Neurofeedback Games to Improve Cognitive Abilities

2014 International Conference on Cyberworlds Neurofeedback Games to Improve Cognitive Abilities Yisi Liu Fraunhofer IDM@NTU Nanyang Technological University Singapore LIUYS@ntu.edu.sg Olga Sourina Fraunhofer IDM@NTU Nanyang Technological University Singapore eosourina@ntu.edu.sg Xiyuan Hou Fraunhofer IDM@NTU Nanyang Technological University Singapore HOUX0003@ntu.edu.sg Abstract Neurofeeback can be used to enhance cognitive abilities related to multi-tasking such as working memory, attention, etc. We propose and implement a neurofeedback system which includes a number of neurofeedback algorithms and a Shooting game. To make neurofeedback more effective, an Individual Alpha Peak (IAP) frequency and individual alpha bandwidth are calculated and applied in the algorithms. We do preliminary study on the effectiveness of the proposed neurofeedback system with three subjects taking 6 sessions each. The neurofeedback protocols based on the power of individual upper alpha or beta-1/theta ratio are used. Our hypothesis is that after the neurofeedback by playing the Shooting game, the individual alpha peak frequency increases. The results show that all subjects had a higher individual alpha peak frequency after the that indicated an enhancement of the subjects cognitive abilities related to multi-tasking. Keywords-EEG; Neurofeedback ; Neurofeedback game; individual alpha peak frequency; individual alpha bandwidth I. INTRODUCTION Multi-tasking skills are needed in daily life and in a number of situations. There are two levels of multitasking such as context multitasking and pure multitasking. For example, normal car driving needs context multitasking from the driver. If he/she needs to use navigator or just listen another passenger then this is an example of pure multitasking. Another example of multitasking is the situation when a pilot of the plane needs to listen and response to the air traffic controller. Human can have better performance (e.g. a safer driving) if his/her ability of multitasking is improved. Multi-tasking is related to different cognitive abilities such as memory and attention, and neurofeedback can be used to improve such abilities. Neurofeedback is one of biofeedback that uses Electroencephalogram (EEG) to measure the real-time brain activity and teach the user to do self-regulation. During the, visual feedback such as color or velocity change of the objects can be used to indicate whether the current brain state is the desired one or not. Then, the subject can maintain/adjust the current state based on the feedback. Traditionally, this technique is used as the treatment for patients with mental illnesses. For example, the neurofeedback was given to the children with learning disability (LD), and after the, the cognitive performance of these children was enhanced [1]. Recently, it is also used as a method for healthy people to exercise their brain and to boost their cognitive abilities. For example, healthy adults attended neurofeedback and their cognitive performance has been improved significantly after the [2, 3]. Usually during the neurofeedback, the target is increasing or suppressing the activity in certain EEG frequency range. For example, the amplitude in the 12-15 Hz band is targeted to be increase [4]. In [5, 6], it is shown that the individual alpha frequency is a positive indicator of the cognitive abilities, which means a better performance can be obtained if the individual alpha peak frequency and individual alpha bandwidth are increased. Different neurofeedback protocols such as increase of upper alpha band power or the beta/theta ratio are applied. The results show that after such the individual alpha peak (IAP) frequency is larger and the individual alpha bandwidth are wider [6]. In this paper, we propose and implement a neurofeedback system to enhance cognitive abilities related to multitasking. The system includes a neurofeedback algorithms module that allows the users to choose different neurofeedback protocols and a neurofeedback game module which is a Shooting game. We also did an experiment with three subjects to assess the effectiveness of the neurofeedback system. Our is based on the IAP frequency and individual alpha frequency range. Once the individual alpha frequency range is obtained