Evaluating Virtual Reality Mood Induction Procedures with Portable EEG Devices

E alua ing Vi ual Reali y Mood Induc ion P ocedu es wi h Po able EEG De ices Alejand o RODRÍGUEZa,1 Bea iz REY a,b and Ma iano ALCAÑIZ a,b a Ins i u o In e uni e si a io de In es igación en Bioingenie ía y Tecnología O ien ada al Se Humano, Uni e si a Poli ècnica de València, I3BH/LabHuman, Camino de Ve a s/n, 46022 Valencia, España Ins i u o In e uni e si a io. b Cibe , Fisiopa ología de Obesidad y Nu ición, CB06/03 Ins i u o de Salud Ca los III, Spain. Abs ac . Vi ual En i onmen s (VEs) ha e been used as mood induc ion p ocedu es. In his con ex , i is necessa y o ha e ins umen s o analyze he emo ional s a e du ing VE exposu e. Objec i e echniques such as EEG should be e alua ed o his pu pose. The aim in his wo k was o s udy he changes in he b ain ac i i y wi h a po able EEG de ice du ing a nega i e mood induc ion based on a VE. A i ual pa k was used o induce a nega i e mood (sadness) in en pa icipan s. Changes in he b ain ac i i y o subjec s we e compa ed be ween wo momen s (be o e and a e emo ional induc ion). Ob ained esul s we e in acco dance wi h p e ious scien i ic li e a u e ega ding on al EEG asymme y, which suppo s he possibili y o using he po able EEG as a eliable ins umen o measu e emo ions in VE. Keywo ds. EEG, Emo i Epoc, Nega i e Mood Induc ion, Vi ual En i onmen In oduc ion Vi ual En i onmen s (VEs) ha e been used as mood induc ion p ocedu es (MIPs) in p e ious emo ional s udies. The e icacy o Vi ual Reali y (VR) as a MIP (VR-MIP) has been es ed [1], whe e he esul s showed ha VR was an excellen MIP because subjec s el mo e p esence han in adi ional MIPs du ing he mood induc ion expe ience. As VEs can be used o mood induc ion, i is necessa y o ha e ools ha allow us o e alua e and assess mood changes o subjec s du ing he VR expe ience. Cu en ly, assessmen ins umen s a e based on subjec i e ques ionnai es ha ask he subjec s abou he moods hey el . These ques ionnai es, al hough hey ha e p o ed o be e y use ul, p esen some limi a ions. Fo example, i he ques ionnai es a e used in combina ion wi h a VE, hey may only be used be o e and a e he expe ience, bu can ne e be used du ing he i ual exposu e wi hou in e up ing i . The e o e, i is undamen al o ha e ins umen s o analyze emo ional s a es while moods a e being induced in subjec s h ough VE [2]. B ain ac i i y measu es based on elec oencephalog am (EEG) can be a good ins umen . EEG allows us o measu e he 1 Co esponding Au ho . Annual Re iew o Cybe he apy and Telemedicine 2013 B.K. Wiede hold and G. Ri a (Eds.) © 2013 In e ac i e Media Ins i u e 131 luc ua ions in b ain elec ical ac i i y associa ed wi h emo ions in a non-in asi e way; wi h a empo al esolu ion o milliseconds EEG is a echnique ha would be easily combinable wi h VR-MIPs since i would no impose es ic ions wi h he i ual s imuli ha could be p esen ed in a s udy. Mo eo e , new po able EEGs de ices ha a e appea ing on he ma ke can be easily combined wi h VEs. The ea u es o EEG measu es allow us o o e come ques ionnai e limi a ions, hus complemen ing he in o ma ion ob ained by means o hese ins umen s. In ou p esen wo k, a VE was used o induce nega i e emo ions (sadness) in pa icipan s while a po able EEG moni o ed hei b ain ac i i ies. The main goal was o analyze i a po able EEG de ice was app op ia e o e alua e he on al asymme y o he EEG [3] in he con ex o VE s udies. 