- Email: [email protected]
Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task
Accepted Manuscript Title: Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task Authors: Chien-Chang Hsu, Ching-Wen Cheng, Yi-Shiuan Chiu PII: DOI: Reference:
S0304-3940(17)30032-0 http://dx.doi.org/doi:10.1016/j.neulet.2017.01.022 NSL 32564
To appear in:
Neuroscience Letters
Received date: Revised date: Accepted date:
27-8-2016 6-12-2016 10-1-2017
Please cite this article as: Chien-Chang Hsu, Ching-Wen Cheng, Yi-Shiuan Chiu, Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task, Neuroscience Letters http://dx.doi.org/10.1016/j.neulet.2017.01.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Highlights Beta waves are significantly correlated with external stimulus and cognitive responses. EEG feature analysis and k-means algorithm are used to analyze the difference between young musicians and non-musicians. The feature value of response time, response intensity, and response power of the young musicians were superior than the non-musicians.
1
Title: Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task Author names and affiliations: Author 1: Chien-Chang Hsu Department of Computer Science and Information Engineering, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Author 2: Ching-Wen Cheng Department of Computer Science and Information Engineering, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Author 3: Yi-Shiuan Chiu Department of Psychology, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Corresponding author: Name: Chien-Chang Hsu Address: Department of Computer Science and Information Engineering, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Telephone: 886-2-29052797 Fax: 886-2-29053876 Email: [email protected] Conflict of Interest Disclosure Statement: We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have been influenced its outcome. We confirm that the manuscript has been read and approve by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order listed in the manuscript has been approved by all of the authors.
2
Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task Chien-Chang Hsu1, Ching-Wen Cheng1, Yi-Shiuan Chiu2 1 Department of Computer Science and Information Engineering 2 Department of Psychology Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan
Abstract Electroencephalograms can record wave variations in any brain activity. Beta waves are produced when an external stimulus induces logical thinking, computation, and reasoning during consciousness. This work uses the beta wave of major scale working memory N-back tasks to analyze the differences between young musicians and non-musicians. After the feature analysis uses signal filtering, Hilbert-Huang transformation, and feature extraction methods to identify differences, k-means clustering algorithm are used to group them into different clusters. The results of feature analysis showed that beta waves significantly differ between young musicians and non-musicians from the low memory load of working memory task. Keywords: Electroencephalograms, beta waves, musicians, major scale, working memory task, feature extraction, clustering analysis
1. Introduction Musical competence depends on the ability to perform cognitive processing of musical sounds, and domain-general cognitive processes, including verbal and non-verbal skills [1,2]. Musicians also have enhanced working memory according measurements during N-back task [3]. They found that the consistent group difference was shown on the more difficult 2-back condition by behavioral and fMRI measures. This study developed a feature analysis method of EEG that can differentiate between musicians and non-musicians by 0-back and 1-back tasks. 3
Electroencephalograms (EEGs) have been widely used to diagnose various diseases and are now being used in innovative diagnostic methods [4-9]. The brainwave frequencies of EEGs are usually divided into five wavebands: delta (0-4 Hz), theta (4-8 Hz), alpha (8-12 Hz), sigma (12-16 Hz), and beta (16-32 Hz). Many studies agree that, in various situations and conditions, beta waves significantly correlate with stimulatory and cognitive responses to external environments and mental states when awake, which vary among people according to individual differences and environmental factors [10-13]. For example, Hale [14] reported an abnormally increased rightward beta (16-21 Hz) asymmetry at the inferior parietal region in patients with attention deficit hyperactivity disorder performing continuous performance tasks. A study by Zaretskaya [15] analyzed the cognition and attention of participants and their correlation with beta waves when the patients viewed various rotational movement patterns of white and black dots. The analytical results showed that the beta-band power in the parietal areas of the participants decreased after the experimental tasks were completed. Perrier [16] indicated that the beta power of drivers with insomnia were lower than those of drivers without insomnia disorder. Cognition and responses are often assessed by performance of N-back memory tasks. In N-back memory tasks, the working memory in the brain is used as the primary EEG data. Participants perform a series of stimulatory tasks in pre-established test procedures so that data for their short-term memory can be collected [17-19]. In aural working memory task, it compares the working memory EEG physiological signals generated by participants in response to stimuli such sounds, music, and various tones and melodies. By using N-back memory tasks and event-related potentials to analyze the relationship between noise and arousal, Han et al. [20] indicated that the response speed of participants was faster with low arousal than with high arousal, and that the response speed decreased as task difficulty increased. In a study of the effect of music lessons on the aural working memories of children diagnosed with hearing disabilities, Rochette [21] found that long-term music lessons improved auditory performance. Strait [22] compared musical aptitude between people with 4
average and below average reading abilities by using intermediate measures of music audition. The comparisons showed that music aptitude and reading and writing abilities correlate with aural working memory performance and attention. The EEG feature analysis developed in this work was used to analyze the major scale working memories of musicians and non-musicians. A series of major scale working memory tasks were used to stimulate the EEG signals of the participants. Specifically, the feature analysis was used to analyze changes in the EEG features of the musicians and non-musicians performing the tasks. Finally, the post experimental EEG features were discussed based on different working memory tasks.
2. Methods 2.1 Participants Sixteen trained musicians (4 males, age range: 18-30 yrs, mean: 22.9 yrs) from the department of music at Fu-Jen catholic university were enrolled to join the experiments. The mean duration of musicians had commenced musical study was 11.68 years. All musicians actively pursued multiple instruments while maintaining one principal instrument (piano: 6 persons, cello: 2 persons, viola: 2 persons, violin: 2 persons, France horn: 1 person, trombone: 1 person, clarinet: 1 person, oboe: 1 person). Sixteen non-musician students (4 males, age range: 20-30 yrs, mean age: mean: 21.6 yrs) from different department of the same university volunteered to participate in the experiment. All participants reported that they are right handedness. None had been diagnosed with a severe physical or mental illness. 2.2 Measures All participants performed N-back memory tasks of tones in this study. According to the Western tonal music theory, the five tones of major scales were used as the stimuli. The task included two blocks: 0-back and 1-back. Participants were asked to decide whether later the presenting tones matched the single specific tone presented at first (0-back) or the previous one tones (1-back). In 5
each block, 128 stimuli were given (32 × 4 rounds). Each block was separated by a break of 10-20 s. After the fixation, each stimulus was shown for 1200 msec, and the interstimuli interval was 800 msec. In each block, the presentation of five tones was random and could not composed into any familir music or songs. 2.3 EEG The electroencephalogram was measured in a quiet room at Fu-Jen catholic university. Data were recorded with a 32 Ag/AgCl electrodes mounted on an electrode cape using SynAmps amplifiers (Neuroscan Labs, Sterling, USA) and arranged according the international 10/20 systems. Sampling rate was 500 Hz. A 0.1-30 Hz bandpass filter was applied. 2.4 EEG feature analysis The EEG feature analysis contains two main tasks: signal preprocessing and feature analysis. First, the signal preprocessing step mainly captured EEG signals, filtering noise, and calculating the proportions of the beta waves in the signals. After the C3-A2 signals were extracted from the original EEG signals, the noises were filtered, and the baseline was corrected [9]. To filter the noise, a band pass filter was used to extract the EEG signals from the 0.5-32 Hz band of the original signals for baseline correction by using the following equation: 𝐿𝑛𝑒𝑤 [t] = 𝐿𝑜𝑙𝑑 [t] − 𝐿𝑏𝑎𝑠𝑒 [𝑡]
(1)
where Lnew [t] represents the corrected signal, Lold [t] represents the filtered C3-A2 signal, and Lbase [t] represents the signal baseline. The equation for the baseline was as follows: 𝐿𝑏𝑎𝑠𝑒 [𝑡] = 𝑚𝑒𝑑𝐹𝑖𝑙𝑡(𝑚𝑒𝑑𝐹𝑖𝑙𝑡(𝐿𝑜𝑙𝑑 [𝑡], 𝑇1 ), 𝑇2 )
(2)
where Lbase [t] is the signal baseline, medFilt() is the median filter function, and T1 and T2 are the numbers of frequency points of the filtered samples. The proportions of the beta waves in the EEG signals [9] were calculated using Hilbert-Huang transform [23] and smoothed using the following equation: 𝛽𝑟𝑎𝑡𝑖𝑜 [𝑡] =
𝑅𝛽 [𝑡] 𝑅𝛿 [𝑡]+𝑅𝛾 [𝑡]+𝑅𝛼 [𝑡]+𝑅∑ [𝑡]+𝑅𝛽 [𝑡]
6
(3)
where βratio [t] represents the proportion of Hilbert-Huang transform beta waves and where Rδ [t], Rγ [t], Rα [t], R∑ [t], and Rβ[t] are the proportions of delta, theta, alpha, sigma, and beta waves, respectively. Figure 1 shows the beta waves in N-back testing. Notably, the starting test time begin at the 30 second.
