ॣڈᐟΠϐတݢૻဦϩ
Analysis of the EEG Signals Response to Musical Signal Stimuli
݅࠶דǵߋݥЎǵ৪ࡌ *
ѠчᙴᏢεᏢ ᙴᏢၗૻࣴز܌
Wei-Chih Lin , Hung-wen Chiu , Chien-Yeh Hsu *
Graduate Institute of Medical Informatics, Taipei Medical University
* Corresponding Author. Email: cyhsu@tmu.edu.tw
ǵʳύЎᄔा ߈ԃٰӭ࣬ᜢޑࣴزჴΑॣޑਏ݀ё аϸࢀӧΓᡏޑ٤ғኧᏵ[1]Ƕՠӧ EEG(တݢ) Бय़ǴᗲϿڀԖ၁ᅰుΕޑѐӧॣڈᐟΠ ΓᜪတࢲޑቹៜǶࡺҁࣴزख़ӧಡ᠋ॣ ਔ EEG( တ ݢ ) ϐ ϩ Б ݤ Ǵ ၮ Ҕ ᓎ ϩ (frequency distribution analysis)аϷጕ܄ᐱҥϡҹ ϩݤǴԑӧϩတݢӧಡ᠋όӕॣ܌ౢғޑ ᡂϯǴ٠යఈૈᏵаϩᜪǵᡉҢॣ፦ᆶғ ૻဦ፦໔ޑ࣬ᜢ܄Ƕ
Զҁࣴزวӧಡ᠋ॣ(Soft)аϷڙ ෳޣҁي܌഻ངޑॣ(Favorite)ਔǴӧ Alpha
band аϷ Theta band ϩǴวڙෳޣഗယ
(Parietal)P3ǵP4ǵPz Οᗺϐ EEG Power ڀԖܴᡉ ϲǴԶӧའᄾॣ(Rock)аϷؒԖ᠋ॣޑ ნΠ(baseline)Ǵഗယ EEG Power ࠅևΠफ़ޑ ݩǶ೭٤วᡉҢрॣޑৡ౦ዴჴૈϸᔈӧ Γᡏ EEG Ƕ
ᜢᗖຒǺ တݢǵॣǵᓎϩǵጕ܄ᐱҥϡҹ ϩݤǶ
Abstract
In recent years, many studies have shown that music can reflect quantifiable physiological effects on human. However, in the part of EEG, not many research projects have made use of EEG to verify the music influence on human brain activity. This research has emphasized on the development of analytical tools and methods for bio-signals, especially focused on the part of EEG with musical signal stimuli. We can use frequency distribution analysis and linear separation algorithms such as the Independent Component Analysis (ICA) to analyze the data collected from the scalp electrodes. The ultimate goal is to analyze EEG responses of subjects with different musical signal stimuli. It is expected that different musical signal stimuli can be classified and therefore the correlation between music and bio-signal characteristics can be demonstrated.
In this study, we have found out that when subjects were listening to Soft and subject-preferred
music, the EEG power at parietal points-P3ǵP4ǵPz located on alpha band and theta band-were both on the rise. When subjects were listening to rock music and at baseline, however, the EEG power at parietal was decreasing. These findings have confirmed that the differences in music stimulation have reflected physiological effects on human EEG.
