Virtual Avatar for Emotion Recognition in Patients with Schizophrenia: A Pilot Study

CLINICAL TRIAL published: 26 Augus 2016 doi: 10.3389/ nhum.2016.00421 Vi ual A a a o Emo ion Recogni ion in Pa ien s wi h Schizoph enia: A Pilo S udy Samuel Ma cos-Pablos1*, Emilio González-Pablos2,Ca los Ma ín-Lo enzo2, Luis A. Flo es1,Jaime Gómez-Ga cía-Be mejo3and Edua do Zalama3 1Ca i Founda ion, Pa que Tecnológico de Boecillo, Valladolid, Spain, 2Resea ch Uni , He manas Hospi ala ias Cen o Sociosani a io Palencia, Palencia, Spain, 3ITAP-DISA, Uni e si y o Valladolid, Valladolid, Spain Edi ed by: Mikhail Lebede , Duke Uni e si y, USA Re iewed by: Pablo Billeke, Uni e sidad del Desa ollo, Chile Fe an A gelague , In ia, F ance *Co espondence: Samuel Ma cos-Pablos [email p o ec ed] Recei ed: 25 Ap il 2016 Accep ed: 08 Augus 2016 Published: 26 Augus 2016 Ci a ion: Ma cos-Pablos S, González-Pablos E, Ma ín-Lo enzo C, Flo es LA, Gómez-Ga cía-Be mejo J and Zalama E (2016) Vi ual A a a o Emo ion Recogni ion in Pa ien s wi h Schizoph enia: A Pilo S udy. F on . Hum. Neu osci. 10:421. doi: 10.3389/ nhum.2016.00421 Pe sons who su e om schizoph enia ha e di icul ies in ecognizing emo ions in o he s’ acial exp essions, which a ec s hei capabili ies o social in e ac ion and hinde s hei social in eg a ion. Pho og aphic images ha e adi ionally been used o explo e emo ion ecogni ion impai men s in schizoph enia pa ien s, bu hey lack o he dynamism ha is inhe en o acial exp essi eness. In o de o o e come hose incon eniences, o e he las yea s di e en au ho s ha e p oposed he use o i ual a a a s. In his wo k, we p esen he esul s o a pilo s udy ha explo ed he possibili ies o using a ealis ic- looking a a a o he assessmen o emo ion ecogni ion de ici s in pa ien s who su e om schizoph enia. In he s udy, 20 subjec s wi h schizoph enia o long e olu ion and 20 con ol subjec s we e in i ed o ecognize a se o acial exp essions o emo ions showed by bo h he said i ual a a a and s a ic images. Ou esul s show ha schizoph enic pa ien s exhibi ecogni ion de ici s in emo ion ecogni ion om acial exp essions ega dless he ype o s imuli (a a a o images), and ha hose de ici s a e ela ed wi h he psychopa hology. Finally, some imp o emen s in ecogni ion a es (RRs) o he pa ien g oup when using he a a a we e obse ed o sadness o su p ise exp essions, and hey e en ou pe o m he con ol g oup in he ecogni ion o he happiness exp ession. This leads o conclude ha , apa om he dynamism o he shown exp ession, he RRs o schizoph enia pa ien s when employing anima ed a a a s may depend on o he ac o s which need o be u he explo ed. Keywo ds: schizoph enia, acial ecogni ion o emo ions, ealis ic i ual a a a INTRODUCTION Schizoph enia is a se ious men al diso de ha a ec s abou 1% o he popula ion. I is he mos expensi e men al illness in e ms o di ec heal h cos s, indi ec p oduc i i y losses, and he impac on he amily and socie y. The e is also a no able impai men o he pe son pe o mance in a eas such as employmen , amily and social in eg a ion (APA, 2002). This diso de is cha ac e ized by a a ie y o symp oms: hallucina ions, delusions, a ec i e la ening, diso ganized beha io , apa hy, social wi hd awal, cogni i e impai men , and o he s. Among hose symp oms, social cogni ion is a majo cons uc o in es iga e in schizoph enia. Social cogni ion has been de ined as he human abili y and capaci y o pe cei ing o he s’ in en ion and p edisposi ion (B o he s, 1990). Fo G een e al. (2005), social cogni ion F on ie s in Human Neu oscience | www. on ie sin.o g 1Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia e e s o he men al ope a ions unde lying social in e ac ions, which include p ocesses in ol ed in pe cei ing, in e p e ing, and gene a ing esponses o he in en ions, disposi ions, and beha io s o o he s. In he case o pa ien s wi h schizoph enia, impai men s in social cogni ion a e p esen be o e he onse o psychosis, sugges ing ha social cogni ion may be a ai ma ke o he illness. Al hough he e is s ill a lack o comple e consensus abou he conc e e scope and signi icance o hose impai men s, some ini ia i es ha e a isen which y o s udy cogni i e impai men s in schizoph enia ha include social domains. Examples o such s udies a e he Measu emen and T ea men Resea ch o Imp o e Cogni ion in