Fig. 1 Beta waves in N-back test at 30 seconds
The feature analysis included the features of response speed (S), response intensity (I), and response power (P) of the beta waves. The response speed (S) was defined as the first descending time of the beta waves after the beginning of the N-back memory tasks.
S Ratio (i 1) Ratio (i) 0, if i 30
(4)
where Ratio(i) represents the amplitude of beta wave at the ith second of N-back task and 30 is the starting time of N-back task. The response intensity (I) was defined as the area under the beta waves of the participants at the beginning of the N-back tasks. The interval from the starting point of the test to 30 s after the starting point was adopted, and the area under the curve in the interval was calculated. I
60
i 30
Ratio (i)
(5)
where Ratio(i) represents the amplitude of beta waves at the ith second of N-back task. The response power (P) was computed as the power spectral density [24] during the first 30 seconds of the test. 7
P=|
1
∞
∫ 𝑓(𝑡)𝑒 −𝑖𝜔𝑡 𝑑𝑡 |2 ＝ (F(ω) ∗ F ∗ (ω))/2𝜋 √2𝜋 −∞
(6)
where f(t) is the finite energy signal, ω is the angular frequency, t is a random real number, F(ω) is the continuous Fourier transform of f(t), and F*(ω) is the conjugate function of F(ω).
3. Results After the EEG signals were preprocessed, the beta wave proportions were calculated to compare the waveforms of the participants before and during the N-back memory tasks. Tables 1 to 3 compare the feature values in N-back experiments between musicians and non-musicians. The mixed design ANOVA in SPSS was used to analyze the group and task difficulty on response speed, response intensity, and response power. Group significantly affected all three parameters, and indicated at the bottom of Table 3. Table 4 shows the clustered result of all participants by k-means algorithm [25]. Notably, all participants were clustered into two clusters, cluster 1 and 2. All the trained musicians were clustered in the cluster 2. The case number of cluster 1 and 2 was 11 and 21, respectively. Five non-musicians were clustered as potential musicians, i.e., case 3, 4, 13, 14, and 15. The accuracy of classification was 84.38%. The mixed design ANOVA in SPSS was used to analyze the cluster and task difficulty on response speed, response intensity, and response power. Again, the main effects of cluster were significantly on all three parameters and shown at the bottom of Table 4. The members for each group and those for each cluster were quite overlapped and significantly correlated [r =.72, p <.01]. The probability, predicted by the feature analysis, did not differ significantly from the probability of the original groups [X2 (1) = 3.13, p >.05], revealing that the feature analysis effectively placed the participants into their original group.
8
Case# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Case# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Table 1 Feature values of musicians Response speed (S) Response intensity (I) Response power (P) 0-back (V1) 1-back (V2) 0-back (V3) 1-back (V4) 0-back (V5) 1-back (V6) 8.1 10.1 1.1 1.7 0 0 6.9 9.2 0.8 1.4 0 0 4.9 7.8 0.4 1.2 5 4 8.6 4.7 1.4 0.4 0 0 6.5 5.6 0.8 0.5 0 0 6.8 9.0 0.7 1.4 1 1 6.1 5.0 0.6 0.4 3 7 7.3 6.8 1.2 0.9 3 0 5.2 5.5 0.4 0.5 3 0 10.0 4.2 2.3 0.3 0 0 7.7 8.0 0.9 1.2 4 0 4.5 5.0 0.3 0.4 4 5 3.2 5.3 0.2 0.4 0 7 5.0 6.0 0.4 0.9 6 0 9.6 7.0 1.6 0.9 0 2 9.3 9.5 1.5 1.4 9 8 Table 2 Feature values of non-musicians Response speed (S) Response intensity (I) Response power (P) 0-back (V1) 1-back (V2) 0-back (V3) 1-back (V4) 0-back (V5) 1-back (V6) 2 20 5.7 4.1 0.5 0.3 19 7 6.1 5.8 0.7 0.6 8 0 4.8 5.4 0.4 0.5 0 12 6.1 4.3 0.9 0.3 0 28 4.1 4.6 0.3 0.4 10 10 3.9 4.2 0.3 0.3 0 17 6.7 4.8 0.8 0.4 12 17 3.7 5.0 0.2 0.5 30 13 7.5 5.0 0.9 0.5 0 18 4.7 4.0 0.4 0.3 21 18 6.7 7.5 1.0 1.2 20 6 6.2 8.7 0.7 1.3 0 8 6.2 7.1 0.7 0.9 4 9 4.8 2.4 0.5 0.1 7 5 5.7 4.8 0.6 0.4 0 24 3.9 4.5 0.2 0.4 9
Table 3 Feature values of musicians and non-musician significantly differed for three parameters (mean standard deviation) Response speed (S)
Response intensity (I)
Response power (P)
Group Musicians Non-musicians Main effect of group
0-back (V1) 2.42.6
1-back (V2) 0-back (V3) 2.12.9 6.82.0
8.39.6 13.37.5 F (1,30) = 36.39, p < .01
1-back (V4) 6.81.9
5.41.2 5.11.5 F (1,30) = 10.41, p < .01
0-back (V5) 0.90.6
1-back (V6) 0.90.5
0.60.3 0.50.3 F (1,30) = 9.30 , p < .01
Table 4 Mean values of two clusters computed by feature analysis significantly differed for three parameters (mean standard deviation) Response speed (S)
Response intensity (I)
Response power (P)
Cluster# 1 (21 persons) 2 (11 persons) Main effect of cluster
0-back (V1) 2.73.0 10.410.4
1-back (V2) 3.23.9 16.26.8
F (1,30) = 192.57, p < .01
0-back (V3) 6.51.8 5.41.4
1-back (V4) 6.32.0 5.31.5
F (1,30) = 458.28, p < .01
0-back (V5) 0.80.5 0.50.3
1-back (V6) 0.80.5 0.60.4
F (1,30) = 112.98, p < .01
4. Discussion and conclusion This study developed a feature analysis method of EEG to differentiate between musicians and non-musicians on the easier working memory N-back conditions. The results of this study revealed that musicians and non-musicians had similar EEG features upon the initial stimulation of the test, indicating that the participants’ EEG signals also generated features following the stimulations during the working memory tasks. Compared to non-musicians, the musicians had a faster response speed, a higher response intensity, and a higher response power. Before these results can be explained, the possible cognitive functions of beta waves and the features in our analysis must be clarified. Beta waves are characteristic of a strongly engaged mind. Beta frequency responses have been observed in processing difficult tasks [26], motor planning [27], memory retrieval [28], and memory load [28,29]. A person with a high proportion of beta brainwaves may be able to think fast, concentrate, feel excited, and remain in a state of high cognitive functioning [26]. Moreover, a high intensity of beta brainwaves may cause muscular tension and high blood pressure. In this investigation, feature analysis of response speed, response intensity, and response power 10