for each subject, the individual upper alpha, beta, and theta frequency range can be defined. Then, the power of individual upper alpha or beta-1/theta ratio is calculated and enhanced during the neurofeedback by playing the game. The paper is organized as follows. In Section II, related work such as introduction to five frequency bands in EEG, review on neurofeedback algorithms is given. In Section III, the proposed neurofeedback system is described. In Section IV, a case study with three subjects who used the neurofeedback system and played the Shooting game is shown. Section V gives the results of the case study. Section VI concludes the paper. 978-1-4799-4677-8/14 $31.00 2014 IEEE DOI 10.1109/CW.2014.30 161

II. RELATED WORK A. EEG Ranges In neurofeedback systems, different algorithms related to EEG frequency bands are used. Five EEG frequency bands are defined based on their frequency ranges as follows: from low to high frequency, namely delta, theta, alpha, beta, and gamma. Delta waves (0.5-4Hz) The delta waves are related to deep sleep [7]. Theta waves (4-8Hz) The theta waves are related to drowsiness. It is also found to be associated with learning and memorial processes and attention recently. Neurofeedback to increase theta power can benefit an artist as his/her creativity and performance technique are enhanced [8]. Training to decrease theta power is used to improve the verbal IQ, executive functions and attention for seniors [9] or in the treatment for children with mental disorder such as Attention Deficit Hyperactivity Disorder (ADHD) [10, 11]. Alpha waves (8-12Hz) Generally, alpha is related to relaxation. Recently, it was also shown that neurofeedback to increase individual upper alpha power can help improve the ability of memorization [12] and visuospatial skills [2, 3] for healthy people. Beta waves (12-30 Hz) Beta band is related to emotional arousal. If a subject is in a scared state, the power of beta band may increase [7]. It is also believed to be relevant to motor functions [13]. Training to increase beta power can help ADHD children who have problems with attention [10, 11] and can also help healthy subjects reduce the reaction time in the Test of Variables of Attention test [14]. Gamma waves (above 30Hz) These rhythms are used to confirm certain brain diseases [7]. Despite the above-mentioned standard frequency ranges, use of individual frequency ranges can improve efficiency of neurofeedback [3, 6]. The individual alpha peak can be obtained from the EEG recording during eyes closed, and the individual alpha bandwidth can be obtained by comparing the EEG recorded during eyes closed and eyes open [15]. Based on the individual alpha bandwidth, theta and beta ranges can be decided correspondingly. The individual frequency range can make the neurofeedback strategy more efficient. For example, in [6], an example is given that the individual alpha peak frequency of a subject was 11.7 Hz which was larger than the norms of alpha peak frequency in the similar age group (that is around 9.61 Hz), but the individual alpha bandwidth for this subject was narrower (1.52 Hz) than the standard definition which ranges from 8 to 12 Hz. When the standard frequency range is used in the neurofeedback, the subject can have a headache and the irritability [6]. This problem can be solved when the individual frequency ranges are used. The effectiveness of neurofeedback with use of individual frequency ranges is much higher. B. Review on Neurofeedback Training Neurofeedback based on individual alpha frequency has been proved to be effective for cognitive performance [2, 3] and working memory performance enhancement [16]. In such neurofeedback, the individual upper alpha (UA) power is targeted to be increased. For example, in [2], the subjects needed to increase the individual upper alpha power and to decrease theta power; in [3], the protocol was enhancing the individual upper alpha power. In both works, the cognitive performance of healthy participants was increased significantly after the. The based on the suppression of theta/alpha power ratio is another effective method for cognitive performance enhancement. In [1], children with learning disabilities (LD) needed to decrease their theta/alpha power ratio to get a reward during the. When it comes to ADHD children, the neurofeedback