- Email: [email protected]
The effect of accomplishment and failure on P300 potentials evoked by neutral stimuli
\ PERGAMON
Neuropsychologia 26 "0888# 302Ð319
The e}ect of accomplishment and failure on P299 potentials evoked by neutral stimuli Andrzej Michalski Department of Neurophysiology\ Nencki Institute of Experimental Biology\ 2 Pasteur St\ 91!982 Warsaw\ Poland Received 04 September 0886^ accepted 06 March 0887
Abstract The e}ects of emotional states of {being successful| vs[ {being unsuccessful| were studied by measuring the P299 component of event!related potentials "ERPs#[ Nine subjects were instructed to reduce their P299 amplitude using feedback[ Feedback was random but the relative probability of di}erent signals created the situations of {being successful| or {being unsuccessful|[ The probability of {small| feedback was 9[6 in {successful| and 9[04 in {unsuccessful| trials[ ERPs recorded without the feedback were used as a reference[ Potentials\ evoked by light stimuli in a standard {odd!ball| procedure\ were recorded from Fz\ Cz and Pz scalp sites[ The amplitudes of P299 components were reduced in {unsuccessful| trials whereas in {successful| trials they did not di}er signi_cantly from responses recorded without the feedback[ There were no signi_cant di}erences in peak latencies[ These _ndings indicate that tonic emotional states a}ect the processing of neutral stimuli and that late components of ERPs can be useful indices in the analysis of these alterations[ The results also indicate that the e}ects of positive and negative emotional states are not always reciprocal[ Manipulated feedback is suggested as an useful model in the studies of emotions[ Data can also facilitate the interpretation of the real feedback e}ects[ Þ 0888 Elsevier Science Ltd[ All rights reserved[ Keywords] P299^ Success^ Failure
0[ Introduction There have been numerous attempts to study the e}ect of emotions on the P299 potential[ In a number of experi! ments facial expressions have been used\ since these are stimuli considered to be phylogenetically specialized for emotional communication ð7Ł[ The results were con! tradictory[ In some studies the neutral faces evoked lower amplitudes than emotional ones ð0\ 1\ 19Ł^ in others they evoked the highest responses ð23Ł[ Testing stimuli other than faces\ Radilova and co!workers ð16Ð18Ł showed that erotic stimuli produced more positive P299 waves than non!erotic pictures[ Johnston and co!workers ð05Ð07Ł showed that both pictures of nude models of opposite sex and pictures of dermatological diseases evoked grater P299 amplitudes than neutral pictures[ This _nding was con_rmed by Palomba and co!authors ð12\ 13Ł[ Apart from the di}erences in experimental methods\ one element of the procedure was common in all these studies] the same stimulus that evoked P299 potential also trig! gered the emotional response[
Corresponding author[ Tel[] ¦37 11 548 74 60^ fax] ¦37 11 11 42 31^ e!mail] michanÝnencki[gov[pl
To resolve the discrepancy between the results of experiments on hemispheric localization of emotions\ Sobotka and co!authors ð21Ł suggested di}erent mech! anisms for the brief reactions to emotional stimuli and the more prolonged emotional states that a}ect processing of the neutral stimuli[ The analysis of available data led these authors\ in agreement with the hypothesis originally developed by Gardner ð09Ł\ to conclude that patterns of brain functions associated with judgments of the emotional content of stimuli\ such as facial expressions\ indicated the dominance of right hemisphere for both positive and negative emotions[ On the other hand\ the charged emotional states associated with brain damage ð8\ 29Ł\ injections of sodium amytal ð22Ł or watching emotional _lms ð4Ł have been reported to produce pat! terns of activation that support an alternative hypothesis of Goldstein ð00Ł\ that the left hemisphere is dominant for positive emotions and right hemisphere for negative emotions[ Such di}erences suggest that event!related potentials\ especially the late components\ may also be di}erentially a}ected by prolonged emotional states and by brief responses to the emotional component of evok! ing stimuli[ The aim of the present experiment was to analyze the e}ects of tonic emotional states "being successful vs being
9917Ð2821:88:,*see front matter Þ 0888 Elsevier Science Ltd[ All rights reserved PII] S 9 9 1 7 Ð 2 8 2 1 " 8 7 # 9 9 0 9 5 Ð 6
303
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