Schizoph enia (MATRICS; G een e al., 2000, 2008), and he Cogni i e Neu oscience o T ea men Resea ch o Imp o e Cogni ion in Schizoph enia (CNTRICS; Ca e e al., 2009). Following hese amewo ks, h ee majo a eas o social cogni ion in Schizoph enia can be conside ed: he heo y o mind, a ibu ion s yle and he acial ecogni ion o emo ions (Penn e al., 2008). The p esen s udy ocuses on he las o hose a eas. Facial Recogni ion o Emo ions Conce ning emo ion ecogni ion, subjec s wi h schizoph enia show a de ici in acial emo ion disc imina ion and iden i ica ion. Also, his de ici has been obse ed in o he diseases such as dep ession, mania, demen ia, b ain inju y, au ism, e c. (Mandal e al., 1998; Penn e al., 2001; Pinkham e al., 2003). Se e al wo ks ha e ocused on he abili y o schizoph enic pa ien s o ecognize ace emo ions. Rega ding he clinical condi ion o pa ien s wi h schizoph enia, he de ici in he acial ecogni ion o emo ions is a pe manen ea u e o he disease. This de ici is p esen om he beginning o he illness and ends o s ay in mo e ch onic s ages (Kee e al., 2003). Pa ien s ha e mo e di icul y o ecognize nega i e emo ions such as ange o ea , in e p e ing hem w ongly as neu al exp essions (Kohle e al., 2003). In ela ion o he symp oms, i seems ha subjec s wi h delusions o pe secu ion and pa anoia ha e be e ou comes han subjec s wi h schizoph enia, bu wi hou pa anoia (Lewis and Ga e , 1995). Pa ien s wi h nega i e symp oms would ha e u he de e io a ion in hese es s (G een and Nuech e lein, 1999). Wi h espec o gende , women misin e p e neu al aces as sad ones mo e o en han men, and males epo ed neu al aces as ang y ones (Weiss e al., 2007). The abili y o ecognizing acial emo ions seems o be ela ed o o he cogni i e unc ions, as well as o a ious aspec s o social unc ioning and he quali y o li e (Kee e al., 2003; Kohle e al., 2010). Fo example, in Sil e e al. (2002) some au ho s ela e social cogni ion impai men s o al e a ions in he memo y, abs ac hinking, language p ocessing and a en ion. Apa om emo ional ecogni ion impai men diagnosis, many psychosocial in e en ion p og ams in he ield o social cogni ion (including emo ion ecogni ion) ha e been used o yea s: Social Skills T aining (SST; Libe man e al., 1985), In eg a ed Psychological The apy (IPT; Rode e al., 1996) and T aining in Emo ional In elligence (TEI; Vau h e al., 2001). Also, speci ic social cogni ion ea men echniques ha e also been de eloped mo e ecen ly: Social Cogni ion and In e ac ion T aining (SCIT; Penn e al., 2005), Social Cogni ion Enhancemen T aining (SCET; Choi and Kwon, 2006), Me aKogni i es T aining (MKT; Mo i z and Woodwa d, 2007) and T aining o A ec Recogni ion (TAR; Wölwe e al., 2005). Mos o hose s udies and ea men p ocedu es ocus on indi idual mechanisms and obse a ional pe spec i es. Howe e , du ing he las yea s, esea che s ha e s a ed o s udy schizoph enic pa ien s du ing in e ac i e si ua ions (Pa k e al., 2011; Lee e al., 2013). The main goal is o make use o i ual en i onmen s and a a a s o p o ide new objec i e me hods o assessing pa ien s’ in e pe sonal beha io cha ac e is ics in a social se ing. In e ms o emo ion ecogni ion, one d awback o cu en diagnosis and ea men p ocedu es esides in he ac ha he e is a lack o a me hodology ha allows co ela ing he exis ing me hods o encoding and pa ame e izing acial mo emen s wi h he di e en cogni i e impai men s. To da e, se e al me hods ha e been p oposed o acial mo ion pa ame e iza ion: Facial Ac ion Coding Sys em (FACS; Ekman and F iesen, 1978; Ekman e al., 2002), Maximally Disc imina i e Facial Mo emen Coding Sys em (MAX; Iza d, 1983), and AFFEX (T onick e al., 2002). Howe e , mos p og ams o he iden i ica ion and ea men o a ec ion ecogni ion diso de s in pa ien s wi h schizoph enia ha e been based on he use o s a ic images o acial emo ions. These images a e ob ained om di e en exis ing acial coding sys ems, he FACS being one o he mos accep ed ones. The use o his kind o image esul s in a ha dly-pa ame e izable iden i ica ion and ea men p ocess. In addi ion, s a ic images lack he impo an in o ma ion ha is inhe en o ace o ace social in e ac ions, so he pe son unde s udy (ei he su e ing om schizoph enia o no ) may no co ec ly di e en ia e he ac ual emo ion (Lemay e al., 2000; Weh le e al., 2000; Adams