protocol is suppressing theta band power and enhancing beta band power, in other words, the theta/beta ratio should be decreased [10, 11]. The theta/beta can be used in the for adults as well. It is proved that after such, the cognitive performance of the subjects was improved [6]. Another neurofeedback protocol is based on SensoryMotor Rhythm (SMR, 12 15 Hz) activity. In such neurofeedback, the subjects need to increase the amplitude in SMR band without concurrent increase in theta and high beta band. the, the commission errors in the divided attention task were reduced, and perceptual sensitivity in the continuous performance tasks was improved [4]. In [17], the cognitive performance of healthy subjects was improved after 8 sessions of neurofeedback based on the enhancement of SMR amplitude. When the subjects are too anxious or stressed and they need to relax, for example, the performers can be under the stress before the show, the neurofeedback can be designed to relief the anxiety. The protocol is to raise the power of theta over alpha with auditory feedback and eyes closed. such, it shows that music performance was improved [18]. In [19], it is found that by increasing the frontal theta amplitude in the neurofeedback among healthy elderly people, the executive attention and working memory were improved. Neurofeedback can be done in different forms such as games, simple visual feedback such as color changes, or audio feedback such as a beeping sound. These changes can give feedback on whether the current brain activity is the desired one or inhabited one. For example, the speed of a car depends on the attention level of the player [20]. It means if the player has a high level of attention, he/she can drive faster. In [21], the color and the movement speed of the icon in the pacman-type neurofeedback game indicates whether the brain activity of the subject reaches the reward criteria based on the corresponding neurofeedback protocol. In our work, we proposed and implemented a neurofeedback system that enables different types of neurofeedback protocols. We also develop a 162

neurofeedback game named Shooting to make the neurofeedback experience more engaging. III. NEUROFEEDBACK TRAINING SYSTEM The procedure of neurofeedback is illustrated in Fig. 1. The subject s EEG signals are acquired by EEG device in real time and filtered by artifact removal methods such as band-pass filter. Then, the features are extracted from the EEG data and compared with the threshold to judge whether the brain activity of the subject is desired or not. Next, the result is sent to the neurofeedback games to give the feedback to the user. Our proposed neurofeedback system consists of two modules. One is the neurofeedback algorithms module and the other is the neurofedback game. Both modules are described in the following sections. threshold is automatically decreased or increased based on the subject s performance. If the subject cannot reach the reward criteria for 100 times, the threshold is decreased. In the contrast, if the subject can reach the reward criteria for 100 times, the threshold is increased. In the Fixed Threshold mode, the threshold is predefined by entering the value or by using the scroll bar to choose the threshold value in the Thresholds and Brain State block. The Training Direction can be decided based on the needs of the neurofeedback. For example, if the target of the is to increase upper alpha power, Reinforce should be selected from the menu. Then, when the calculated feature values are larger than the threshold, reward feedback will be given. If the target is to decrease theta power, Inhibit should be selected. Then, when the calculated feature values are smaller than the threshold, the reward feedback is given. Figure 2. Neurofeedback Training Protocol Selection Menu. Figure 1. The procedure of Neurofeedback. A. Neurofeedback Algorithms The proposed neurofeedback system allows different neurofeedback methods be chosen. The screenshot is shown in Fig. 2. In the Methods block we can see that the implemented neurofeedback algorithms include fractal dimension [22], power of standard EEG bands, power of customized band, and beta/theta power ratio with customized beta and theta frequency range. Different from the traditional frequency band based neurofeedback, the method