Keywords: Electroencephalogram, Music Stimuli, Frequency Distribution Analysis, Independent Component Analysis ΒǵጔҗᆶҞޑ ԾђаٰॣӧΓᜪޑғࢲύޔ՞Ԗख़ ाޑӦՏǴԶ߈ԃٰӭ࣬ᜢޑࣴزΨჴΑॣ ޑ ਏ ݀ ё а ϸ ࢀ ӧ Γ ᡏ ޑ ٤ ғ ኧ Ᏽ [1-3]ǶԶԿҞࣁЗǴᗲϿԖࣴزుΕޑϩတ ݢǴٰӑॣჹΓᡏတࢲޑቹៜǶڂࠠϩ ғ ૻ ဦ த ٬ Ҕ ਔ ᓎ ϩ (Time-frequency analysis)բࣁϩ EEG کॣૻဦޑБݤ[4]Ǵஒ
EEGբഡճယᙯඤϩȐFourier transformȑǴஒ
ፄᚇޑݢࠠᙯϯԋᓎୱǴ٠٩ᓎޑόӕϩձᄒ ڗр Alpha bandǵBeta bandǵGamma bandǵTheta bandǴᙖԜ࣮рόӕ band ޑமࡋ Ǵளډတݢӧ ᓎޑϩթ٠җԜϩǶҁࣴزი໗ ς߃ࡌҥଆಡ᠋ॣਔғߞဦϐໆෳБݤ [5]Ƕ ନᓎϩѦǴҗܭໆෳ EEG ૻဦਔǴሡճ ҔܫᓐҜޑႝཱུ(electrodes)ѐԏတ㚊ઓ ϡޑࢲǴՠࢂӢࣁεတǵλတຯᓐҜԖλࢤ ຯᚆǴ೭ύ໔кᅈΑ Volumn- conductanceǴ܌а ᓐҜޑҺՖঁႝཱུԏډӧঁ࣬ε ୱϣတઓϡޑࢲǶЪӢࣁ೭ᅿᏤ٠όੋ Ϸਔ໔ᒨۯ(time-delay)Ǵ܌аጕ܄ϩᚆݤǴٯӵ ᐱҥϡҹϩ(Independent Component Analysis, ICA)൩ࡐӝҔٰϩᓐҜԏډޑတႝݢǴ όሡा٣ଷ٣ౢғႝՏޑวғᏔԖ൳ঁǴ ൩ёவ୴షӝޑૻဦύϩрൂᐱޑӚձޑૻ ဦٰྍ[6,7]Ƕ ICAҔٰϩတݢૻဦਔǴӧόӕႝཱུԏ ޑEEGૻဦࢂᒡΕޑંତ̔ǴԶICAޑᒡр̑ж ߄όӕޑᐱҥϡન(Independent Component)Ƕ̑ޑ
ӈ (Row) ж߄όӕޑઓᆛၡ܈ࢂတѦ (Extra-brain)ૻဦྍȐٯӵઁ࿊Ǵᖍय़ԼԺםȑ ܌ౢғޑݢࠠǶӢࣁICAࢂঁΜϩᙁൂޑጕ܄ سǴ̑=̓̔Ǵ܌аךॺΨёаஒঁঁޑᐱҥ ϡનǴൂᐱӦቹډᓐҜޑႝཱུǴΨ൩ࢂ̔= ̓-1̑[8,9]Ƕ ࡺҁࣴزख़ӧಡ᠋ॣਔ EEG(တݢ)ϐϩ БݤǴаᓎϩǵጕ܄ᐱҥݤٰϩတݢӧ όӕॣڈᐟΠ܌ౢғޑᡂϯǴ٠යఈૈᏵԜ ፦уаϩᜪǴᡉҢॣ፦ᆶғૻဦ፦໔ޑ ࣬ᜢ܄Ƕ ΟǵБݤᆶ ҁࣴزӼ௨ڙෳޣಡ᠋ॣǴໆෳڙෳޣӧ ௗԏॣڈᐟރᄊΠǴԏڗऊ 15 ϩដޑတݢၗ Ƕॣڈᐟޑᒧϩձࣁॣ(Soft)ǵའᄾ ॣ(Rock)аϷڙෳޣԾᒧॣ(Favorite)Ǵу ό᠋ॣ(Baseline)ᕴӅѤᅿნǶჴᡍჹຝϐᒧ ڗаѠчᙴᏢεᏢᏢғࣁЬЪค᠋ϐምᛖǵค ҺՖςޕቹៜတݢޑ੯ੰǵ҂ۓයܺҔᛰނЪ ᜫཀଛӝୖᆶҁჴᡍޣǶჴᡍՉޑӦᗺࣁೀ ӼᓉόڙѦࣚυᘋޑࣴز࠻Ƕတݢᐒҁࣴز௦Ҕ
Stellate HarmonieǴ٬Ҕ 10-20 systems ୯ሞྗݤ
(კ)ᘏڗ 21 ঁ Channel တݢૻဦǶڙ၂ޣٰ ډࣴز࠻ࡕǴҁࣴزޣࣁڙ၂ޣௗတݢᐒǴ٠ ࣁڙ၂ޣᔎԸᐒǴڙ၂ޣ֤ӧԖङёԾฅܫ ޑශηௗڙჴᡍნǶࣁΑᗉխჴᡍნኞ ܫׇ܌ԋޑᇤৡǴࡺჴᡍნ௦ҔᒿᐒޑБ ԄևǶ
კǺThe 10-20 System of Electrode Placement ( рԾhttp://faculty.washington.edu/chudler/10 20.html [10])
ӧϩတݢၗБय़Ǵஒᘏڗډޑတݢૻဦ ϩࣁ RockǵSoftǵBaseline аϷڙෳޣԾᒧॣ ѤঁნǴϩځਔᓎૻဦϐ܄Ǵໆෳрတݢ ૻဦϐ Alpha bandǵBeta bandǵGamma bandǵTheta