of beta waves, was conducted to classify musicians and non-musicians. The response speed, which is determined by the time of the first descent of the beta waves after the beginning of the N-back memory tasks, may reflect the speed with which a participant’s state of arousal decreases and the participant adapts to the demands of the task or the associated cognitive loading. The response intensity, which is given by the area under the beta waves, may reveal the accumulation energy by a participant during the task. The response power is the power spectral density of beta waves and refers to the spectral energy distribution of beta waves. The EEG beta wave responses of musicians shifted more rapidly than those of non-musicians and their psychological states changed from tense to relax more rapidly, too. In addition, the musicians herein were more focused and sensitive to tasks than the non-musicians. Possibly, during working memory tasks, musicians used more effective strategies for encoding and manipulating sounds, even with no melody [30]. Therefore, they might have performed the memory tasks more efficiently than non-musicians. However, we could not exclude the possibility that the EEG beta wave responses seen in musicians might be a phenomenon associated with their better motor control and ability to sustain attention, which might result in less effort needed on the task. This should be clarified in future studies. The improved working memory of musicians has been associated with greater brain activities in the right lateral prefrontal cortex, the right lateral parietal cortex, the bilateral posterior dorsal prefrontal cortex, and the bilateral anterior cingulate gyrus [3], as well as a larger P300 amplitude [31]. Furthermore, long-term extensive music training has been found to influence the anatomy of the brain. Professional musicians have been found to have more gray matter in motor-related and auditory-related areas [32]. Music practice has been found to provide cognitive improvements not only in working memory [3,30,31,33-37], but also in reasoning [1], visual attention [38], and processing speed [39]. In the future, the feature analysis model could be used to differentiate the EEG physiological signals of 11
musicians from those of non-musicians in other cognitive tasks. In summary, the working memory EEG analysis can reveal short-term amplitude responses but may not reveal long-term EEG responses and changes. This study analyzed response speed, response intensity, and response power to identify changes in the EEG features in the 30-s interval beginning from the initial stimulation of the test. Moreover, the working memory EEG feature analysis was unitary and typically used to compare the amplitudes of the stimulations in the original EEG signals or to identify similarities in the waveforms. By capturing and analyzing multiple EEG features and eliminating noise by using the EEG proportions, this study can be used as the basis to realize the development of the talent musician discovering system.
Acknowledgements This work is partly supported by the Fu-Jen catholic university and ministry of science and technology of Taiwan under grants MOST 103-2221-E-030-018.
References [1] M. Forgeard, E. Winner, A. Norton and G. Schlaug, "Practicing a musical instrument in childhood is associated with enhanced verbal ability and nonverbal reasoning," PLoS One. 3(10)(2008), e3566. doi:10.1371/journal.pone.0003566. [2] T. Fujioka, L.J. Trainor, B. Ross, R. Kakigi and C. Pantev, "Musical training enhances automatic encoding of melodic contour and interval structure," J Cogn Neurosci. 16(6) (2004) 1010-1021. doi:10.1162/0898929041502706. [3] K.J. Pallesen, E. Brattico, C.J. Bailey, A. Korvenoja, J. Koivisto, A. Gjedde and S. Carlson, "Cognitive control in auditory working memory is enhanced in musicians," PLoS One. 5(6) (2010) e11120. doi:10.1371/journal.pone.0011120. [4] J. Wasifa, D. Saptarshi, M. Koushik, P. Indranil and K. Doga, “Brain connectivity analysis from 12
EEG signals using stable phase-synchronized states during face perception tasks,” Physica A 434 (2015) 273-295. doi: 10.1016/j.physa.2015.03.087. [5] G. Arnulfo, J. Hirvonen, L. Nobili, S. Palva and J.M. Palva, “Phase and amplitude correlations in resting-state activity in human stereotactical EEG recordings,” Neuroimage 112 (2015) 114-127. doi: : 10.1016/j.neuroimage.2015.02.031. [6] M. Evan, G. Hatice and K. Nick Bryan, “Using affective and behavioural sensors to explore aspects of collaborative music making,” Int. J. Human-Computer Studies 82 (2015) 31-47. doi: 10.1016/j.ijhcs.2015.05.002. [7] A. Mathieu, S. Mazza, D. Petit, A. Décary, J. Massicotte-Marquez, J. Malo and J. Montplaisir, “Does age worsen EEG slowing and attention deficits in obstructive sleep apnea syndrome?” Clin Neurophysiol. 118(7) (2007) 1538-1544. doi: 10.1016/j.clinph.2007.04.009 [8] C.C. Kung, T.H. Hsieh, J.Y. Liou, K.J. Lin, F.Z. Shaw and S.F. Liang, “Musicians and non-musicians’ different reliance of features in consonance perception: A behavioral and ERP study,” Clin Neurophysiol. 125(5) (2014) 971-978. doi: 10.1016/j.clinph.2013.10.016. [9] C.C. Hsu, J.H. Wu, H.C. Chiu and C.M. Lin, “Evaluating the sleep quality of obstructive sleep apnea patients after continuous positive airway pressure treatment,” Comput Biol Med. 43(7) (2013) 870-878. doi: 10.1016/j.compbiomed.2013.04.006. [10] B. Guntekin, D.D. Emek-Savas, P. Kurt, G.G. Yener and E. Basar, “Beta oscillatory responses in healthy subjects and subjects with mild cognitive impairment,” Neuroimage Clin. 3 (2013) 39-46. doi: 10.1016/j.nicl.2013.07.003. [11] A. Wróbel, “Beta activity: a carrier for visual attention,” Acta Neurobiol Exp (Wars). 60(2) (2000) 247-260. [12] L.F. Basile, J.R. Sato, M.Y. Alvarenga, N.Jr. Henrique, H.A. Pasquini, W. Alfenas, S. Machado, B. Velasques, P. Ribeiro, R. Piedade, R. Anghinah and R.T. Ramos, “Lack of systematic topographic difference between attention and reasoning beta correlates,” PLoS One. 8(3) (2013) e59595. doi: 10.1371/journal.pone.0059595. 13
[13] F. van Ede, F. de Lange, O. Jensen and E. Maris, “Orienting attention to an upcoming tactile event involves a spatially and temporally specific modulation of sensorimotor alpha- and beta-band
oscillations,”
J
Neurosci.