using fractal dimension considers the non-linear property of EEG. It does not need to adjust the frequency range individually and can be used directly with all subjects (in adaptive threshold mode) [22]. During the neurofeedback, the features in the selected methods such as power from different frequency bands are extracted and compared with the threshold (the reward/inhabit criteria). Then, the result is sent to the game module via a TCP port to trigger certain command. The threshold setting is defined in the Mode block: Adaptive Threshold or Fixed Threshold can be selected. In the Adaptive Threshold mode, the B. Neurofeedback Games Different neurofeedback games can be used with the neurofeedback algorithms module. A Shooting game is designed and implemented in our work. The user can choose the difficulty levels of the game. In total three levels are designed and implemented (Fig. 3). In all levels, the user has to shoot the enemies (robots). When the subject s brain activity is recognized as undesired one, the robots are blue and run around (Fig. 4). The robots change the color to red and stay still if the brain activity is the desired one. Then, the user can shoot the robots and they will be destroyed (Fig. 5). In Fig 4 and 5, the screenshots of the first level of the game are given. To make the game more attractive, two more levels are implemented as shown in Fig. 6 and 7. In the second and third levels, the background of the game is changed, and the entire map is broadened. IV. METHODS Preliminary study was done to assess the effectiveness of the proposed neurofeedback system. 163

A. Subjects Three subjects (two male and one female) with age ranging from 24-30 years old participated in the study. None of them had mental illness before. B. Procedure For all subjects, the procedure of each neurofeedback session consists of three parts. In the first part, EEG with one minute eyes closed and one minute eyes open are recorded. These data are used to calculate the individual frequency range for each subject. In the second part, the subjects need to play the neurofeedback game for entire 40 minutes (3 times game playing, 10 minutes each time with 5 minutes of resting in between). For Subject 1, the neurofeedback protocol is to increase the individual upper alpha power, for Subject 2 and 3 the neurofeedback protocol is to increase the individual beta-1/theta power ratio. In the third part, EEG with one minute eyes closed and one minute eyes open are recorded again. These data are used to analyze the effects. All three subjects had 6 neurofeedback sessions. Figure 5. Screenshot of the first level with the robots turn red color corresponding to desired brain state of the player. Figure 6. Screenshot of the second level. Figure 3. Screenshot of the level selection page. Figure 7. Screenshot of the third level. Figure 4. Screenshot of the first level with the robots turn blue color corresponding to undesired brain state of the player. C. EEG Device The 14 electrodes Emotiv device [23] is used to obtain EEG signals. This device is wireless and easy to set up. The locations of the electrodes are standardized by the American Electroencephalographic Society [24] which include AF3, F7, F3, FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, AF4. The technical parameters of the device are as follows: bandwidth 164

- 0.2-45Hz, digital notch filters at 50Hz and 60Hz; A/D converter with 16 bits resolution and sampling rate of 128Hz. Channel P8 is selected in the neurofeedback system to calculate the features according to the research in [6] where parietal lobe channel was used in the neurofeedback. D. EEG Analysis The individual alpha peak frequency can be obtained from the EEG recorded with eyes closed. It is the peak frequency in the eyes-closed EEG (Fig. 8). It is found that the EEG spectrum is suppressed when the eyes are opening compared to the eyes-closed condition. Thus, the individual alpha band width is calculated using both eyes-closed and eyes-open EEG data, and it is defined as the frequency range which has at least 20% suppression of the alpha peak frequency power [6]. An example of the power spectrums of the 1 minute eyes-closed and eyes-open EEG from Subject 2 is plotted in Fig. 8. From the figure, we can see that the alpha peak frequency is 9.33 Hz, and the individual