bandӚᓎޑૈໆኧॶǴӕਔࣁΑ࠼ᢀޑᢀჸӚ ڙෳޣૈໆޑᡂϯǴࡺஒӚૈໆኧॶჹᕴૈ ໆ࣬ନၲډ҅ೕϯҞޑǶ٩Ԝኧॶբࣁတݢૻ ဦޑ፦Ǵӕਔ٠ஒတݢૻဦၸ ICA բೀǴ ᙖԜᢀჸ܌ϩဂ໔ᆶ܌᠋ॣޑ࣬ᜢ܄ǶԶӧ ीϩБय़ǴҗܭԖਏၗࣁΐಔǴࡺҁࣴز ճҔค҆ኧϩȐNonparametric testȑWilconxon
Singned Ranks TestٰᔠᡍӧόӕჴᡍნΠޑတ
ݢᓎࢂցԖ܌όӕǶ Ѥǵ่݀ᆶፕ २Ӄךॺᢀჸ EEG ӧӚᅿნΠόӕᓎ ޑᡂϯݩǴკΒᡉҢတ 21 channels ӧѤᅿჴ ᡍნΠ Alpha band ਔޑૈໆॶǶёа࣮ـځύ 3ǵ4ǵ7ǵ11ǵ14ǵ15ǵ18 channel Ԗၨεޑૈໆ КॶǴϩձж߄တ P4ǵO2ǵT6ǵPzǵP3ǵ O1ǵT5 Ӛᗺޑ Alpha ݢၨࣁமਗ਼Ƕ ˃ˁ˃˃˃ ˈ˃˃˃ˁ˃˃˃ ˄˃˃˃˃ˁ˃˃˃ ˄ˈ˃˃˃ˁ˃˃˃ ˅˃˃˃˃ˁ˃˃˃ ˅ˈ˃˃˃ˁ˃˃˃ ˄ ˅ ˆ ˇ ˈ ˉ ˊ ˋ ˌ ˄˃ ˄˄ ˄˅ ˄ˆ ˄ˇ ˄ˈ ˄ˉ ˄ˊ ˄ˋ ˄ˌ ˅˃ ˅˄ Channel Power ˥̂˶˾ ˦̂˹̇ ˕˟ ˙˴̉
კΒǺAlpha band ύӚ channel ޑ EEG PowerǴკ ΠБ 3ǵ4ǵ7ǵ11ǵ14ǵ15ǵ18 ϩձж߄တ P4ǵO2ǵT6ǵPzǵP3ǵO1ǵT5 ӚᗺǶ ԶӧკΟǵკѤύǴ߾ࢂϩձж߄ӄڙෳ ޣӧ P3 аϷ P4 channel ύѤঁნΠ܌ౢғૈໆ Кѳ֡ॶǶҗॊკҁࣴزวӧಡ᠋ॣ ϷڙෳޣԾᒧॣਔǴӧ Alpha band ਔǴځڙ ෳޣޑ EEG Power ܴᡉၨଯǶЀځӧܾယ
(Occipital)ǵ ഗ ύ ѧ (Midline Parietal) ǵ ഗ ယ
(Parietal)ǵࡕ㙠(Posterior Temporal)ϩǶࡺ׳ ஒৡ౦ၨεޑഗယՏڗрϩǴว P3 ᗺ ӧ Alpha band ک Theta band ޑރᄊǴಡ᠋ ॣ а Ϸ ڙ ෳ ޣ ҁ ي ܌ ഻ ང ޑ ॣ ਔ Ǵ EEG PowerڀԖܴᡉϲ(߄)ǴԶӧའᄾॣаϷؒ Ԗ᠋ॣޑნΠǴഗယ EEG Power ࠅևΠ फ़ޑݩ(კΟǵკѤ)Ƕӕኬޑӧ P4ǵPz ٿᗺΨ วӕኬޑǶ೭٤วᡉҢрॣޑৡ౦ዴ ჴૈϸᔈӧ EEG Ƕՠࢂीϩрޑܴᡉࡋό ଯǴёૈࢂࣴزኬҁኧόى܌ԋޑቹៜǶ ߄ǺP3 ᗺӧѤᅿნΠϐ Power Mean ॶǶ Alpha Theta
Rock Mean=0.12²0.05 Mean=0.04²0.01
Soft Mean=0.23²0.27 Mean=0.06²0.07
Baseline Mean=0.12²0.05 Mean=0.04²0.02
P3 Average ˃ ˃ˁ˃˅ ˃ˁ˃ˇ ˃ˁ˃ˉ ˃ˁ˃ˋ ˃ˁ˄ ˃ˁ˄˅ ˃ˁ˄ˇ ˃ˁ˄ˉ ˃ˁ˄ˋ ˃ˁ˅
Rock Soft BL Fav
Power Ratio Alpha Beta Gamma Theta კΟǺѤᅿნӧᓐ P3 ᗺ܌ౢғ EEG Power ϐᡂϯǶ P4 average ˃ ˃ˁ˃˅ ˃ˁ˃ˇ ˃ˁ˃ˉ ˃ˁ˃ˋ ˃ˁ˄ ˃ˁ˄˅ ˃ˁ˄ˇ ˃ˁ˄ˉ ˃ˁ˄ˋ ˃ˁ˅
Rock Soft BL Fav
Power Rat io Alpha Beta Gamma Theta კѤǺѤᅿნӧᓐ P4 ᗺ܌ౢғ EEG Power ϐᡂϯǶ ԶҁࣴزΨஒတݢૻဦբ ICA(კϖ)Ǵ٠ஒϩ ࡕ܌ᕇளӚᐱҥϡҹ Power மࡋᛤᇙԋϩթკ (კϤ)ǴԜკࣁ Rock ნനࡕϖࣾаϷ Sock ნനϖࣾޑ EEGǴ ICA ࡕ܌ளډ 21 ঁတݢ ᐱҥϡҹǴҗკύךॺёჸӧಃ 10ǵ11ǵ18ǵ 19ǵ20channelǴΨ൩ࢂတޑഗယ(Parietal)ǵ ࡕ㙠(Posterior Temporal)ǵܾယ(Occipital)ޑ ϩǴёวа 180 ࣾࣁΠϩࣚޑ RockǵSoft ნǴၸ ICA ೀၸࡕޑૻဦԖΑܴᡉόӕǶ კϖǺEEG ICA ࡕ܌ள 21 ঁᐱҥϡҹǴ175~180 ࣁ Rock ნǴ180~185 ࣁ Soft ნǶ კϤǺICA ࡕӚᐱҥϡҹϩթமࡋკ(आՅϩࣁ PowerၨமǴᙔՅϩࣁ Power ၨ১) ϤǵୖԵЎ