1. Ma e ials and Me hods Ten heal hy pa icipan s, 5 men and 5 women, we e e alua ed in his s udy, all o hem igh -handed, wi hin he age o 23-27 yea s old, and wi h no mal o co ec ed- o- no mal ision. The subjec s we e uni e si y s uden s wi hou any psychological diso de ; none o which we e expe s in he use o he echnologies used in his s udy. The pa icipan s signed an in o med consen allowing hei da a o be used in his s udy. The EEG signal was moni o ed by means o an EEG po able de ice, Emo i EPOC [4], which had 14 EEG channels using he s anda d 10/20 layou and 2 e e ence channels (CMS/DRL), Fig.1. Figu e 1. Emo i EPOC de ice and Emo i EPOC elec odes posi ioning on s anda d 10/20 layou . EEG eco dings we e analyzed o -line using cus om so wa e w i en in MATLAB. This cus om so wa e was based on EEGLAB [5] ools, which was applied o clean he EEG-da a and analyze changes in alpha-spec al powe . In he expe imen al session, he pa icipan s had o na iga e h ough a VE designed o induce hem in o a nega i e mood (sadness) [2]. This VE consis ed o a i ual pa k, al eady es ed in p e ious s udies, and can be used as an e ec i e mood-induc ion p ocedu e [1]. This i ual pa k was shown in a e o-p ojec ed sc een and pa icipan s we e able o na iga e using a wi eless pad (Logi ech Rumblepad). Some pic u es o he VE and he con igu a ion o he expe imen al session a e shown in Fig. 2. The p o ocol o his s udy can be consul ed in [2]. Pa icipan s spen app oxima ely one hou in he labo a o y and a PANAS ques ionnai e was comple ed by each; be o e A. Rod íguez e al. / E alua ing Vi ual Reali y MIPs wi h Po able EEG De ices132 and a e he i ual expe imen . An analysis o hese ques ionnai es allowed us o di ide he subjec s in o wo g oups; subjec s in whom he mood was success ully induced (sad g oup) and subjec s who we e no sad a e he exposu e (non-sad g oup). Figu e 2. Subjec na iga ing in he VE and cap u e o he VE used o induce sadness The na iga ion pe iod was p eceded and ollowed by es pe iods du ing which pa icipan s had o wa ch a black sc een du ing wo minu es and hi y seconds. Be ween he wo es pe iods, he i ual mood-induc ion p ocedu e s a ed, using he same p o ocol as p e iously desc ibed [1-2]. Finally, he pa icipan s we e in i ed o isualize a ilm o induce hem in o a posi i e mood be o e inishing he expe imen . Na u al log Alpha powe and asymme ic coe icien alue o F3 and F4 we e calcula ed. The asymme ic coe icien alues we e calcula ed h ough na u al log o he a io F4/F3 [3]. F3 and F4 we e used because hey a e he senso s mo e o en used in scien i ic li e a u e o analyze he alence o he mood expe ienced by pa icipan s. A h ee way ANOVA wi h epea ed measu es in wo o he ac o s was applied o Na u al log Alpha alues. The ac o s we e he momen (ini ial es and inal es ), he hemisphe e (le / igh ) and he g oup (sad/non-sad). A wo way ANOVA wi h epea ed measu e in one o he ac o s was applied o he asymme ic coe icien alues. The ac o s we e he momen (ini ial es and inal es ) and he g oup (sad/non-sad). 2. Resul s In his sec ion, we p esen he esul s o ou cu en s udy. The PANAS ques ionnai e showed ha a sad mood was induced in six pa icipan s (sad g oup) while in ou subjec s i was no possible o induce a sad emo ion wi h his VE (non-sad g oup). Resul s showed signi ican di e ences be ween igh and le hemisphe es (F(1,8)=12.058; p=0.008), a end close o signi icance o he in e ac ion ac o be ween hemisphe e, momen and g oup (F(1,8)=5.267; p=0.051) in he na u al log alpha powe . Pai -wise compa isons only showed signi ican di e ences be ween hemisphe es in he sad g oup a e he i ual induc ion (p=0.014). The mean alue and s anda d de ia ion o each senso o di e en momen s and g oups can be obse ed in Table 1. A. Rod íguez e al. / E alua ing Vi ual Reali y MIPs wi h Po able EEG De ices 133 Table 1. Mean and s anda d de ia ion o F3 and F4 senso s. G oup F3 senso F4 senso be o e a e be o e a e Mean S d Mean S d Mean S d Mean S d Sad 3.368 1.244 1.700 1.116 2.045 1.006 3.765 1.372 Non-Sad 1.723 1.257 1.968 0.901 1.646 0.901 2.417 1.127 On he o he hand, he asymme ic coe icien showed a end o he in e ac ion ac o be