alpha frequency range is from 8.94 Hz to 10.92 Hz, which is not the same as the standard frequency range (8-12 Hz). the individual alpha frequency range is calculated, the corresponding upper alpha frequency range, beta-1 and theta frequency range can be defined and entered to the neurofeedback algorithms system. The upper alpha frequency range is defined by the individual alpha peak frequency and upper alpha frequency, Beta-1 ranges from the upper alpha to 18 Hz, and theta ranges from 3Hz to lower alpha [6]. In Fig. 8, upper alpha frequency is 9.33-10.92Hz, beta-1 frequency is 10.92-18Hz, and theta frequency is 3-8.94 Hz. The hypothesis of our neurofeedback is that after, the individual alpha peak frequency increases and the individual alpha bandwidth becomes wider. V. DATA PROCESSING AND ANALYSIS RESULTS each neurofeedback, one minute eyesclosed and eyes-open EEG data were recorded for all three subjects. The initial individual alpha peak frequency of Subject 1 is calculated as 11.02 Hz; the lower alpha frequency is 8.80 Hz; the upper alpha frequency is 13.16 Hz (individual alpha band width 4.36 Hz). In total, Subject 1 attended 6 neurofeedback sessions based on upper alpha power, which means if his upper alpha power exceeded the pre-defined threshold, he got reward in the shooting game: the robots in the game turned red, and disappear after the subject shoots them; otherwise the robots are blue and run around. and after each session, the eyes-closed and eyes-open EEG data were recorded and analyzed. The alpha peak frequency, upper alpha frequency and lower alpha frequency, and the alpha bandwidth (upper alpha frequency -lower alpha frequency ) are given in Table I. The changes of alpha peak frequency and alpha bandwidth across different neurofeedback sessions are illustrated in Fig. 9 and 10. From these two figures, we can observe some interesting phenomenon: 1) the individual alpha peak of Subject 1 did not always increase immediately after neurofeedback. However, if we compare the alpha peak before the first session and before the last session, it is increased from 11.02 Hz to 12.17 Hz; 2) the individual alpha bandwidth is always increased immediately after the neurofeedback, however, if the changes across different days are compared, there is no increase between the first and the last. Figure 8. Spectrum of 1 minute eyes-closed and eyes-open EEG from Subject 2. Figure 9. The changes of alpha peak frequency through different neurofeedback sessions from Subject 1. 165

1 2 3 4 5 6 Figure 10. The changes of alpha bandwidth through different neurofeedback sessions from Subject 1. TABLE I. Training THE CHANGES ACROSS 6 NEUROFEEDBACK TRAINING SESSIONS OF SUBJECT 1 Alpha Peak Lower alpha Upper alpha Alpha Bandwidth 11.02 8.80 13.16 4.36 11.47 7.63 12.85 5.22 11.78 9.46 12.35 2.89 11.75 7.69 12.69 5.00 11.77 10.40 12.96 2.53 11.39 7.47 12.66 5.19 11.44 8.88 12.61 3.73 10.57 7.74 12.08 4.34 11.36 8.86 12.35 3.49 9.706 7.74 11.69 3.95 12.17 9.49 12.52 3.03 10.25 8.28 12.71 4.42 The initial individual alpha peak frequency of Subject 2 is calculated as 8.94Hz; the lower alpha frequency is 8.64Hz; the upper alpha frequency is 10.44 Hz (individual alpha band width 1.80 Hz). Compared with Subject 1, the individual alpha peak frequency of Subject 2 is smaller and the alpha bandwidth is narrower. The protocol for Subject 2 is to increase beta- 1/theta ratio during the. The feedback in the game is the same as for Subject 1. The alpha peak frequency, lower and upper alpha frequency, and the alpha bandwidth of Subject 2 before and after each neurofeedback session are given in Table II. The changes of alpha peak frequency and alpha bandwidth across different neurofeedback sessions are illustrated in Fig. 11 and 12. In Fig. 11 we can see that when compare the alpha peak of Subject 2 before the first and before the last, it is increased from 8.94 Hz to 9.55 Hz. Different from Subject 1, the individual alpha peak increased after each session in most of the case. In Fig. 12, we can see that the individual alpha bandwidth of Subject 2 increases immediately after the neurofeedback in most of the cases, and when we have a look at the changes across different days, the alpha bandwidth is increased from 1.8 Hz to 2.55 Hz. The difference between Subject 1 and 2 could be due to 1) the beta-1/theta ratio