be removed [35]. However if one is also interested in measuring the breathing rate, it is possible to get it directly from the ECG signal and thus do not require additional measuring equipment [35].
2.5
The muscles
There are different types of muscles cell that are used for different functions in the body, these are skeletal, heart and smooth muscles cells. The heart muscle cell is of course located in the heart and are responsible for the heart contractions [38]. Smooth muscle cells are located in for example the blood vessel walls and are responsible of contracting and dilating the blood vessels [38]. The skeletal muscle cells are the cells that make up the skeletal muscles, i.e. the muscles that are responsible for all movements of the body [38].
The skeletal muscle cells are connected to each other in fibres and the fibres are connected to each other to make up the muscle [38]. The muscle is then attached to the bones in the skeleton by tendons [38]. To make a movement a muscle need to contract which makes it shorter. To start a muscle contraction the brain decide that the muscle shall move. It does so by sending a nerve impulse from the brain, down the spinal cord and out to the muscle that shall contract [35]. The nerve and the muscle are connected in a neuromuscular junction, where the information from the nerve is transferred to a muscle fibre [35]. One nerve can be attached to one or more muscle fibre depending on how fine movements the muscle perform [35]. The muscles in the thighs have lot of muscles fibers connected to each nerve while the muscles around the eyes have very few [35]. The connection between one nerve and a number of muscle fibers is called a motor unit [38].
As the muscle do not contract equally much all the time there is a way of tuning the amount of contraction needed in the muscle. This is done by only activating a few motor units when the muscle only has to be a little tense while more tension mean that more motor units are activated [35]. There is an additional way to increase the tension when most of the motor units are activated and that is to increase the frequency of the action potential meaning that the fibers are activated more often [35].
2.6. The lungs
2.5.1
Measure muscle activity
Electrical activity of the muscles can be measured by electrodes. The technique is called electromyography (EMG). The activity can be measured either invasively by placing needle electrodes into the muscle to look at the activity in specific motor units or with electrodes on the skin that measure more general activity [35]. Invasive electrodes are more common in clinical applications because they can provide more detailed information [35]. However the surface electrodes can be used to see how much and when a muscle is activated [35], while not causing any harm to the person that is measured on and is thus more suitable to be used in this project.
The interesting aspects of the surface EMG signal is the amplitude of the signal, which give information on how much a muscle is activated and also if it experiences fatigue [35]. Fatigue can also be detected in the power spectrum of the EMG signal, i.e. the frequency components in the signal [35].
Artifacts exists in EMG signals in the same way as in all signals. Power line interference affect the signal and has to be filtered out [35]. Another problem is if the electrodes move on the skin or that the skin stretches [35]. There is also a risk that ECG signals can interfere with the EMG signal if the electrodes are placed at a muscle near the heart [35].
2.6
The lungs
The lungs are involved in the breathing and are filled up with air when a person inhale, this makes it possible for oxygen in the air to get into the blood and carbondioxide to leave the blood so that it can be exhaled [38]. The breathing is controlled by the autonomic nervous system but can also be controlled by voluntarily [38]. Depending on the oxygen need in the body the breathing is slower or faster, during rest a normal person take about 12 breaths per minute, while exercise increase the need and the breathing rate can be up to 40-45 breaths per minute [18].
Normal relaxed breathing mainly involves the diaphragm and causes the belly to move [38]. Heavy breathing on the other hand involve muscles round the ribs and belly to help push the air out and fill the lungs faster [38]. During normal breathing there is a small pause between the exhale of one breath and the inhale of a new breath, this is called the natural respiratory pause and when shooting a shot it should be fired in this pause so that the breathing do not cause the weapon to move [34].
2.6.1
Measure breathing rate
Breathing rate can be measured in many different ways, for example using a band that is fitted around the torso and monitor the stretch of that band. Another way is to measure with a device called spirometer, with a spirometer it is possible to measure not only the respiration rate but also the volume of the breaths and flow rate of the air [18]. It is also possible to monitor it through a device that monitor heart activity, as the heart rate and position of the heart is affected by the breathing cycle which can be seen as a change in the measured data [35]. This means that PPG or ECG can be used to monitor the breathing rate [35, 30].