unsuccessful# on P299 component of the ERPs evoked by non!emotional stimuli "red and yellow ~ashes of light#[ Since random feedback procedures were used to manipu! late the emotional states\ the results will also be relevant to the interpretation of the real feedback e}ects ð20Ł[ 1[ Methods 1[0[ Subjects Volunteers were informed that they were being asked to participate in standard feedback experiments and their consent was obtained[ Before the experiment\ each par! ticipant was asked to evaluate his:her level of anxiety and mood state using a ten!point scale[ The {normal| level of anxiety and the {neutral| mood corresponded to value 4 on the scale[ Data were collected from nine volunteers of both sexes "seven females and two males#\ aged 18Ð37 years[ They were all familiar with the laboratory and they all reported their anxiety level and mood state to range between 3 and 5 on the scale[ Two additional participants "at the end of the series# expressed their concern that {the experiment was strange|\ and they were excluded from the experiment[ 1[1[ Recording EEG signals were recorded with disc electrodes glued at Fz\ Cz and Pz positions\ referenced to linked mastoids and supplemented by vertical and horizontal EOG[ Sig! nals were sampled with 1937 Hz frequency\ 01 bit resol! ution\ digitally _ltered 9[05Ð29 Hz and reduced to 145 Hz by averaging the adjacent points "Elmiko Paperless EEG system#[ Data were stored in epochs containing 149 ms before and 0 s after the stimulus onset[ Epochs were rejected by computer program if EOG amplitude exceeded 39 mV[ Rejected epochs were replaced with the new ones[ Flashes of spatially overlapped arrays of red and yel! low LED diodes "1> 1> of the visual angle\ 09 cd:m1 luminosity on 0 cd:m1 background\ 099 ms duration# were used as stimuli in a standard odd!ball procedure "stimulus rate*1[4 s^ target probability*9[01^ minimum number of good target stimulus repetitions*29#[ Sub! jects were asked to mentally count the target stimuli and report the number seen at the end of the recording[ Data were rejected if the error in counting was bigger than 19)[ 1[2[ Procedures To reduce the number of variables\ the yellow LED diode was always used as a target and the task instruction was always to reduce the amplitude of P299 potentials "downtraining#[ Downtraining was selected because it
had produced bigger e}ects in an earlier experimental attempt to use the real feedback paradigm "unpublished data from the same laboratory#[ Subjects were asked to reduce the P299 amplitude using feedback provided by the spoken words {small|\ {medium| or {large|\ pro! nounced by the experimenter 0 s after the onset of each target stimulus[ The words were taken from tables which ensured their random distribution\ regardless of the amplitude of real potentials[ In the {successful| trials the probability of {small| was 9[6\ {medium|\ 9[04 and {large|\ 9[04[ In the {unsuccessful|\ trials the probability of {small| was 9[04\ {medium|\ 9[04 and {large|\ 9[6[ The exper! imental session always contained a sequence of _ve rec! ordings] 0[ recording without the feedback\ 1[ randomly selected {successful| or {unsuccessful| feed! back recording\ 2[ recording without the feedback\ 3[ complementary {successful| or {unsuccessful| feedback recording\ 4[ recording without the feedback[ 1[3[ Data analysis Epochs of EEG were averaged\ stored and printed using Elmiko software package[ Di}erence curves were computed for each pair of target and non!target ERPs[ Grand!averaged waveforms were computed by averaging data for all subjects in each experimental condition[ Stat! istical analysis followed two basic strategies] 0[ Amplitudes and latencies of peaks were measured using cursors on the computer screen and the data were transferred to SYSTAT program for multi!factor analysis of variance "ANOVA#[ 1[ Grand!averaged curves\ computed for each exper! imental condition\ were compared[ Within the regions of visible di}erences the corresponding points of grand!averaged curves were expanded into sets of paired data "points of individual di}erence curves# and compared with Student{s t!test[ Direct comparison of the curves avoided the arbitrary peak identi_cation\ on the other hand\ it did not allow the {latency cor! rections| "the scatter of peak latencies could a}ect the grand averaged waveshapes#[ 2[ Results The responses to target stimuli\ grand!averaged over all subjects\ are shown in Fig[ 0[ Three sets of curves represent the data recorded from Fz "top#\ Cz "middle# and Pz "bottom# electrodes[ Dotted lines show the wav! eforms obtained without the feedback[ Dashed lines show the potentials recorded with the dummy feedback that contained 69) of {small| messages[ Since the instruction