31(6)
(2011)
2016-2024.
doi:
10.1523/JNEUROSCI.5630-10.2011. [14] T.S. Hale, S.L. Smalley, P.D. Walshaw, G. Hanada, J. Macion, J.T. McCracken, J.J. McGough and S.K. Loo, “Atypical EEG beta asymmetry in adults with ADHD,” Neuropsychologia. 48(12) (2010) 3532-3539. doi: 10.1016/j.neuropsychologia.2010.08.002. [15] N. Zaretskaya and A. Bartels, “Gestalt perception is associated with reduced parietal beta oscillations,” Neuroimage. 112 (2015) 61-69. doi:10.1016/j.neuroimage.2015.02.049. [16] J. Perrier, T.R. Leufkens, J. Bulla, S. Jongen, M.L. Bocca, J.G. Ramaekers and A. Vermeeren, “Electroencephalography during on-the-road driving in older untreated insomnia patients and normal sleepers,” Biol Psychol. 109 (2015) 20-28. doi: 10.1016/j.biopsycho.2015.04.002. [17] A. Gevins, L.K. McEvoy, M.E Smith, C.S. Chan, L. Sam-Vargas, C. Baum and A.B. Ilan, “Long-term and within-day variability of working memory performance and EEG in individuals,” Clin Neurophysiol. 123(7) (2012) 1291-1299. doi: 10.1016/j.clinph.2011.11.004. [18] P. Ruhnau, N. Wetzel, A. Widmann and E. Schröger, “The modulation of auditory novelty processing by working memory load in school age children and adults: a combined behavioral and
event-related
potential
study,”
BMC
Neurosci.
11
(2010)
126.
doi:
10.1186/1471-2202-11-126. [19] P.J. Maes, M.M. Wanderley and C. Palmer, “The role of working memory in the temporal control of discrete and continuous movements,” Exp Brain Res. 233(1) (2015) 263-73. doi: 10.1007/s00221-014-4108-5. [20] L. Han, Y. Liu, D. Zhang, Y. Jin and Y. Luo, “Low-arousal speech noise improves performance in
n-back
task:
an
ERP
study,”
PLoS
One.
8(10)
(2013)
e76261.
doi:
10.1371/journal.pone.0076261. [21] F. Rochette, A. Moussard and E. Bigand, “Music lessons improve auditory perceptual and 14
cognitive performance in deaf children,” Front Hum Neurosci. 8(2014) 488. doi: 10.3389/fnhum.2014.00488. [22] D.L. Strait, J. Hornickel and N. Kraus, “Subcortical processing of speech regularities underlies reading and music aptitude in children,” Behav Brain Funct. 7(2011) 44. doi: 10.1186/1744-9081-7-44. [23] N.E. Huang and S.S. P Shen, Hilbert Huang transform and its applications: 2nd Edition, World Scientific, Singapore, 2014. [24] J. Leis, Digital signal processing using MATLAB for students and researchers, Wiley, New York, 2011. [25] R Core Team, R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria, 2013. http://www.R-project.org/. [26] J.B. Brookings, G.F. Wilson and C.R. Swain, “Psychophysiological responses to changes in workload during simulated air traffic control,” Biol Psychol. 42(3) (1996) 361-377. doi: 10.1016/0301-0511(95)05167-8. [27] J. Kaiser, N. Birbaumer and W. Lutzenberger, “Event-related beta desynchronization indicates timing of response selection in a delayed-response paradigm in humans,” Neurosci Lett. 312(3) (2001) 149-152. doi:10.1016/S0304-3940(01)02217-0. [28] M. Karrasch, M. Laine, P. Rapinoja and C.M. Krause, “Effects of normal aging on event-related desynchronization/synchronization during a memory task in humans,” Neurosci Lett. 366(1) (2004) 18-23. doi:10.1016/j.neulet.2004.05.010. [29] M. Pesonen, C.H. Björnberg, H. Hämäläinen and C.M. Krause, “Brain oscillatory 1-30 Hz EEG ERD/ERS responses during the different stages of an auditory memory search task,” Neurosci Lett. 399(1-2) (2006) 45-50. doi:10.1016/j.neulet.2006.01.053. [30] K. Schulze, K. Mueller and S. Koelsch, “Neural correlates of strategy use during auditory working memory in musicians and non-musicians,” Eur J Neurosci. 33(1)(2011) 189-196. doi:10.1111/j.1460-9568.2010.07470.x. 15
[31] E.M. George and D. Coch, “Music training and working memory: An ERP study,” Neuropsychologia 49(5) (2011) 1083-1094. doi:10.1016/j.neuropsychologia.2011.02.001. [32] K.L. Hyde, J. Lerch, A. Norton, M. Forgeard, E. Winner, A.C. Evans and G. Schlaug, “Musical training shapes structural brain development,” J Neurosci. 29(10) (2009) 3019-3025. doi:10.1523/jneurosci.5118-08.2009. [33] S.L. Bengtsson, Z. Nagy, S. Skare, L. Forsman, H. Forssberg and F. Ullen, “Extensive piano practicing has regionally specific effects on white matter development,” Nat Neurosci. 8(9) (2005) 1148-1150. doi: 10.1038/nn1516. [34] S. Bergman Nutley, F. Darki and T. Klingberg, “Music practice is associated with development of working memory during childhood and adolescence,” Front Hum Neurosci. 7(926) (2014) eCollection. doi:10.3389/fnhum.2013.00926. [35] J.A. Bugos, W.M. Perlstein, C.S. McCrae, T.S. Brophy and P.H. Bedenbaugh, “Individualized piano instruction enhances executive functioning and working memory in older adults,” Aging Ment Health. 11(4) (2007) 464-471. doi:10.1080/13607860601086504. [36] L. Suárez, S. Elangovan and A. Au, “Cross‐ sectional study on the relationship between music training and working memory in adults,” Aust J Psychol. 68(1) (2016) 38-46. doi: 10.1111/ajpy.12087. [37] L. Vandervert, “How music training enhances working memory: A cerebrocerebellar blending mechanism that can lead equally to scientific discovery and therapeutic efficacy in neurological disorders,” Cerebellum Ataxias. 2(1) (2015) 11. doi:10.1186/s40673-015-0030-2. [38] A.C. Rodrigues, M.A. Loureiro and P. Caramelli, “Long-term musical training may improve different forms of visual attention ability,” Brain Cogn. 82(3) (2013) 229-235. doi: 10.1016/j.bandc.2013.04.009. [39] E.G. Schellenberg and W.W. Michael, “Music and cognitive abilities,” In D. Deutsch (Ed.) The Psychology of Music: 3rd, vol. 12, pp. 499-550, Amsterdam, Elsevier, 2013.