ween momen and g oup (F(1,8)=5.267; p=0.051). Pai -wise compa isons showed signi ican di e ences be ween he momen s be o e and a e he i ual induc ion (p=0.014) o he sad g oup (Table 2). Table 2. Mean and s anda d de ia ion asymme ic coe icien EEG F on al. G oup Asymme ic Coe . be o e a e Mean S d Mean S d Sad 2.065 1.273 1.137 7.532 Non-Sad 4.485 8.218 5.945 1.078 Figu e 3 shows six opog aphic scalp maps, which gi e a isual ep esen a ion abou he dis ibu ion o a e ages na u al log alpha alues a ound scalp in be o e induc ion, a e induc ion and he di e ence be ween hem. Figu e 3. Topog aphic scalp map o a e age o na u al log alpha alues 3. Discussion and Conclusion The goal o his s udy was o de e mine i po able EEG de ices would be able o e alua e asymme ic EEG on al heo ies in an emo ional i ual s udy. S a is ical esul s showed g ea e ac i a ion, dec ease o alpha powe [6], in he igh hemisphe e a e i ual mood induc ion o he sad g oup. Fu he mo e, he asymme y analysis showed signi ican di e ences be ween he momen s be o e and A. Rod íguez e al. / E alua ing Vi ual Reali y MIPs wi h Po able EEG De ices134 a e he i ual induc ion o he sad g oup. Finally, hese esul s we e co obo a ed wi h a isual analysis in igu e 3, which shows ha he ac i i y in he igh hemisphe e is g ea e han in he le hemisphe e o he g oup in which he sad emo ion was induced o he subjec s. These esul s showed e idence o he asymme y in he on al EEG o he pa icipan s a e a sad induc ion, which is cohe en wi h p e ious esul s in scien i ic li e a u e which indica es ha he igh hemisphe e is ac i a ed when a subjec has been induced a nega i e mood [3, 6]. This s udy demons a es ha po able EEG de ices can be used as a complemen a y ool in emo ional induc ion s udies wi h i ual eali y, because hei lexibili y and eliabili y allow ob aining esul s abou he b ain ac i a ion o he pa icipan s caused by he mood induc ion wi h VR-MIPs in a mo e objec i e way. Acknowledgmen s. This s udy was unded by Vice ec o ado de In es igación de la Uni e si a Poli ècnica de València, Spain, PAID-06-2011, R.N. 1984; by Minis e io de Educación y Ciencia, Spain, P ojec Game Teen (TIN2010-20187) and pa ially by p ojec s Consolide -C (SEJ2006-14301/PSIC), “CIBER o Physiopa hology o Obesi y and Nu i ion, an ini ia i e o ISCIII” and Excellence Resea ch P og am PROMETEO (Gene ali a Valenciana. Conselle ía de Educación, 2008-157). The wo k o Alejand o Rod íguez was suppo ed by he Spanish MEC unde an FPI G an BES-2011-043316. Re e ences [1] Baños, R., Liaño, V., Bo ella, C., Alcañiz, M., Gue e o, B., & Rey, B. (2006). Changing induced moods ia i ual eali y. In: W.A. Ijssels eijn, Y. de Ko , C. Midden, B. Eggen, E. an den Ho en, eds, Pe suasi e Technology: Lec u e No es in Compu e Science, Be lin/Heilde be g: Sp inge -Ve lag (2006), 7-15. [2] Rey,B., Rod íguez,A.& Alcañiz,M. (2012). Using po able EEG de ices o e alua e emo ional egula ion s a egies du ing i ual eali y exposu e. S udies in heal h echnology and in o ma ics,181, 223-227. [3] Richa d S. Lewis, Nicole Y. Weekes &T acy H. Wang(2007). The e ec o a na u alis ic s esso on on al eeg asymme y, s ess, and heal h. Biological Psychology, 75(3): 239 - 247. ISSN 0301-0511. [4] Zhang, B., Wang, J., Fuhlb igge, T. (2010). A Re iew o he Comme cial B ain-Compu e In e ace Technology om Pe spec i e o Indus ial Robo ics. P oceedings o he 2010 IEEE In e na ional Con e ence on Au oma ion and Logis ics. [5] Delo me A & Makeig S (2004) EEGLAB: an open sou ce oolbox o analysis o single- ial EEG dynamics. Jou nal o Neu oscience Me hods 134:9-21 [6] Coan, J. A., & Allen, J. J. (2004). F on al EEG asymme y as a mode a o and media o o emo ion. Biological psychology, 67(1), 7-50. A. Rod íguez e al. / E alua ing Vi ual Reali y MIPs wi h Po able EEG De ices 135