In this project respiration will be measured with a ECG sensor that measure the impedance change across the chest by placing three electrodes on the body [31]. It works by sending a very small high frequency current into the body, that become amplitude modulated by the breathing [31]. The technique is called impedance pneumography (IP) [18].
3
Material and method
This chapter intend to describe what has been done during the project. The project started with a literature study, followed by an experiment. The literature study is done individually while the experiment is done together with Amanda Hansson.
3.1
Literature study
In the beginning of the project a literature study was preformed. The goal of the literature study was to find suitable physiological parameters that were related to shooting perfor-mance and also techniques that could be used to measure these parameters. Another goal with the literature study was to find studies done by others that had investigated relations between physiological parameters and target shooting or aiming sports.
Firstly a number of body functions were identified to be related to target shooting. This was done by talking to experts and reading material of basic target shooting techniques. As a result of this study the areas brain activity, eye movements, muscle activity, heart ac-tivity and respiration were chosen for further analysis. This literature study also included identifying suitable measuring techniques that could be used in the simulator environment. When this was identified, the next stage of the literature study started, i.e. finding studies that had looked at target shooting and one or more of the physiological parameters. To limit the search, studies that had used other measuring techniques than those found suitable or measured other parameters were excluded. Although the intended use of the results in this thesis is mainly military training it was found that studies only including military target shooting was too limiting. Therefore studies including other types of target shooting and aiming sports were included. To limit the search further it only included studies that focused on:
• Learning target shooting/aiming sports
• Attention and focus while performing target shooting/aiming sports
• Comparison between expert and non-expert/novices performing target shoot-ing/aiming sports
3.2. Experiment setup
• Feedback about physiological parameters while performing target shooting/aiming sports
In the beginning of this stage in the literature study the main focus was on brain activity, and as time progressed the focus shifted to the other areas that were identified. Searching for suitable studies were done mainly in Google Scholar and Linköping University library. Examples of key words that were used when searching can be found in appendix A. These searches resulted in a number of studies that were found to be interesting for the aim of this thesis. These studies are presented as literature studies in chapter 4.
As a result of the literature studies it was decided that a small experiment that measured some of the parameters found in the literature study would be performed. The reason for this was to see if any of the results found in the literature study could be found in the sim-ulation environment at Saab. All parameters except the brain activity were found suitable to measure in the simulator environment. Brain activity is very complex to measure and the quality of the experiment must be high and well thought through in order to get a usable result. For this reason brain activity measurements got excluded from the experiment and the only results presented regarding brain activity will be those in the literature study. The literature study showed that many previous studies had compared expert and novices in different tasks and it was therefore decided that this would be done in the experiment performed in this thesis as well. Suitable sensors were found for measuring respiration, heart rate, muscle contractions and eye movements. These were purchased and based on the literature it was decided that an experiment that investigated the relationship between the measured values from the sensors and that performance results of experts and novices were to be conducted. The reason for this was to try to find a pattern in the measured data that corresponded to a good performance.
3.2
Experiment setup
An experiment was conducted using 4 military expert marksmen (4 male), 1 civilian compet-ing marksmen (1 male), 4 non-active marksmen1(4 male) and 4 novice marksmen2(2 male, 2 female). All participant except one was right handed. The experiment was conducted in the GC-IDT simulator3. Each participant performed the experiment 1 time.
The participants arrived at Saab Training and Simulation in Huskvarna. Upon arrival they were given information about the setup and received an information sheet describing the experiment and gave their consent to be in the experiment by agreeing to the information on the sheet. This information sheet can be found i appendix B. After that they where given a introduction to the GC-IDT and how it works and got to try if for approximately 5 minutes. They also received information on what type of tasks that they should perform.