[1] Koelsch S, Mulder J (2002) Electric brain responses to inappropriate harmonies during listening to expressive music. Clin Neurophysiol, 113: 862-869.
[2] Bhattacharya J, Petsche H (2001) Universality in the brain while listening to music, Proc R Soc
Lond B Biol Sci, 268: 2423-33.
[3] Bhattacharya J, Petsche H, Pereda E (2001) Interdependencies in the spontaneous EEG while listening to music. Int J Psychophysiol, 42: 287-301.
[4] Cohen, L. (1998) Time-Frequency Distribution- A review. Proc IEEE, 77: 941-981
[5]ߋӼ▲ (2004) ॣჹတݢϷЈᡂ౦܄ޑቹ
ៜǶѠчᙴᏢεᏢᙴᏢࣴز܌ǴᅺγፕЎǶ [6] Jung TP, Makeig S, Bell AJ, and Sejnowski TJ
(1998) Independent Component Analysis of Electroencephalographic and Event-related Potential Data. in Central Auditory Processing
and Neural Modeling (eds P Poon and J Brugge),
Plenum Press, NY, pp189-197.
[7] Jung TP, Humphries C, Lee TW, Makeig S, McKeown M, Iragui V, and Sejnowski TJ (1998), Extended ICA removes artifacts from electroencephalographic recordings. Advances
in Neural Information Processing Systems, 10:
894-900.
[8] Makeig S, Bell AJ, Jung TP, and Sejnowski TJ (1996), Independent component analysis of electroencephalographic data. Advances in
Neural Information Processing Systems, 8:
145-151.
[9] Makeig S, Westerfield M, Jung T P, Covington J, Townsend J, Sejnowski T J and Courchesne E (1999) Independent components of the late positive event-related potential in a visual spatial attention task. J Neuroscience, 19: 65-80.