16
S0304-3940(17)30032-0 http://dx.doi.org/doi:10.1016/j.neulet.2017.01.022 NSL 32564
To appear in:
Neuroscience Letters
Received date: Revised date: Accepted date:
27-8-2016 6-12-2016 10-1-2017
Please cite this article as: Chien-Chang Hsu, Ching-Wen Cheng, Yi-Shiuan Chiu, Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task, Neuroscience Letters http://dx.doi.org/10.1016/j.neulet.2017.01.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Highlights Beta waves are significantly correlated with external stimulus and cognitive responses. EEG feature analysis and k-means algorithm are used to analyze the difference between young musicians and non-musicians. The feature value of response time, response intensity, and response power of the young musicians were superior than the non-musicians.
1
Title: Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task Author names and affiliations: Author 1: Chien-Chang Hsu Department of Computer Science and Information Engineering, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Author 2: Ching-Wen Cheng Department of Computer Science and Information Engineering, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Author 3: Yi-Shiuan Chiu Department of Psychology, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Corresponding author: Name: Chien-Chang Hsu Address: Department of Computer Science and Information Engineering, Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan Telephone: 886-2-29052797 Fax: 886-2-29053876 Email: [email protected] Conflict of Interest Disclosure Statement: We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have been influenced its outcome. We confirm that the manuscript has been read and approve by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order listed in the manuscript has been approved by all of the authors.
2
Analyze the beta waves of electroencephalogram signals from young musicians and non-musicians in major scale working memory task Chien-Chang Hsu1, Ching-Wen Cheng1, Yi-Shiuan Chiu2 1 Department of Computer Science and Information Engineering 2 Department of Psychology Fu-Jen Catholic University 510 Chung Cheng Rd., Hsinchuang Dist., New Taipei City 242, Taiwan
Abstract Electroencephalograms can record wave variations in any brain activity. Beta waves are produced when an external stimulus induces logical thinking, computation, and reasoning during consciousness. This work uses the beta wave of major scale working memory N-back tasks to analyze the differences between young musicians and non-musicians. After the feature analysis uses signal filtering, Hilbert-Huang transformation, and feature extraction methods to identify differences, k-means clustering algorithm are used to group them into different clusters. The results of feature analysis showed that beta waves significantly differ between young musicians and non-musicians from the low memory load of working memory task. Keywords: Electroencephalograms, beta waves, musicians, major scale, working memory task, feature extraction, clustering analysis
1. Introduction Musical competence depends on the ability to perform cognitive processing of musical sounds, and domain-general cognitive processes, including verbal and non-verbal skills [1,2]. Musicians also have enhanced working memory according measurements during N-back task [3]. They found that the consistent group difference was shown on the more difficult 2-back condition by behavioral and fMRI measures. This study developed a feature analysis method of EEG that can differentiate between musicians and non-musicians by 0-back and 1-back tasks. 3
Electroencephalograms (EEGs) have been widely used to diagnose various diseases and are now being used in innovative diagnostic methods [4-9]. The brainwave frequencies of EEGs are usually divided into five wavebands: delta (0-4 Hz), theta (4-8 Hz), alpha (8-12 Hz), sigma (12-16 Hz), and beta (16-32 Hz). Many studies agree that, in various situations and conditions, beta waves significantly correlate with stimulatory and cognitive responses to external environments and mental states when awake, which vary among people according to individual differences and environmental factors [10-13]. For example, Hale [14] reported an abnormally increased rightward beta (16-21 Hz) asymmetry at the inferior parietal region in patients with attention deficit hyperactivity disorder performing continuous performance tasks. A study by Zaretskaya [15] analyzed the cognition and attention of participants and their correlation with beta waves when the patients viewed various rotational movement patterns of white and black dots. The analytical results showed that the beta-band power in the parietal areas of the participants decreased after the experimental tasks were completed. Perrier [16] indicated that the beta power of drivers with insomnia were lower than those of drivers without insomnia disorder. Cognition and responses are often assessed by performance of N-back memory tasks. In N-back memory tasks, the working memory in the brain is used as the primary EEG data. Participants perform a series of stimulatory tasks in pre-established test procedures so that data for their short-term memory can be collected [17-19]. In aural working memory task, it compares the working memory EEG physiological signals generated by participants in response to stimuli such sounds, music, and various tones and melodies. By using N-back memory tasks and event-related potentials to analyze the relationship between noise and arousal, Han et al. [20] indicated that the response speed of participants was faster with low arousal than with high arousal, and that the response speed decreased as task difficulty increased. In a study of the effect of music lessons on the aural working memories of children diagnosed with hearing disabilities, Rochette [21] found that long-term music lessons improved auditory performance. Strait [22] compared musical aptitude between people with 4
average and below average reading abilities by using intermediate measures of music audition. The comparisons showed that music aptitude and reading and writing abilities correlate with aural working memory performance and attention. The EEG feature analysis developed in this work was used to analyze the major scale working memories of musicians and non-musicians. A series of major scale working memory tasks were used to stimulate the EEG signals of the participants. Specifically, the feature analysis was used to analyze changes in the EEG features of the musicians and non-musicians performing the tasks. Finally, the post experimental EEG features were discussed based on different working memory tasks.
2. Methods 2.1 Participants Sixteen trained musicians (4 males, age range: 18-30 yrs, mean: 22.9 yrs) from the department of music at Fu-Jen catholic university were enrolled to join the experiments. The mean duration of musicians had commenced musical study was 11.68 years. All musicians actively pursued multiple instruments while maintaining one principal instrument (piano: 6 persons, cello: 2 persons, viola: 2 persons, violin: 2 persons, France horn: 1 person, trombone: 1 person, clarinet: 1 person, oboe: 1 person). Sixteen non-musician students (4 males, age range: 20-30 yrs, mean age: mean: 21.6 yrs) from different department of the same university volunteered to participate in the experiment. All participants reported that they are right handedness. None had been diagnosed with a severe physical or mental illness. 2.2 Measures All participants performed N-back memory tasks of tones in this study. According to the Western tonal music theory, the five tones of major scales were used as the stimuli. The task included two blocks: 0-back and 1-back. Participants were asked to decide whether later the presenting tones matched the single specific tone presented at first (0-back) or the previous one tones (1-back). In 5
each block, 128 stimuli were given (32 × 4 rounds). Each block was separated by a break of 10-20 s. After the fixation, each stimulus was shown for 1200 msec, and the interstimuli interval was 800 msec. In each block, the presentation of five tones was random and could not composed into any familir music or songs. 2.3 EEG The electroencephalogram was measured in a quiet room at Fu-Jen catholic university. Data were recorded with a 32 Ag/AgCl electrodes mounted on an electrode cape using SynAmps amplifiers (Neuroscan Labs, Sterling, USA) and arranged according the international 10/20 systems. Sampling rate was 500 Hz. A 0.1-30 Hz bandpass filter was applied. 2.4 EEG feature analysis The EEG feature analysis contains two main tasks: signal preprocessing and feature analysis. First, the signal preprocessing step mainly captured EEG signals, filtering noise, and calculating the proportions of the beta waves in the signals. After the C3-A2 signals were extracted from the original EEG signals, the noises were filtered, and the baseline was corrected [9]. To filter the noise, a band pass filter was used to extract the EEG signals from the 0.5-32 Hz band of the original signals for baseline correction by using the following equation: 𝐿𝑛𝑒𝑤 [t] = 𝐿𝑜𝑙𝑑 [t] − 𝐿𝑏𝑎𝑠𝑒 [𝑡]
(1)
where Lnew [t] represents the corrected signal, Lold [t] represents the filtered C3-A2 signal, and Lbase [t] represents the signal baseline. The equation for the baseline was as follows: 𝐿𝑏𝑎𝑠𝑒 [𝑡] = 𝑚𝑒𝑑𝐹𝑖𝑙𝑡(𝑚𝑒𝑑𝐹𝑖𝑙𝑡(𝐿𝑜𝑙𝑑 [𝑡], 𝑇1 ), 𝑇2 )
(2)
where Lbase [t] is the signal baseline, medFilt() is the median filter function, and T1 and T2 are the numbers of frequency points of the filtered samples. The proportions of the beta waves in the EEG signals [9] were calculated using Hilbert-Huang transform [23] and smoothed using the following equation: 𝛽𝑟𝑎𝑡𝑖𝑜 [𝑡] =
𝑅𝛽 [𝑡] 𝑅𝛿 [𝑡]+𝑅𝛾 [𝑡]+𝑅𝛼 [𝑡]+𝑅∑ [𝑡]+𝑅𝛽 [𝑡]
6
(3)
where βratio [t] represents the proportion of Hilbert-Huang transform beta waves and where Rδ [t], Rγ [t], Rα [t], R∑ [t], and Rβ[t] are the proportions of delta, theta, alpha, sigma, and beta waves, respectively. Figure 1 shows the beta waves in N-back testing. Notably, the starting test time begin at the 30 second.