After the introduction, the participants were given a code, which was used to name the data, and then performed the experiment individually. The codes depended on the group the participant belonged to, military expert marksmen received a code starting 01, novice marksmen were given a code starting with 02, civilian competing marksmen received a code starting with 03 and the non-active marksmen were given a code starting with 04. The exper-iment started with placing the electrodes on the body. The ECG and respiration electrodes here placed in accordance with figure 3.1. EMG electrodes were placed in accordance with 1Non-active meaning that they had received basic shooting training or actively trained more than 5 years ago
and not practiced regularly since.
2E.g. people that have not had any prior shooting training 3See section 2.1.2.
3.2. Experiment setup
figure 3.2 and EOG electrodes were placed according to figure 3.3. Skin preparation was performed on all sites where electrodes should be placed, this included cleaning it with alcohol and, if it was necessary, shaving the area. The electrodes were then connected to the measuring equipment and when everything was connected it was tested to check that the signal quality was satisfactory.
3.2. Experiment setup
Figure 3.1: Placement of ECG and respiration measuring electrodes4.
Figure 3.2: Placement of EMG measuring electrodes5.
Figure 3.3: Placement of EOG measuring electrodes6.
4Image created by RexxS and released to Public domain, positions of electrodes are added to the original image. 5Image created by RexxS and released to Public domain, positions of electrodes are added to the original image. 6Image created by Freepik, positions of electrodes are added to the original image.
3.2. Experiment setup
ECG and respiration were measured with a Shimmer 3 (Shimmer, Dublin, Ireland) which was programmed to measure respiration, with a gain of 3, sampling rate 256 Hz and bits 24. EMG was also measured with a Shimmer 3 which was programmed to measure muscle activity, with a gain of 12, sampling rate 504.12 Hz for participant 0101, 496.48 Hz for partici-pants 0102-0207,0410-0413 and 489.07 Hz for the other participartici-pants. The EMG sensor uses a technique called common mode rejection [32]. The reason for this is to remove noise from the surroundings that interfere with the signal [32]. It works by having several electrodes on the body and subtract the signal from one electrode with another, which removes the common noise component while leaving the EMG signal [32].
EOG was measured with a BlueGain EOG Biosignal Amplifier (Cambridge Research Systems Ltd., Rochester, UK) with sampling rate 500 Hz. The Shimmer 3 devices were programmed and the data from the Shimmer 3 devices was collected with a program called Consensys Pro Software (Shimmer, Dublin, Ireland). The data from the BlueGain sensor was collected using a program called EOGtest (Cambridge Research Systems Ltd., Rochester, UK). Electrodes used for ECG, EMG and respiration measurements were Covidien Kendall Disposable Sur-face EMG/ECG/EKG Electrodes of Ag/AgCI type, with gel and a diameter of 35 mm. For the EOG measurements Ambu Neuroline 700 of Ag/AgCl type were used.
Thereafter the experiment started. The experiment consisted of six tasks, which are called X, A, B, C, D, E from now on. A more detailed description of the task can be found in table 3.1. All tasks were performed in the standing position. Between each task the the participant got to rest for approximately 3 minutes.
Table 3.1: Detailed description of the tasks
Task Position Distance
Number of shots
Description
X - -
-Control task, where maximal and minimal muscle activity, maximal and normal respiration were mea-sured
A Standing 100 m 5 No time requirement, no stress triggers
B Standing 100 m 5
Targets were pop-up, meaning they were shown for 2 seconds then disappeared for 1 second. The par-ticipants had to shoot 5 shots on 4 showings of the targets
C Standing 100 m 1 The participants should shoot 1 shot as fast and ac-curate as possible
D Standing 100 m 1
The participants were told that the shot should be scored and compared to other participants and that the winner would receive a price
E Standing 100 m 5
The participants had to cycle as fast as they could for 1 minute on an exercise bike, then had 1 minute to shoot 5 shots
During these tasks data were collected from the sensors. For participant 0410-0413 the only parameters measured were ECG, respiration and EMG as the EOG equipment did not work during those measurements. The data was transferred from the sensors via Bluetooth to a computer nearby. The computer was also used to manually mark the shots in the data so that is was possible to determine where in the data the shot had taken place. This was used in the data analysis stage, which is described more in section 3.3. Scores from each shot was also collected manually.