e al., 2003; Klucha e and Sams, 2004; Joyal e al., 2014). The e o e, he sea ch o o he e alua ion and ea men ools o hese pa ien s is an in e es ing challenge. One powe ul ool o enhance expe imen al con ol o e such dynamic social in e ac ions has been he use o a a a s o i ual eali y (Wu e al., 2014), and also he use o pic u e mo phing in e aces (Fuku a e al., 2014). Fo example, a ecen s udy shows ha he use o i ual eali y ools may p esen a numbe o ad an ages when dealing wi h pa ien s su e ing om schizoph enia, such as symp om assessmen , iden i ica ion o symp om ma ke s, o he in es iga ion o he di e en ial p edic ion o symp oms (F eeman, 2008). I has also been shown ha he use o i ual cha ac e s (a a a s) can cha ac e ize, in a simple way, dis o ions in he social pe cep ion o pa ien s (Kim e al., 2007). O he s udies show ha he dis o ions o eali y and memo y can be assessed using i ual en i onmen s (So kin e al., 2008; Wenige and I le, 2008). This ype o en i onmen s has also been applied as a he apy o con e sa ional aining p og ams in pa ien s wi h schizoph enia (Ku e al., 2006, 2007). Focusing on he emo ional ecogni ion h ough acial exp essions, he use o a a a s has a numbe o ad an ages F on ie s in Human Neu oscience | www. on ie sin.o g 2Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia o e con en ional me hods based on s a ic images, as hey allow including he dynamic aspec o he acial exp essi eness, as well as he con ol and pa ame e iza ion o he exp ession iming and in ensi y, o he symme y le el and he du a ion o each exp ession. The e o e, he use o a a a s could p o ide an e icien ool o he diagnosis and ea men o de iciencies in social and emo ional beha io . Howe e , while a a a s a e well sui ed o s udy human cogni ion (Joyal e al., 2014), he e a e s ill many open ques ions conce ning bo h he heo e ical aspec s o employing i ual a a a s in psychology and hei p ac ical applica ion (Ayme ich e al., 2014). As such, u he in es iga ion needs o be done in o de o assess he sui abili y o employing i ual a a a s in schizoph enia diagnosis and ea men . Expe imen al Goals and Hypo heses The use o i ual agen s cons i u es a po en ial in e media e link be ween s anda dized es s and eal-li e unc ioning, and also o e s he possibili y o being used as an inno a i e medium o cogni i e emedia ion (Oke e al., 2015). As such, he objec i e o he p esen s udy is o explo e he possibili ies o using a ealis ic- looking a a a p e iously in oduced in Ma cos e al. (2010) o he assessmen o emo ion ecogni ion de ici s in pa ien s who su e schizoph enia. Al hough hey a e ou o he scope o his pilo s udy, some o he bene i s o an a a a o e s a ic images include: o allow he pa ame e iza ion and ine uning o he shown exp essions; o o co ela e he di e en exp ession in ensi y, iming and o he cha ac e is ics o acial exp essi eness wi h he pa ien ecogni ion a es (RRs) and psychopa hology. In his con ex , we compa ed he esponse o a g oup o pa ien s wi h schizoph enia and a con ol g oup using bo h a classical emo ion ecogni ion es based on s a ic images, and a new es based on a dynamic a a a . Conside ing p e ious s udies (Ku e al., 2006, 2007; Dyck e al., 2010) he e is e idence ha schizoph enia pa ien s would also show de ici s in he ecogni ion o emo ional exp essions displayed by a i ual a a a . As such, ou i s hypo hesis was: Hypo hesis 1: Compa ed wi h con ols, pa ien s will show wo se ecogni ion sco es, ega dless he ype o s imuli (a a a o images). The bene i s o dynamic acial exp essions ha e been discussed la ely. Some s udies sugges ha he dynamic p ope ies o human acial mo emen s may play a su p isingly small ole in people’s abili y o in e he emo ional s a es o o he s om hei acial exp essions (Gold e al., 2013). O he s show ha he e is no di e ence in he use o dynamic o s a ic images as a combina ion o hype sensi i i y o s a ic emo ions and hyposensi i i y o dynamic emo ions in people wi h schizoph enia migh unde lie he absence o di e ences in esponse o hese s imuli (Fuku a e al., 2014). On he o he hand, he e is also e idence ha he use o dynamic acial exp essions, apa om being ecologically mo e alid and he e o e mo e app op ia e o emo ion esea ch, helps in he pe cep ion o emo ions (K umhube e al., 2013). Based on he abo e assump