is more efficient than upper alpha ; 2) there is more room for improvement for Subject 2 as at the beginning the alpha peak frequency of Subject 2 is smaller and alpha bandwidth is narrower than Subject 1. TABLE II. 1 2 3 4 5 6 Training THE CHANGES ACROSS 6 NEUROFEEDBACK TRAINING SESSIONS OF SUBJECT 2 Alpha Peak Lower alpha Upper alpha Alpha Bandwidth 8.94 8.64 10.44 1.80 9.14 8.50 10.64 2.14 9.17 8.92 10.46 1.54 9.27 8.78 11.22 2.44 9.13 8.99 10.77 1.78 9.33 8.72 10.88 2.16 9.33 8.94 10.92 1.98 9.49 8.57 10.3 1.73 9.52 9.21 10.03 0.82 9.46 8.77 10.03 1.26 9.55 8.05 10.6 2.55 9.46 8.85 10.32 1.47 The initial individual alpha peak frequency of Subject 3 is calculated as 9.24 Hz; the lower alpha frequency is 8.72Hz; the upper alpha frequency is 9.67 Hz (individual alpha band width 0.95 Hz). In total she attended 6 neurofeedback sessions based on beta-1/theta ratio. The detailed alpha peak frequency, lower and upper alpha frequency, and the alpha bandwidth of Subject 3 before and after each neurofeedback session are given in Table III. The changes of alpha peak frequency and alpha bandwidth across different neurofeedback sessions are illustrated in Fig. 13 and 14. In Fig. 13, it shows that when we compare the alpha peak frequency of Subject 3 before the first and before the last, the alpha peak is increased from 9.24 Hz to 9.66 Hz. As it is illustrated in Fig. 14, the individual alpha bandwidth is boosted after the 4th session, from 0.95 Hz to 166

1.63 Hz. However, there is no increase of individual alpha bandwidth between the first and the last. Figure 11. The changes of alpha peak frequency among different neurofeedback sessions from Subject 2. Figure 14. The changes of alpha bandwidth among different neurofeedback sessions from Subject 3. Figure 12. The changes of alpha bandwidth among different neurofeedback sessions from Subject 2. TABLE III. 1 2 3 4 5 6 Training THE CHANGES ACROSS 6 NEUROFEEDBACK TRAINING SESSIONS OF SUBJECT 3 Alpha Peak Lower alpha Upper alpha Alpha Bandwidth 9.24 8.72 9.67 0.95 8.83 8.46 9.44 0.98 9.25 8.86 9.86 1.00 9.55 9.1 9.94 0.84 10.03 9.03 10.33 1.30 9.27 8.3 9.58 1.28 9.53 8.74 9.97 1.23 9.25 8.78 10.41 1.63 9.35 8.8 9.85 1.05 8.99 8.8 9.8 1.00 9.66 8.91 9.81 0.90 9.02 8.88 9.55 0.67 Figure 13. The changes of alpha peak frequency among different neurofeedback sessions from Subject 3. VI. CONCLUSION In this work, we proposed and implemented a neurofeedback system. It consists of the neurofeedback algorithms module and neurofeedback game Shooting. Compared with the other existing systems, the proposed one allows the users to select different neurofeedback protocols such as EEG power (upper alpha, theta/beta ratio, etc.) and fractal dimension based. It also contains a shooting game which makes the more attractive and interesting. We also did the preliminary study with 3 subjects taking 6 neurofeedback sessions each to assess the effectiveness of the proposed neurofeedback system. Different neurofeedback 167

protocols were applied: Subject 1 was trained to increase the individual alpha power and Subject 2 and 3 were trained to enhance beta-1/theta ratio. The results show that all subjects had a higher individual alpha peak frequency after the that indicates an enhancement of the subjects cognitive abilities related to multi-tasking such as attention, memory, etc. Subject 2 had a wider alpha bandwidth after the that indicates an improvement of the subject s cognitive abilities related to creativity. In the next step, we will increase number of subjects and compare alpha, theta/beta and fractal dimension based neurofeedback protocols to get deeper understanding of the effect of the neurofeedback. ACKNOWLEDGMENT This research was done for Fraunhofer IDM@NTU, which is funded by the National Research Foundation (NRF) and managed through the multi-agency Interactive & Digital Media Programme Office (IDMPO) hosted by the Media Development Authority of Singapore (MDA). The Shooting game is original and designed by NTU final year student Qiuyu Xiang using UDK game engine. REFERENCES [1] T. Fernández, et al., "Changes in EEG Current Sources Induced by Neurofeedback in Learning Disabled Children. An Exploratory Study," Applied Psychophysiology and Biofeedback, vol. 32, pp. 169-183, 2007. [2] B. Zoefel, R. J. Huster, and C. S. Herrmann, "Neurofeedback of the upper