3.3. Data analysis
3.3
Data analysis
All collected data was divided into sections of maximum 5 seconds before a shot and 1 second after a shot in a task. Some shots were taken with a shorter time interval and in those cases the sections started directly after the previous shot and ended 1 second after the shot. Thereafter the signals were filtered. A more detailed description of the signals processing applied to the signals will be described below.
3.3.1
Signal processing EOG
The only signal processing done on the EOG signals was to remove the mean of the signal. Thereafter the gradient of the signal was determined using the Matlab-function gradient7.
3.3.2
Signal processing Respiration
First of all the mean of the signal was removed to get a new mean of approximately zero. The respiration signal was then filtered using a lowpass FIR8filter with filter order 40, cut-off frequency 2 Hz and using a Kaiser window with shape parameter 3. Thereafter the gradient of the signal was determined using the Matlab-function gradient9.
3.3.3
Signal processing EMG
The processing of the EMG signals start by removing the mean of the signal and thereafter filtering it. The filters used of the EMG signal is a bandpass FIR filter with filter order 20, cut-off frequencies 20 Hz and 240 Hz and using a Kaiser window with shape parameter 3. The signal is also filtered with an IIR10notch filter to remove 50 Hz power line noise.
After filtering the signal is rectified by taking the absolute value of all the values in the signal. The rectified signal is then lowpass filtered using a FIR filter with filter order 40, cut-off frequency 5 Hz and using a Kaiser window with shape parameter 3. Thereafter the area under the curve is calculated of 1 second periods and the value is divided by the value for the same parameter for task X, were the participants preformed a maximum activity of the muscle. The resulting value is thus a percentage of the maximum tension of the muscle.
3.3.4
ECG Signal processing
Signal processing of the ECG signals started with choosing the ECG lead with best signal quality. This were done through visual inspection of the original signals. The one that was found to have the best quality were used in further processing. The mean of the signal was removed to get a new mean of approximately zero.
The signals were then filtered using a bandpass FIR filter with cut-off frequencies 1 Hz and 20 Hz, filter order 60 and using a Kaiser window with shape parameter 3. The filter was used to filter the ECG using zero-phase filtering. For all signals from task E and the signal from participant 0205 from task B the lower cut-off frequency was changed to 3 Hz as these signals contained more baseline wandering than the other signals.
7https://se.mathworks.com/help/matlab/ref/gradient.html [2017-06-22] 8FIR = finite impulse response
9https://se.mathworks.com/help/matlab/ref/gradient.html [2017-06-22] 10IIR = infinite impulse response
3.3. Data analysis
After filtering, peaks in the signal corresponding to the R-peaks in the ECG were detected and the pulse in beats per minute were calculated using equation 3.1.
Beats per minute=60Total number of samplesNumber of peaks
Sampling frequency
(3.1)
Another parameter that were calculated were the time between individual heart beats and also the heart rate for these individual beats. The last parameter that was calculated was a percentage of where in the interval between two R-peaks the shot was taken.
3.3.5
Statistical analysis
The statistical analysis was done on the processed data. The tests performed were principle component analysis (PCA) and correlation, a detailed description on how the statistical tests has been done can be found in Amanda Hansson’s thesis "Accelerating learning during rifle shooting focusing on the shooter".
4
Results
4.1
Literature study
The following sections describe the results from the literature study regarding the previous studies that have been done about target shooting or aiming sports and physiological param-eters.