ions, he second hypo hesis we wan ed o con as was: Hypo hesis 2: The a a a would u n ou o ha e be e ecogni ion sco es han s a ic images. Finally, many s udies ha e ela ed he acial emo ion ecogni ion impai men s o schizoph enia pa ien s wi h hei psychopa hology (Dyck e al., 2010; Gold e al., 2013) and o he s. Fo ha eason, ou hi d hypo hesis is: Hypo hesis 3: I p esen , ecogni ion de ici s o he pa ien s g oup a e co ela ed wi h he psychopa hology. MATERIALS AND METHODS Fo he expe imen , bo h pa ien s and con ols we e asked o en e an o ice indi idually and o ake a sea a a desk which held a compu e sc een. The he apis hen explained hem ha some acial exp essions co esponding o he six basic emo ions: joy, sadness, ea , ange , su p ise and disgus we e going o be shown on he sc een. I was also explained ha exp essions we e going o be pe o med by an a a a ha will s a in a neu al pose o be shown in a s a ic image, and ha he goal was o y and guess he exp ession shown. The he apis insis ed on he ac ha he e we e no good o bad answe s, and ha he main pu pose o he es was o collec he pa icipan s’ pe cep ion. The he apis hen asked he pa icipan s i hey had unde s ood he ins uc ions o he es , epea ing hem in he case he pa icipan did no . The he apis ied o main ain a neu al one o oice h oughou he expe imen , in o de no o in luence he esponses. When he pa icipan was eady, he he apis showed he di e en emo ional exp essions on he PC sc een. Bo h he s a ic images and he a a a we e in e spe sed andomly and p esen ed o he use , and he he apis wai ed up o 20 s o a esponse o be gi en. Images showed a s a ic emo ional exp ession (see Figu e 1). On he o he hand, he exp ession pe o med by he a a a s a ed om a neu al pose and ended in he exp ession maximum. A a a main ained he maximum o he exp ession un il he he apis mo ed on o he nex exp ession ei he because o a esponse gi en o i he conside ed a imeou . Co ec answe s we e labeled as 1, whe eas bo h w ong o lack o answe we e labeled as 0. Subjec s Pa icipan s in he s udy consis ed on wo g oups: pa ien and con ol. The pa ien g oup consis ed o 20 subjec s ea ed o schizoph enia a he San Luis hospi al (Palencia, Spain). On an a e age, hey a e middle-aged pa ien s and wi h a long e olu ion o he disease. A e he esea ch eam checked ha hei cogni i e le el allowed hem o unde s and ins uc ions and he pu pose o he es , hey we e asked o olun a ily pa icipa e in he s udy. The con ol g oup consis ed o 20 subjec s andomly ec ui ed among heal hy wo ke s o he same cen e . Table 1 con ains a su ey o he demog aphics o bo h g oups. Symp oma ology was assessed by e sion A o he Posi i e and Nega i e Synd ome Scale (PANNS; Kay e al., 1987). F on ie s in Human Neu oscience | www. on ie sin.o g 3Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia FIGURE 1 | The chosen s a ic emo ions. Ma e ials and Equipmen S a ic Images Fo he s a ic images, six pho os om he Paul Ekman model e lec ing six uni e sal emo ions pe o med by di e en ac o s we e selec ed. Images we e selec ed so ha he shown exp ession was isually simila o he one pe o med by he a a a , bu also looking ha he ac o s ha pe o med hem we e om di e en aces and gende s in o de o a oid biased esul s. Figu e 1 shows he selec ed images o he di e en exp essions. A a a In e ace The de eloped a a a ollows a bio-inspi ed design looking o ealis ic esul s and consis ing o wo majo s eps: modeling and anima ion. The modeling s ep is ca ied ou using a lase scanne ha allows h ee-dimensional in o ma ion o a eal ace o be ob ained. Measu ed poin s a e hen meshed in o de o ob ain a 3D su ace model. Face ex u e (colo ) is ob ained om sepa a e pho og aphs ha a e mapped on o he 3D su ace. This allows a deg ee o ealism highe han ha p o ided by o he modeling me hods o be achie ed. Ce ain elemen s (eyes, ee h, o al ca i y) ha e been modeled sepa a ely and hen included in he whole model, since hey could no be scanned. The anima ion has been done h ough pseudo-muscles g ouped in o a hie a chical skele on. The pseudo-muscles achie e a good app oxima ion o he eal acial muscles and ha e he ad an age o equi ing a lowe compu a ional e o , which makes hem ideal o in e ac i e eal- ime applica ions. The pseudo-muscle anima ion p ocess has been ca ied ou aking in o accoun bo h