Fig. 1 Beta waves in N-back test at 30 seconds
The feature analysis included the features of response speed (S), response intensity (I), and response power (P) of the beta waves. The response speed (S) was defined as the first descending time of the beta waves after the beginning of the N-back memory tasks.
S Ratio (i 1) Ratio (i) 0, if i 30
(4)
where Ratio(i) represents the amplitude of beta wave at the ith second of N-back task and 30 is the starting time of N-back task. The response intensity (I) was defined as the area under the beta waves of the participants at the beginning of the N-back tasks. The interval from the starting point of the test to 30 s after the starting point was adopted, and the area under the curve in the interval was calculated. I
60
i 30
Ratio (i)
(5)
where Ratio(i) represents the amplitude of beta waves at the ith second of N-back task. The response power (P) was computed as the power spectral density [24] during the first 30 seconds of the test. 7
P=|
1
∞
∫ 𝑓(𝑡)𝑒 −𝑖𝜔𝑡 𝑑𝑡 |2 ＝ (F(ω) ∗ F ∗ (ω))/2𝜋 √2𝜋 −∞
(6)
where f(t) is the finite energy signal, ω is the angular frequency, t is a random real number, F(ω) is the continuous Fourier transform of f(t), and F*(ω) is the conjugate function of F(ω).
3. Results After the EEG signals were preprocessed, the beta wave proportions were calculated to compare the waveforms of the participants before and during the N-back memory tasks. Tables 1 to 3 compare the feature values in N-back experiments between musicians and non-musicians. The mixed design ANOVA in SPSS was used to analyze the group and task difficulty on response speed, response intensity, and response power. Group significantly affected all three parameters, and indicated at the bottom of Table 3. Table 4 shows the clustered result of all participants by k-means algorithm [25]. Notably, all participants were clustered into two clusters, cluster 1 and 2. All the trained musicians were clustered in the cluster 2. The case number of cluster 1 and 2 was 11 and 21, respectively. Five non-musicians were clustered as potential musicians, i.e., case 3, 4, 13, 14, and 15. The accuracy of classification was 84.38%. The mixed design ANOVA in SPSS was used to analyze the cluster and task difficulty on response speed, response intensity, and response power. Again, the main effects of cluster were significantly on all three parameters and shown at the bottom of Table 4. The members for each group and those for each cluster were quite overlapped and significantly correlated [r =.72, p <.01]. The probability, predicted by the feature analysis, did not differ significantly from the probability of the original groups [X2 (1) = 3.13, p >.05], revealing that the feature analysis effectively placed the participants into their original group.
8
Case# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Case# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Table 1 Feature values of musicians Response speed (S) Response intensity (I) Response power (P) 0-back (V1) 1-back (V2) 0-back (V3) 1-back (V4) 0-back (V5) 1-back (V6) 8.1 10.1 1.1 1.7 0 0 6.9 9.2 0.8 1.4 0 0 4.9 7.8 0.4 1.2 5 4 8.6 4.7 1.4 0.4 0 0 6.5 5.6 0.8 0.5 0 0 6.8 9.0 0.7 1.4 1 1 6.1 5.0 0.6 0.4 3 7 7.3 6.8 1.2 0.9 3 0 5.2 5.5 0.4 0.5 3 0 10.0 4.2 2.3 0.3 0 0 7.7 8.0 0.9 1.2 4 0 4.5 5.0 0.3 0.4 4 5 3.2 5.3 0.2 0.4 0 7 5.0 6.0 0.4 0.9 6 0 9.6 7.0 1.6 0.9 0 2 9.3 9.5 1.5 1.4 9 8 Table 2 Feature values of non-musicians Response speed (S) Response intensity (I) Response power (P) 0-back (V1) 1-back (V2) 0-back (V3) 1-back (V4) 0-back (V5) 1-back (V6) 2 20 5.7 4.1 0.5 0.3 19 7 6.1 5.8 0.7 0.6 8 0 4.8 5.4 0.4 0.5 0 12 6.1 4.3 0.9 0.3 0 28 4.1 4.6 0.3 0.4 10 10 3.9 4.2 0.3 0.3 0 17 6.7 4.8 0.8 0.4 12 17 3.7 5.0 0.2 0.5 30 13 7.5 5.0 0.9 0.5 0 18 4.7 4.0 0.4 0.3 21 18 6.7 7.5 1.0 1.2 20 6 6.2 8.7 0.7 1.3 0 8 6.2 7.1 0.7 0.9 4 9 4.8 2.4 0.5 0.1 7 5 5.7 4.8 0.6 0.4 0 24 3.9 4.5 0.2 0.4 9
Table 3 Feature values of musicians and non-musician significantly differed for three parameters (mean standard deviation) Response speed (S)
Response intensity (I)
Response power (P)
Group Musicians Non-musicians Main effect of group
0-back (V1) 2.42.6
1-back (V2) 0-back (V3) 2.12.9 6.82.0
8.39.6 13.37.5 F (1,30) = 36.39, p < .01
1-back (V4) 6.81.9
5.41.2 5.11.5 F (1,30) = 10.41, p < .01
0-back (V5) 0.90.6
1-back (V6) 0.90.5
0.60.3 0.50.3 F (1,30) = 9.30 , p < .01
Table 4 Mean values of two clusters computed by feature analysis significantly differed for three parameters (mean standard deviation) Response speed (S)
Response intensity (I)
Response power (P)
Cluster# 1 (21 persons) 2 (11 persons) Main effect of cluster
0-back (V1) 2.73.0 10.410.4
1-back (V2) 3.23.9 16.26.8
F (1,30) = 192.57, p < .01
0-back (V3) 6.51.8 5.41.4
1-back (V4) 6.32.0 5.31.5
F (1,30) = 458.28, p < .01
0-back (V5) 0.80.5 0.50.3
1-back (V6) 0.80.5 0.60.4
F (1,30) = 112.98, p < .01
4. Discussion and conclusion This study developed a feature analysis method of EEG to differentiate between musicians and non-musicians on the easier working memory N-back conditions. The results of this study revealed that musicians and non-musicians had similar EEG features upon the initial stimulation of the test, indicating that the participants’ EEG signals also generated features following the stimulations during the working memory tasks. Compared to non-musicians, the musicians had a faster response speed, a higher response intensity, and a higher response power. Before these results can be explained, the possible cognitive functions of beta waves and the features in our analysis must be clarified. Beta waves are characteristic of a strongly engaged mind. Beta frequency responses have been observed in processing difficult tasks [26], motor planning [27], memory retrieval [28], and memory load [28,29]. A person with a high proportion of beta brainwaves may be able to think fast, concentrate, feel excited, and remain in a state of high cognitive functioning [26]. Moreover, a high intensity of beta brainwaves may cause muscular tension and high blood pressure. In this investigation, feature analysis of response speed, response intensity, and response power 10