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
304
Fig[ 0[ Grand!averaged ERPs evoked by target stimuli[ Potentials were recorded with Fz "top#\ Cz "middle# and Pz "bottom# electrodes[ Dotted lines show the ERPs recorded without the feedback[ Dashed lines show the ERPs recorded with the feedback and the instruction to reduce the P299 amplitude[ Feedback was random but it contained 69) of {small| signals[ The subjects were {successful|[ Solid lines show the ERPs recorded with the feedback and the instruction to reduce the P299 amplitude[ Feedback was random but contained only 04) of {small| signals[ The subjects were {unsuccessful|[
was to reduce the P299 amplitude\ the subject was {suc! cessful|[ This situation will be referred to as a {positive context|[ Solid lines show the recordings obtained with the dummy feedback that contained 04) of {small| mess! ages[ Since the instruction was still to reduce the P299
amplitude\ the subject was {unsuccessful|[ This situation will be referred to as a {negative context|[ In addition to the early waves\ all the recordings show the well developed late components within the P199 and P299 regions[ It is clear that the experimental procedures
305
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
a}ected the P299 waves much more than the earlier com! ponents[ Both types of feedback reduced the P299 ampli! tude but in the case of negative context the reduction was greater[ The P299 peak was de_ned as the positive value of greatest magnitude between 149 msÐ399 ms following the stimulus onset\ and the P199 peak was de_ned as the largest positive value recorded between 049 msÐ189 ms after the stimulus onset[ Amplitudes\ measured in indi! vidual ERPs relative to 149 ms pre!stimulus baseline\ were analyzed with two way ANOVA "electrode context#[ Context variable included the positive context\ negative context and no feedback condition[ Di}erences
in both P299 and P199 peak amplitudes due to the context type were insigni_cant[ Latencies of both peaks\ analyzed with the identical two way ANOVA\ also did not show any signi_cant e}ects of the context[ The grand!averaged potentials evoked by non!target stimuli are illustrated in Fig[ 1 "_gure conventions are the same as in Fig[ 0#[ Besides the early components\ recordings from all electrodes showed well!developed P199 waves and only greatly reduced P299 components[ The comparison of Figs 1 and 0 indicates that di}erences\ due to the type of the context\ were much greater in the target responses[ In the individual non!target recordings\ the very small
Fig[ 1[ Grand!averaged ERPs evoked by non!target stimuli[ Figure conventions are the same as in Fig[ 0[
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
size of the P299 components made it impossible to ident! ify and measure this component reliably[ Thus\ only P199 waves were analyzed[ Two way ANOVA "electrode context# showed that context type clearly did not produce signi_cant di}erences\ either in amplitudes or in latencies of P199 waves in non!target responses[ However\ in spite of their statistical insigni_cance\ the di}erences associated with the negative context in target
306
responses looked promising[ The concentration of di}erences within the P299 region indicated that they were not entirely due to the random scatter[ In an attempt to remove some of the sources of uncontrolled amplitude ~uctuations\ di}erence curves were computed by sub! tracting ERPs evoked by non!target stimuli from those evoked by target stimuli[ These di}erence curves\ grand! averaged for all subjects\ are shown in Fig[ 2 "_gure
Fig[ 2[ Grand!averaged di}erence curves computed by subtracting ERPs evoked by non!target stimuli from ERPs evoked by target stimuli[ Figure conventions are the same as in Fig[ 0[ Black diamonds indicate the points of signi_cant di}erences between the curves recorded in {unsuccessful| trials "solid lines# and the control\ no!feedback trials "dotted lines#[
307
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