he ac ual muscula physiognomy and ana omy. Once posi ioned, he pseudo-muscles a e used o p ope ly bend he h ee-dimensional mesh. The subsequen de o ma ion is also based on he ana omy o he human ace in he egions o muscle inse ions in o he skin. Figu e 2 shows a scheme o he di e en s ages o he p ocess o TABLE 1 | Demog aphics o bo h con ol and pa ien g oups. Age Sex PANNS posi i e PANNS nega i e PANNS psychopa ology PANNS o al Mean S.D. Male Female Mean S.D. Mean S.D. Mean S.D. Mean S.D. Con ol 42.7 10.8 7 13 – – – – – – – – Pa ien s 55.0 8.3 11 9 21.8 7.5 27.4 9.3 47.1 11.3 97.0 27.8 F on ie s in Human Neu oscience | www. on ie sin.o g 4Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia FIGURE 2 | S ages o he p ocess o cons uc ion o he a a a . Modeling (le ) and anima ion ( igh ). modeling and anima ion. A mo e de ailed desc ip ion o he a a a cons uc ion and de elopmen is gi en in Ma cos e al. (2010). Figu e 3 shows he se o pseudo-muscles used and hei posi ion on he i ual ace. The ac ion o hese pseudo-muscles p oduce de o ma ions on he i ual 3D model in he same way as eal acial muscles do on he ace su ace. The pseudo-muscle mo emen s a e based on he minimum ac ion uni s (AUs) om he FACS. Complex exp essions can be gene a ed h ough he combina ion o AUs wi h di e en in ensi ies, as shown in Figu e 4. Fo his expe imen , we ha e gene a ed six basic dynamic exp essions: happiness, disgus , ange , ea , sadness and su p ise (See Figu e 5). Also, he con ol in e ace o he a a a has been simpli ied in he cu en s udy, in such a way ha he he apis only has o p ess a keyboa d key o make he a a a pe o m a gi en exp ession. Al hough a iables like in ensi y and iming can be modi ied, hey ha e been kep cons an in o de o s anda dize he exp essions shown o he pa icipan s o he expe imen . FIGURE 3 | Pseudo-muscles used in he a a a . 1. F on ; 2. Co uga o supe cilii; 3. P oce us; 4. The eye muscles and o bicula o he eye; 5. Quad a uslabiisupe io is; 6. Zygoma icus majo ; 7. Zygoma icus mino ; 8 Neck; 9. Buccina o ; 10. Giggly; 11. O bicula mou h; 12. Jaw; 13. T iangula om bea d. P ocedu e The s udy was ca ied ou indi idually o each pa icipan . Facial s imuli in he o m o s a ic images and eal- ime dynamically gene a ed a a a exp essions we e p esen ed o each pa icipan , h ough a compu e sc een, in a andom o de . In o al, pa icipan s we e asked o iden i y he emo ional exp essions exhibi ed by six s a ic images and six a a a exp essions, each co esponding o he six uni e sal emo ional exp essions: ange , ea , disgus , su p ise, sadness and happiness. Responses we e eco ded as success ul i he emo ional exp ession is co ec ly iden i ied o as a ailu e i he emo ional exp ession is w ongly iden i ied o no esponse is gi en o a 20 s ime limi . E hics This c oss-sec ional s udy was app o ed by he Clinical Resea ch E hics Commi ee o he Hospi al i was pe o med in. In o med, olun a y, w i en consen was ob ained om he pa ien s and con ols a e a ho ough explana ion o he p ocedu es, in acco dance wi h he Decla a ion o Helsinki. ANALYSIS AND RESULTS All analyses we e pe o med using he so wa e SPSS, wi h he alpha le el a 0.05 o s a is ical signi icance, unless o he wise s a ed. Desc ip i e S a is ics In e ms o desc ip i e s a is ics, a compa ison o RRs o each o he emo ional exp essions o he wo g oups o pa icipan s has been done, by conside ing he esponses ob ained o he a a a agains hose ob ained o s a ic images. The RRs o he each g oup (pa ien s/con ol) a e shown in Figu e 6. Fo he g oup o pa ien s, i can be obse ed ha he success a e o exp essions o ange and happiness a e highe o he s a ic images, while exp essions o sadness and su p ise a e mo e easily ecognizable o he anima ed a a a . The exp ession o ea has simila RRs in bo h cases. Fo he con ol g oup, sadness, happiness and ange we e be e ecognized in he case o images, while sadness and su p ise exp essions we e be e ecognized in FIGURE 4 | Neu al exp ession and ange exp ession wi hin he ac ion uni s (AUs) o he acial ac ion coding sys em (FACS) ha gene a e i . F on ie s in Human Neu oscience | www. on ie sin.o g 5Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia FIGURE 5 | The six emo ional exp essions gene a ed by he a a a . F om op o bo om and om le o igh : happiness, disgus , ea , ange , su p ise and sadness. he case o he a a a . Finally he ea exp ession has a qui e low RR. I has o be no ed ha in he case o he exp ession o disgus , he exp ession gene a ed by he a a a was no ecognized by any pa icipan , hus i was excluded om u he analysis. We plan o ebuild his exp ession o u u e ealiza ions. Also, Table 2 shows a compa ison o emo ion ecogni ion sco es be ween bo h g oups o pa icipan s. I can be seen ha in he case o s a ic images he RRs o he con ol g oup a e be e on a e age o all exp ession han hose o he pa ien g oup. In he case o he a a a , howe e , con ol g oup only ou pe o ms pa ien g oup in he case o su p ise and ange exp essions. Recogni ion Ra es A e aged o All Emo ions The answe s o he pa icipan s we e coded as a binomial a iable labeled as ecogni ion ( ecogni ion success/ ecogni ion ailu e), which in u n was conside ed as he dependen a iable o ou s udy. A se o bi a ia e analysis was i s conduc ed in o de o explo e he ela ionship be ween he dependen a iable and he o he ac o s: g oup (con ol/pa ien s), ace ype (a a a /images), sex,age and emo ion (happiness, sadness, ange , ea , su p ise). Table 3 su eys he esul s o he di e en es s o each conside ed ac o . Simple bina y logis ic eg essions we e employed o he ca ego ical a iables, whe eas a S uden ’s - es was employed o he con inuous age a iable. Independen F on ie s in Human Neu oscience | www. on ie sin.o g 6Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia FIGURE 6 | Recogni ion a es ob ained o he a a a and o he s a ic images o each g oup (pa ien s/con ols). samples - es showed no signi ican age di e ences be ween g oups. As Table 3 shows, a iables ace ype and sex did no esul s a is ically signi ican (p=0.614 and p=0.212 espec i ely o he Wald c i e ion). On he o he hand, s a is ically signi ican esul s we e ob ained o he a iable g oup (p<0.001) wi h an OR =0.466 when conside ing he F on ie s in Human Neu oscience | www. on ie sin.o g 7Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia TABLE 2 | Mean and s anda d e o o he di e en emo ions. Con ol Pa ien s A a a Images A a a Images Ange 0.6/0.112∗0.90/0.069 0.35/0.109 0.60/0.112 Happiness 0.65/0.109 1/0 0.70/0.105 0.95/0.050 Sadness 0.70/0.105 0.25/0.099 0.60/0.112 0.25/0.099 Fea 0.2/0.092 0.50/0.115 0.30/0.105 0.35/0.109 Su p ise 1/0 0.80/0.092 0.45/0.114 0.20/0.092 ∗Mean/SE. con ol g oup as a e e ence, indica ing ha he RR o he con ol g oup is o e all wo imes highe han he one o he pa ien g oup. S a is ically signi ican esul s we e also ob ained o he emo ion a iable. In his case he emo ion a iable is poly omous, so in o de o e alua e a eg ession model, one o i s alues is conside ed as a e e ence (e.g., su p ise), while he es a e conside ed as ‘‘dummy’’ binomial a iables, allowing o compu e hei OR. I can be seen ha conside ing he su p ise exp ession as a e e ence, he e esul s a is ically signi ican alues o all exp essions bu ange , indica ing ha emo ion has a signi ican con ibu ion o he RR. In he case o age, s a is ically signi ican esul s we e ob ained o he - es (p<0.001). A u he logis ic eg ession analysis was pe o med o e alua ing he e ec s o age o e ecogni ion, and using age as independen and non- ca ego ical co a ia e. Resul s indica ed ha age indeed had a signi ican e ec o e he RR (p<0,001 o he Wald c i e ion) and an OR =0.950, which indica es ha ecogni ion success dec eases wi h age. Fu he explo a ions showed no s a is ically signi ican in e ac ions be ween co a ia es. To inc ease he powe o he s a is ical model, non-signi ican co a ia es we e excluded o he analysis. A e aged ac oss all emo ions, a logis ic eg ession analysis was hen conduc ed using age, emo ion and g oup as p edic o s, bu he esul ing model indica ed a weak ela ionship be ween p edic ion and g ouping (Nagelke ke’s R2=0.214), and o e all p edic ion o 69.8% (74.9% o ecogni ion success/63.0% o ecogni ion ailu e). TABLE 3 | Resul s o indi idual es s o he conside ed ac o s. Independen a iable ∗P alue ∗∗ OR ∗∗∗ CI (95%) In . Sup. G oup <0.001 0.466 0.311 0.698 Face ype 0.614 1.107 0.745 1.645 Sex 0.212 1.288 0.865 1.917 Age <0.001 4.6 0.405 6.781 Ange ∗∗∗∗ 1 1 – – Happiness∗∗∗∗ 0.003 2.983 – – Sadness∗∗∗∗ 0.04 0.518 – – Fea ∗∗∗∗ 0.001 0.322 – – ∗Wald c i e ion o dicho onomous a iables, - es