4.1.1
Literature study of brain activity
EEG can be used to measure a lot of different kinds of brain activity, it can indicate patho-logical things in the brain as well as sleep [14]. However, this report intend to present mea-surements that might be useful to accelerate learning and will therefore not cover all possible things that can be measured with EEG. The identified measurement of interest are learning and attentional focus. These will be described below.
Learning and EEG
When talking about learning a skill and monitoring it with EEG the term "neural efficiency" is often talked about [26]. The term means that when a person is good at something the brain activity is lower than if a beginner do the same thing. With training the activate areas in the brain becomes smaller and thereof comes the term. The reason for this is assumed to be that when a person learns a skill the connections in the brain changes so that the neural network that preforms the task become more efficient.
Shooting is not an exception to the "neural efficiency" hypothesis and many studies have shown that there are differences in brain activity between an expert marksman and a novice [26]. Many studies have shown that an increase in alpha power in the left-temporal area of experts that is greater than for the novices during the seconds leading up to the shoot [26]. This means that the area is more relaxed and the reason for this is described to be because of suppression of verbral-analtyic processing [26].
4.1. Literature study
Attention in EEG
Many studies have showed that the focus of attention is important when performing a task. They way that they have shown this is by giving direction on where to focus or letting the participant in the study self-estimate where their attention lays. While this give a general indication of the focus it might not be the whole truth about the attention and a subjective measure of the focus level would be good. Some researchers have measured this by looking at the brain activity using an EEG measurement of people when they perform a task. De-pending on the focus of attention the brain activity seen in the EEG look different and can be related to the focus level.
It is mainly the seconds leading up to the task that has been of interest for the researchers [2, 26, 1]. For example Bertollo et al. [2] investigated the brain activity in the 3 seconds leading up to a shoot and saw that the brain activity differed between the different types in the MAP model. The main difference were in the alpha and theta frequency bands. They observed that the power in the frequency bands were generally higher in type 1 and 4, while there were a decrease in power for type 2 and 3 [2].
Loze et al. [21] compared the brain activity in the occipital region and found that air-pistol experts that have a decrease in alpha power in this region in the time leading up to a shoot perform better than if there is an increase in the alpha power. They suggest that the reason for this is that experts focus more of their attention on performing the movement than actually focus on the aiming.
Using EEG to accelerate learning
To accelerate learning and/or enhance performance using EEG a technique called neurofeed-back is the most used. The technique works by measuring a persons brain activity with EEG and give feedback about the brain activity to the person either as visual, auditory or haptic stimuli [26]. Using the feedback the person tries to master their brain activity and change one or more aspects of their brain activity [26]. It is often the power of a certain frequency band that the person try to change, but it can also be a combination of two frequency bands [26]. Worth mentioning is that the result from using neurofeedback has varied, some research groups have shown promising results while other have not seen any difference between control and neurofeedback groups [26].
The research groups that have shown improvements when using neurofeedback have how-ever done it with the performance of target shooting [26]. Park et al. [26] give two examples of successful results, in one the performance of archers increased with neurofeedback and in the other the performance of riflemen increased. Another study that has shown promising results when using neurofeedback for riflemen is made by Berka et al. [1]. They identified a pre-shot peak performance profile (PSPP) in experts that described the optimal brain activity and heart rate in the 3 seconds preceding the shoot, this profile can be seen in figure 4.1. The PSPP was then used to train novice riflemen by measuring their EEG and ECG and giving feedback with two vibrators attached to the neck of the person. The vibrators vibrated in synchrony with the heart beat that was detected with the ECG, and to give feedback about the alpha power band increase one of the vibrators stopped if the increase was between 5-10 % and the other stopped when the increase was more than 10 % above baseline [1]. This feedback was given both when the novices were sat at a desktop as well as during shooting, they were also given visual feedback about their breath control and trigger pull. Using this, Berka et al. [1] managed to accelerate learning by 2.3 times compared to their control group.