of beta waves, was conducted to classify musicians and non-musicians. The response speed, which is determined by the time of the first descent of the beta waves after the beginning of the N-back memory tasks, may reflect the speed with which a participant’s state of arousal decreases and the participant adapts to the demands of the task or the associated cognitive loading. The response intensity, which is given by the area under the beta waves, may reveal the accumulation energy by a participant during the task. The response power is the power spectral density of beta waves and refers to the spectral energy distribution of beta waves. The EEG beta wave responses of musicians shifted more rapidly than those of non-musicians and their psychological states changed from tense to relax more rapidly, too. In addition, the musicians herein were more focused and sensitive to tasks than the non-musicians. Possibly, during working memory tasks, musicians used more effective strategies for encoding and manipulating sounds, even with no melody [30]. Therefore, they might have performed the memory tasks more efficiently than non-musicians. However, we could not exclude the possibility that the EEG beta wave responses seen in musicians might be a phenomenon associated with their better motor control and ability to sustain attention, which might result in less effort needed on the task. This should be clarified in future studies. The improved working memory of musicians has been associated with greater brain activities in the right lateral prefrontal cortex, the right lateral parietal cortex, the bilateral posterior dorsal prefrontal cortex, and the bilateral anterior cingulate gyrus [3], as well as a larger P300 amplitude [31]. Furthermore, long-term extensive music training has been found to influence the anatomy of the brain. Professional musicians have been found to have more gray matter in motor-related and auditory-related areas [32]. Music practice has been found to provide cognitive improvements not only in working memory [3,30,31,33-37], but also in reasoning [1], visual attention [38], and processing speed [39]. In the future, the feature analysis model could be used to differentiate the EEG physiological signals of 11
musicians from those of non-musicians in other cognitive tasks. In summary, the working memory EEG analysis can reveal short-term amplitude responses but may not reveal long-term EEG responses and changes. This study analyzed response speed, response intensity, and response power to identify changes in the EEG features in the 30-s interval beginning from the initial stimulation of the test. Moreover, the working memory EEG feature analysis was unitary and typically used to compare the amplitudes of the stimulations in the original EEG signals or to identify similarities in the waveforms. By capturing and analyzing multiple EEG features and eliminating noise by using the EEG proportions, this study can be used as the basis to realize the development of the talent musician discovering system.
Acknowledgements This work is partly supported by the Fu-Jen catholic university and ministry of science and technology of Taiwan under grants MOST 103-2221-E-030-018.
References [1] M. Forgeard, E. Winner, A. Norton and G. Schlaug, "Practicing a musical instrument in childhood is associated with enhanced verbal ability and nonverbal reasoning," PLoS One. 3(10)(2008), e3566. doi:10.1371/journal.pone.0003566. [2] T. Fujioka, L.J. Trainor, B. Ross, R. Kakigi and C. Pantev, "Musical training enhances automatic encoding of melodic contour and interval structure," J Cogn Neurosci. 16(6) (2004) 1010-1021. doi:10.1162/0898929041502706. [3] K.J. Pallesen, E. Brattico, C.J. Bailey, A. Korvenoja, J. Koivisto, A. Gjedde and S. Carlson, "Cognitive control in auditory working memory is enhanced in musicians," PLoS One. 5(6) (2010) e11120. doi:10.1371/journal.pone.0011120. [4] J. Wasifa, D. Saptarshi, M. Koushik, P. Indranil and K. Doga, “Brain connectivity analysis from 12
EEG signals using stable phase-synchronized states during face perception tasks,” Physica A 434 (2015) 273-295. doi: 10.1016/j.physa.2015.03.087. [5] G. Arnulfo, J. Hirvonen, L. Nobili, S. Palva and J.M. Palva, “Phase and amplitude correlations in resting-state activity in human stereotactical EEG recordings,” Neuroimage 112 (2015) 114-127. doi: : 10.1016/j.neuroimage.2015.02.031. [6] M. Evan, G. Hatice and K. Nick Bryan, “Using affective and behavioural sensors to explore aspects of collaborative music making,” Int. J. Human-Computer Studies 82 (2015) 31-47. doi: 10.1016/j.ijhcs.2015.05.002. [7] A. Mathieu, S. Mazza, D. Petit, A. Décary, J. Massicotte-Marquez, J. Malo and J. Montplaisir, “Does age worsen EEG slowing and attention deficits in obstructive sleep apnea syndrome?” Clin Neurophysiol. 118(7) (2007) 1538-1544. doi: 10.1016/j.clinph.2007.04.009 [8] C.C. Kung, T.H. Hsieh, J.Y. Liou, K.J. Lin, F.Z. Shaw and S.F. Liang, “Musicians and non-musicians’ different reliance of features in consonance perception: A behavioral and ERP study,” Clin Neurophysiol. 125(5) (2014) 971-978. doi: 10.1016/j.clinph.2013.10.016. [9] C.C. Hsu, J.H. Wu, H.C. Chiu and C.M. Lin, “Evaluating the sleep quality of obstructive sleep apnea patients after continuous positive airway pressure treatment,” Comput Biol Med. 43(7) (2013) 870-878. doi: 10.1016/j.compbiomed.2013.04.006. [10] B. Guntekin, D.D. Emek-Savas, P. Kurt, G.G. Yener and E. Basar, “Beta oscillatory responses in healthy subjects and subjects with mild cognitive impairment,” Neuroimage Clin. 3 (2013) 39-46. doi: 10.1016/j.nicl.2013.07.003. [11] A. Wróbel, “Beta activity: a carrier for visual attention,” Acta Neurobiol Exp (Wars). 60(2) (2000) 247-260. [12] L.F. Basile, J.R. Sato, M.Y. Alvarenga, N.Jr. Henrique, H.A. Pasquini, W. Alfenas, S. Machado, B. Velasques, P. Ribeiro, R. Piedade, R. Anghinah and R.T. Ramos, “Lack of systematic topographic difference between attention and reasoning beta correlates,” PLoS One. 8(3) (2013) e59595. doi: 10.1371/journal.pone.0059595. 13
[13] F. van Ede, F. de Lange, O. Jensen and E. Maris, “Orienting attention to an upcoming tactile event involves a spatially and temporally specific modulation of sensorimotor alpha- and beta-band
oscillations,”
J
Neurosci.