o con inuous a iable age. ∗∗ Odds Ra io (OR) o dicho onomous a iables, mean di e ence o con inuous a iable age. ∗∗∗ Con idence in e al (CI). ∗∗∗∗ E alua ed using Su p ise emo ion as e e ence. Recogni ion Ra es o Each Emo ion Sepa a ed logis ic eg ession analyses we e hen conduc ed o each emo ional exp ession (see Table 4 o a su ey o he esul s). In he case o ange , simple bi a ia e logis ic eg ession analyses showed a s ong ela ionship be ween he dependen a iable ecogni ion and ace ype (p=0.013 o he Wald c i e ion), g oup (p=0.013), and age (p=0.003). No s a is ically signi ican in e ac ions be ween co a ia es we e ound. Logis ic eg ession conduc ed using hose h ee ac o s p oduced a s a is ically signi ican model (chi squa e =19.33, p<0.001 wi h d =3). Nagelke ke’s R2o 0.291 indica ed a weak ela ionship be ween p edic ion and g ouping. P edic ions success o e all was 72.5% (83.7% o ecogni ion success and 54.8% o ecogni ion ailu e). The Wald c i e ion showed ha only ace ype and age made a signi ican con ibu ion o he model. The OR (Exp(B)) indica ed ha success in ecogni ion dec eased wi h age, and ha pa icipan s we e 4161 imes mo e likely o ecognize ange exp ession in he case o images. Howe e , due o he said weakness o he model, we conclude ha in he case o ange he model is unable o co ec ly p edic he ecogni ion success. In he case o happiness bi a ia e logis ic eg ession analyses showed a s ong ela ionship be ween he dependen a iable and ace ype (p=0.006, OR =18.778). No o he co a ia es o in e ac ions we e ound o be signi ican . Fo explo a o y easons, age and g oup we e also included in he model. The eg ession model was s a is ically signi ican (chi squa e =14.738, p<0.002, d =3). O e all p edic ion success was o 82.5%. Wald’s c i e ion showed no signi ican con ibu ion o age and g oup, bu in he case o ace ype i showed a signi ican con ibu ion (p=0.006, OR =18.98) indica ing a signi ican ly highe RR in he case o he images. In he case o sadness only ace ype (p<0.01) and age (p=0.037) we e ound signi ican ly ela ed o he dependen a iable and so we e in oduced in he model. The es o he ull model agains he cons an only model esul ed s a is ically signi ican (chi squa e =18.890, p<0.001 wi h d =2). P edic ion success was 72.5% (69.4% success, 75% ailu e) wi h a Nagelke ke’s R2o 0.281. Again, he RR dec eased wi h age (OR =0.948 o a con inuous a iable). Fo he ace ype he OR using he a a a g oup as a e e ence was 0.156 indica ing ha he ecogni ion success was signi ican ly highe in he case o he a a a . In he case o ea , bi a ia e analyses indica ed no s a is ically signi ican ela ionship be ween he dependen a iable and any o he co a ia es. In he case o su p ise, simple bi a ia e logis ic eg ession analyses showed a ela ionship be ween he dependen a iable ecogni ion and ace ype (p=0.041 o he Wald c i e ion), g oup (p<0.001), and age (p<0.001). No addi ional s a is ically signi ican in e ac ions be ween co a ia es we e ound. Logis ic eg ession conduc ed using hose h ee ac o s p oduced a s a is ically signi ican model (chi squa e =37.903, p<0.001 wi h d =3). Nagelke ke’s R2o 0.512 indica ed a mode a e ela ionship be ween p edic ion and g ouping. P edic ions success o e all was 80% (81.6% o ecogni ion success and 77.4% o ecogni ion ailu e). The Wald c i e ion showed F on ie s in Human Neu oscience | www. on ie sin.o g 8Augus 2016 | Volume 10 | A icle 421 Ma cos-Pablos e al. A a a o Emo ion Recogni ion in Schizoph enia TABLE 4 | Su ey o he esul s o he indi idual emo ions. Exp ession Main e ec s ∗∗ OR Signi icance Ange Face ype (p=0.013) 4.161 Highe RR∗ o images Age (p=0.003) 0.78 RR dec eases wi h age Happiness Face ype (p=0.006) 18.98 Highe RR o images Sadness Face ype (p=0.04) 0.156 Be e RR in he case o he a a a Age (p=0.003) 0.948 RR dec eases wi h age Fea NO SIG NO SIG NO SIG Su p ise Face ype (p<0.001) 0.2 Highe RR in he case o he a a a G oup (p<0.001) 20 Highe RR in he case o he con ol g oup ∗RR, Recogni ion Ra e. ∗∗ Odds Ra io (OR) o dicho onomous a iables, mean di e ence o con inuous a iable age. ha only ace ype and g oup made a signi ican con ibu ion o he model. The OR (Exp(B)) indica ed ha ecogni ion success is likely o be 5×highe in he case o he a a a , and ha he OR is 20×highe in he case o he con ol g