alpha frequency band in EEG improves cognitive performance," NeuroImage, vol. 54, pp. 1427-1431, 2011. [3] S. Hanslmayr, P. Sauseng, M. Doppelmayr, M. Schabus, and W. Klimesch, "Increasing individual upper alpha power by neurofeedback improves cognitive performance in human subjects," Applied Psychophysiology and Biofeedback, vol. 30, pp. 1-10, 2005. [4] T. Egner and J. H. Gruzelier, "Learned self-regulation of EEG frequency components affects attention and event-related brain potentials in humans," NeuroReport, vol. 12, pp. 4155-4159, 2001. [5] W. Klimesch, M. Doppelmayr, H. Schimke, and T. Pachinger, "Alpha frequency, reaction time, and the speed of processing information," Journal of Clinical Neurophysiology, vol. 13, pp. 511-518, 1996. [6] O. Bazanova and L. Aftanas, "Individual EEG alpha activity analysis for enhancement neurofeedback efficiency: Two case studies," Journal of Neurotherapy, vol. 14, pp. 244-253, 2010. [7] S. Sanei and J. Chambers, EEG signal processing. Chichester, England ; Hoboken, NJ: John Wiley & Sons, 2007. [8] J. H. Gruzelier, "EEG-neurofeedback for optimising performance. I: A review of cognitive and affective outcome in healthy participants," Neuroscience & Biobehavioral Reviews, 2013. [9] J. Becerra, et al., "Neurofeedback in healthy elderly human subjects with electroencephalographic risk for cognitive disorder," Journal of Alzheimer's Disease, vol. 28, pp. 357-367, 2012. [10] J. F. Lubar, "Neurofeedback for the management of attentiondeficit/hyperactivity disorders," in Biofeedback: A practitioner's guide (2nd ed.), ed New York: Guilford Press, 1995, pp. 493-522. [11] A. R. Clarke, R. J. Barry, R. McCarthy, and M. Selikowitz, "Electroencephalogram differences in two subtypes of Attention Deficit/Hyperactivity Disorder," Psychophysiology, vol. 38, pp. 212-221, 2001. [12] C. Escolano, M. Aguilar, and J. Minguez, "EEG-based upper alpha neurofeedback improves working memory performance," in Engineering in Medicine and Biology Society, EMBC, 2011 Annual International Conference of the IEEE, 2011, pp. 2327-2330. [13] D. Sammler, M. Grigutsch, T. Fritz, and S. Koelsch, "Music and emotion: Electrophysiological correlates of the processing of pleasant and unpleasant music," Psychophysiology, vol. 44, pp. 293-304, 2007. [14] T. Egner and J. H. Gruzelier, "EEG Biofeedback of low beta band components: frequency-specific effects on variables of attention and event-related brain potentials," Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology, vol. 115, pp. 131-139, 2004. [15] O. Bazanova and L. Aftanas, "Individual measures of electroencephalogram alpha activity and non-verbal creativity," Neuroscience and Behavioral Physiology, vol. 38, pp. 227-235, 2008. [16] C. Escolano, M. Aguilar, and J. Minguez, "EEG-based upper alpha neurofeedback improves working memory performance," in Engineering in Medicine and Biology Society,EMBC, Annual International Conference of the IEEE, 2011, pp. 2327-2330. [17] D. Vernon, et al., "The effect of distinct neurofeedback protocols on aspects of cognitive performance," International Journal of Psychophysiology, vol. 47, pp. 75-85, 2003. [18] T. Egner and J. H. Gruzelier, "Ecological validity of neurofeedback: modulation of slow wave EEG enhances musical performance," NeuroReport, vol. 14, pp. 1221-1224, 2003. [19] J.-R. Wang and S. Hsieh, "Neurofeedback improves attention and working memory performance," Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology, vol. 124, pp. 2406-2420, 2013. [20] A. T. Pope and E. H. Bogart, "Extended Attention Span Training System: Video Game Neurotherapy for Attention Deficit Disorder," Child Study Journal, vol. 26, pp. 39-50, 1996. [21] T. Fuchs, N. Birbaumer, W. Lutzenberger, J. H. Gruzelier, and J. Kaiser, "Neurofeedback treatment for attention-deficit/hyperactivity disorder in children: a comparison with methylphenidate," Applied Psychophysiology and Biofeedback, vol. 28, pp. 1-12, 2003. [22] W. Qiang, S. Olga, and N. Minh Khoa, "Fractal dimension based algorithm for neurofeedback games," in CGI 2010, Singapore, 2010, SP25. [23] Emotiv. Available: http://www.emotiv.com [24] American Electroencephalographic Society, "American electroencephalographic society guidelines for standard electrode position nomenclature," Journal of Clinical Neurophysiology, vol. 8, pp. 200-202, 1991. 168