31(6)
(2011)
2016-2024.
doi:
10.1523/JNEUROSCI.5630-10.2011. [14] T.S. Hale, S.L. Smalley, P.D. Walshaw, G. Hanada, J. Macion, J.T. McCracken, J.J. McGough and S.K. Loo, “Atypical EEG beta asymmetry in adults with ADHD,” Neuropsychologia. 48(12) (2010) 3532-3539. doi: 10.1016/j.neuropsychologia.2010.08.002. [15] N. Zaretskaya and A. Bartels, “Gestalt perception is associated with reduced parietal beta oscillations,” Neuroimage. 112 (2015) 61-69. doi:10.1016/j.neuroimage.2015.02.049. [16] J. Perrier, T.R. Leufkens, J. Bulla, S. Jongen, M.L. Bocca, J.G. Ramaekers and A. Vermeeren, “Electroencephalography during on-the-road driving in older untreated insomnia patients and normal sleepers,” Biol Psychol. 109 (2015) 20-28. doi: 10.1016/j.biopsycho.2015.04.002. [17] A. Gevins, L.K. McEvoy, M.E Smith, C.S. Chan, L. Sam-Vargas, C. Baum and A.B. Ilan, “Long-term and within-day variability of working memory performance and EEG in individuals,” Clin Neurophysiol. 123(7) (2012) 1291-1299. doi: 10.1016/j.clinph.2011.11.004. [18] P. Ruhnau, N. Wetzel, A. Widmann and E. Schröger, “The modulation of auditory novelty processing by working memory load in school age children and adults: a combined behavioral and
event-related
potential
study,”
BMC
Neurosci.
11
(2010)
126.
doi:
10.1186/1471-2202-11-126. [19] P.J. Maes, M.M. Wanderley and C. Palmer, “The role of working memory in the temporal control of discrete and continuous movements,” Exp Brain Res. 233(1) (2015) 263-73. doi: 10.1007/s00221-014-4108-5. [20] L. Han, Y. Liu, D. Zhang, Y. Jin and Y. Luo, “Low-arousal speech noise improves performance in
n-back
task:
an
ERP
study,”
PLoS
One.
8(10)
(2013)
e76261.
doi:
10.1371/journal.pone.0076261. [21] F. Rochette, A. Moussard and E. Bigand, “Music lessons improve auditory perceptual and 14
cognitive performance in deaf children,” Front Hum Neurosci. 8(2014) 488. doi: 10.3389/fnhum.2014.00488. [22] D.L. Strait, J. Hornickel and N. Kraus, “Subcortical processing of speech regularities underlies reading and music aptitude in children,” Behav Brain Funct. 7(2011) 44. doi: 10.1186/1744-9081-7-44. [23] N.E. Huang and S.S. P Shen, Hilbert Huang transform and its applications: 2nd Edition, World Scientific, Singapore, 2014. [24] J. Leis, Digital signal processing using MATLAB for students and researchers, Wiley, New York, 2011. [25] R Core Team, R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria, 2013. http://www.R-project.org/. [26] J.B. Brookings, G.F. Wilson and C.R. Swain, “Psychophysiological responses to changes in workload during simulated air traffic control,” Biol Psychol. 42(3) (1996) 361-377. doi: 10.1016/0301-0511(95)05167-8. [27] J. Kaiser, N. Birbaumer and W. Lutzenberger, “Event-related beta desynchronization indicates timing of response selection in a delayed-response paradigm in humans,” Neurosci Lett. 312(3) (2001) 149-152. doi:10.1016/S0304-3940(01)02217-0. [28] M. Karrasch, M. Laine, P. Rapinoja and C.M. Krause, “Effects of normal aging on event-related desynchronization/synchronization during a memory task in humans,” Neurosci Lett. 366(1) (2004) 18-23. doi:10.1016/j.neulet.2004.05.010. [29] M. Pesonen, C.H. Björnberg, H. Hämäläinen and C.M. Krause, “Brain oscillatory 1-30 Hz EEG ERD/ERS responses during the different stages of an auditory memory search task,” Neurosci Lett. 399(1-2) (2006) 45-50. doi:10.1016/j.neulet.2006.01.053. [30] K. Schulze, K. Mueller and S. Koelsch, “Neural correlates of strategy use during auditory working memory in musicians and non-musicians,” Eur J Neurosci. 33(1)(2011) 189-196. doi:10.1111/j.1460-9568.2010.07470.x. 15
[31] E.M. George and D. Coch, “Music training and working memory: An ERP study,” Neuropsychologia 49(5) (2011) 1083-1094. doi:10.1016/j.neuropsychologia.2011.02.001. [32] K.L. Hyde, J. Lerch, A. Norton, M. Forgeard, E. Winner, A.C. Evans and G. Schlaug, “Musical training shapes structural brain development,” J Neurosci. 29(10) (2009) 3019-3025. doi:10.1523/jneurosci.5118-08.2009. [33] S.L. Bengtsson, Z. Nagy, S. Skare, L. Forsman, H. Forssberg and F. Ullen, “Extensive piano practicing has regionally specific effects on white matter development,” Nat Neurosci. 8(9) (2005) 1148-1150. doi: 10.1038/nn1516. [34] S. Bergman Nutley, F. Darki and T. Klingberg, “Music practice is associated with development of working memory during childhood and adolescence,” Front Hum Neurosci. 7(926) (2014) eCollection. doi:10.3389/fnhum.2013.00926. [35] J.A. Bugos, W.M. Perlstein, C.S. McCrae, T.S. Brophy and P.H. Bedenbaugh, “Individualized piano instruction enhances executive functioning and working memory in older adults,” Aging Ment Health. 11(4) (2007) 464-471. doi:10.1080/13607860601086504. [36] L. Suárez, S. Elangovan and A. Au, “Cross‐ sectional study on the relationship between music training and working memory in adults,” Aust J Psychol. 68(1) (2016) 38-46. doi: 10.1111/ajpy.12087. [37] L. Vandervert, “How music training enhances working memory: A cerebrocerebellar blending mechanism that can lead equally to scientific discovery and therapeutic efficacy in neurological disorders,” Cerebellum Ataxias. 2(1) (2015) 11. doi:10.1186/s40673-015-0030-2. [38] A.C. Rodrigues, M.A. Loureiro and P. Caramelli, “Long-term musical training may improve different forms of visual attention ability,” Brain Cogn. 82(3) (2013) 229-235. doi: 10.1016/j.bandc.2013.04.009. [39] E.G. Schellenberg and W.W. Michael, “Music and cognitive abilities,” In D. Deutsch (Ed.) The Psychology of Music: 3rd, vol. 12, pp. 499-550, Amsterdam, Elsevier, 2013.
16