oup. Co ela ions wi h Psychopa hology Conside ing only he pa ien g oup, success in he ecogni ion o emo ions p o ed o be co ela ed wi h he posi i e deg ee o symp oma ology acco ding o he PANNS. A simple bina y logis ic eg ession ac oss all emo ions and o bo h he a a a and images using ecogni ion success as dependen a iable showed ha i was signi ican ly ela ed wi h he PANNS o al (p=0.039 o he Wald c i e ion). The Odds Ra io (Exp(B) =0.989) indica ed ha he RR slowly dec eases as PANNS inc eases. O he bi a ia e analysis wi h he es o he co a ia es showed no signi ican ela ion wi h he dependen a iable. Indi idual analysis o each exp ession did no show signi ican di e ences in e ms o PANNS sco es. DISCUSSION Consis en wi h ou hypo hesis, desc ip i e esul s show ha schizoph enia pa ien s p esen de ici s in he ecogni ion o i ual emo ional exp essions when compa ed wi h he con ol g oup. As Table 3 shows, he odds a io o he a e age RRs o he con ol g oup is almos wice as high as hose o he schizoph enia pa ien s. Also, a e aged ac oss all emo ions and o bo h con ol g oups, he eg ession model showed ha di e ences in RRs be ween he a a a and he images we e no s a is ically signi ican . These indings ma ch hose p e iously p esen ed by Kim e al. (2007) and Dyck e al. (2010), showing a educ ion in social pe cep ion o schizoph enia pa ien s e en when employing i ual cha ac e s. Fo he o he conside ed co a ia es, sex did no p oduce s a is ically signi ican e ec s o e he RR, while age p oduced signi ican a ia ions, sligh ly dec easing he RRs as age inc eases. In he con ol g oup, his e ec is somehow expec ed as he e ec s o age o e exp ession ecogni ion displayed by i ual cha ac e s has been ela ed in he li e a u e o he ac ha elde ly people is less used o echnology and i ual en i onmen s (Ma cos e al., 2010). I has also been ela ed wi h he ac ha isual acui y ends o decline wi h ad ancing age o e en o he ac ha olde adul s ha e an a en ional bias owa ds old-age aces (Campbell e al., 2015). In he case o he g oup o pa ien s, u he s udies ha ela e age wi h o he ac o s due o he e olu ion o he schizoph enia shall be done in o de o p ope ly add ess his e ec . Rega ding he ela ionship be ween ecogni ion de ici s and psychopa hology, ou indings also e eal an expec ed associa ion be ween symp oma ology and emo ion ecogni ion, as RRs dec ease as PANNS sco es inc ease. Howe e , no s a is ically signi ican di e ences we e ound be ween he na u al images and he i ual ones, sugges ing ha his ela ionship is independen o he ype o acial s imuli employed. These esul s ma ch hose o Fuku a e al. (2014), and could be ela ed o a combina ion o hype sensi i i y and hyposensi i i y o s a ic and dynamic emo ions espec i ely. Howe e , in he case o ou second hypo hesis, he ecogni ion sco es ob ained by he a a a do no ou pe o m hose ob ained om he images as expec ed. Analyzing each emo ional exp ession indi idually, o he ange exp ession s a is ically signi ican esul s we e ound be ween he g oups o pa ien s and con ol as well as when compa ing esul s o he a a a and s a ic images. In his case, he ange exp ession is be e ecognized when using s a ic images and he con ol g oup clea ly ou pe o ms he pa ien g oup. In he case o happiness, s a is ically signi ican di e ences we e ound be ween ace ype, being he exp ession be e ecognized in he case o he s a ic images. In he case o sadness, ace ype was again ound o be a main e ec o e he RR along wi h age. I can be seen, howe e , ha in his case he a a a clea ly ou pe o ms he s a ic images. No s a is ically signi ican main e ec s we e ound o he ea exp ession. Finally, in he case o sadness, bo h he ace ype and he g oup o pa icipan s u ned ou o be s a is ically signi ican o he di e ences ound in he RR. As Kohle e al. (2003) s a es, in he case o images pa ien s ha e mo e di icul y o ecognize nega i e emo ions such as ange , ea , o sadness. Al hough o he s udies ha e shown sligh imp o emen s in he ecogni ion o nega i e emo ions when using i ual aces (Dyck e al., 2010), in ou case his is only ue o sadness bu in a highe deg ee. I is also in e es ing o poin ou ha pa ien s ou pe o m he con ol g oup in he ecogni ion o he happiness exp ession F on ie s in Human Neu oscience | www. on ie sin.o g 9Augus 2016 | Volume 10 | A icle 421