4.1. Literature study
Figure 4.1: a) Show the PSPP for the alpha and theta brain activity in 3 seconds leading up to a shot for the expert marksmen that participated in the study by Berka et al. [1]. b) Show the PSPP for the heart rate activity in 3 seconds leading up to a shot for the expert marksmen that participated in the study by Berka et al. [1].1
Summary of studies using EEG to look at learning and attention
Table 4.1 will give an overview of studies that have looked at brain activity using EEG in task that are similar to shooting. It will also present the results of the studies and the techniques used to process the EEG signals.
4.1. Literature study
Table 4.1: Summary of studies using EEG to measure brain activity during target shooting, aiming sports and tasks similar to target shoot-ing.
Research
group Focus of study
Electrode sites2
Signal
pro-cessing Result Comment
Loze et al. [21]
Brain activity related to the best and worst shots of ex-pert air-pistol marksmen in the seconds 3 seconds leading up to the shot Oz, T3, T4 Bandpass (0.5-70 Hz), visual in-spection for artifacts, FFT with raised cosine window Alpha power increase during pre-shot of best shot at Oz, alpha power decrease before worst shot at Oz, no signifi-cant results for T3 and T4
They suggest that this indi-cate that the marksman does not have maximal visual attention on the aiming Cheng et al. [7] Differences be-tween experts and novices dart throwers in the 2 seconds leading up to the throw Fz, F3, F4, C3, C4, T3, T4, Pz, P3, P4, O1, O2, EOG, FPz (ground) Band pass (1-30 Hz, FIR fil-ter), eye-movement artifact removal, visual in-spection for artifacts, FFT with 50 % overlap, 256 sample Hanning window SMR mean power and beta1 in C3 and C4 significantly higher for ex-perts, no other significant re-sults where found SMR = sensori-motor rhythm (12-15 hz), beta1 (15-18 Hz), Ex-pert has lower activation in sensorimotor area compared to novices, novices prob-ably process more informa-tion than the experts Haufler et al. [15] Differences be-tween experts and novices marksmen in the 6 seconds leading up to the shot F3, F4, C3, C4, T3, T4, P3, P4, O1, O2, EOG Bandpass (1-100 Hz), Notch filter (60 Hz), Eye-movement artifact removal, visual in-spection for arti-facts, FFT, power band averaged Experts had higher theta and alpha power in left hemisphere at temporal, occipital and partial regions (for theta all sights differed), experts had lower beta and gamma power than novices
3 kinds of tasks where one was shoot-ing, looked at many dif-ferent aspects (difference be-tween groups, hemispheric asymmetry, brain regions)
4.1. Literature study
Continuation of Table 4.1 Research
group Focus of study
Electrode sites
Signal
pro-cessing Result Comment
Cooper et al. [10] Relationship between alpha activity and internal and external focus of attention in a task that tested external (i.e. listening to sound) and internal focus (i.e. imagining the sounds) in healthy subjects 28 elec-trodes accord-ing to 10-20 system, FTC1, FTC2, TCP1, TCP2, CP1, CP2, PO1, PO2, EOG Ocular artefact rejection, spline-fit to 8192 points, FFT Alpha power higher in inter-nally directed attention than external, also higher when task demand was higher They suggest that the result mean that in-crease in alpha power happens when non-task relevant areas of the brain are inhibited Berka et al. [1] Accelerate learning using neurofeedback based on expert marksmens performance profile in the 3 seconds leading up to the shot Fz- PO, Cz-PO, Fz-C3, F3-Cz, C3-C4 and Fz, F3, Cz, C3, C4, POz, P3, P4 60-Hz notch filter, re-moval of epochs contain-ing spikes, amplifier saturation, excursions, excessive EMG or eye blinks Increase in alpha power at all electrode sights 2 s before shot for experts, deceleration in heart rate start-ing 2 s before shot in experts, using these results feedback was given to novices which then performed better than a control group Only measured over a single session, gave feedback on more things than just EEG (such as res-piration, heart rate, trigger pull)