conventions are the same as in Figs 0 and 1#[ All di}erence curves show the distinct peaks of P199 and P299 com! ponents[ P199 amplitudes are clearly smaller in the di}erence curves than in target responses\ whereas P299 amplitudes are almost the same[ It is apparent that\ like in target ERPs\ dummy feedback a}ects predominantly the P299 component[ The asymmetry between the e}ects of positive and negative context seems to be even more pronounced in the di}erence curves] within the P299 region the dashed lines "positive context# and the dotted lines "no feedback# are closer to each other than in target ERPs "Fig[ 0#[ In contrast\ the solid lines "negative con! text# show the marked di}erences[ Within the P199 region the di}erence curves show the e}ect that could not be seen in target ERPs] both types of dummy feedback "solid and dashed lines# reduce this component with respect to the no feedback condition "dotted lines#[ The amplitudes and latencies of P299 and P199 com! ponents\ measured in the di}erence waves\ are presented in Table 0[ A two!way ANOVA "electrode context# showed a signi_cant e}ect of the type of context on the P299 amplitude "df 1\ F 2[00\ P ³ 9[94#[ Post!hoc tests showed that the amplitudes of P299 waves recorded with negative context were signi_cantly smaller than the amplitudes recorded without the feedback "Bonferroni\ P ³ 9[933#[ Neither the di}erences in P299 peak latencies nor the di}erences in P199 amplitudes and latencies were signi_cant[ Black diamonds beneath the curves in Fig[ 2 indicate the signi_cant di}erences between the corresponding points of negative context and no!feedback curves[ At every 13!ms\ within the range of 091 msÐ475 ms\ the corresponding points of grand!averaged curves were expanded into sets of paired data "points of individual di}erence curves# and compared with the Student{s t!test[ All diamonds are clearly located within the P299 region[ When the positive and negative context curves were com!
Table 0 The e}ect of positive and negative context on the parameters of P299 and P199 waves\ computed as a di}erence between potentials evoked by target and non!target stimuli[ Standard deviations are shown in parentheses beneath the mean values[ P299
P199
Amplitude ðmVŁ
Latency ðmsŁ
Amplitude ðmVŁ
Latency ðmsŁ
Positive context
08[5 "09[5#
270[4 "16[4#
8[1 "2[1#
199[5 "05[2#
Negative context
04[8 "7[1#
275[2 "23[8#
8[8 "1[7#
193[9 "05[5#
No!feedback recordings
11[0 "8[6#
274[1 "07[9#
09[9 "4[1#
190[2 "05[2#
pared\ only one point of signi_cant di}erence was found in the position of the second diamond in Pz recording[ All the di}erences between the positive context curves and the control no!feedback recordings failed to reach signi_cance[
3[ Discussion The present results demonstrate that being unsuc! cessful in the test "negative context# reduced the ampli! tude of the P299 potential[ Being successful "positive context# did not result in signi_cant di}erences[ It is unlikely that the enlargement caused by positive context\ only failed to reach statistical signi_cance\ because the P299 amplitudes were slightly reduced rather than enlarged in these trials[ Cognitive variables that are known to a}ect P299 amplitude\ such as target stimulus probability or stimulus repetition rate ð01\ 04Ł\ were constant in the experiment[ The variables that could a}ect this amplitude were] the e}ect of the parallel task\ arousal level and emotion[ It was shown that the P299 amplitude was reduced by any parallel activity that competed for brain processing resources with the main experimental task ð5\ 6\ 02\ 03\ 08\ 10\ 14Ł[ Interviews with the subjects taking part in this experiment and the other experiments with real feedback indicated that the process of using feedback information included activities such as] the analysis of the internal states of the brain during the last repetition of the target stimulus\ the attempts to correlate them with the feedback information\ the elaboration of the decision about what to do next\ repetitive veri_cation and modi_cation of this decision etc[ These are all highly complex cognitive processes that most likely recruit a large amount of pro! cessing resources and thereby limit the resources available for the main task of mental counting of target stimuli[ Thus\ the general suppression of P299 amplitude could be expected[ The lack of suppression in positive context trials indicates the presence of other\ superimposed pro! cesses[ The enhancement of P299 amplitude by both positive and negative emotions has been reported in a number of other studies ð0\ 1\ 05Ð07\ 19\ 12\ 13\ 16Ð18Ł[ In all these experiments emotions were triggered directly by target stimuli and there were no parallel activities[ The expla! nation of the present results can be based on the assump! tion that the enhancement\ produced by positive emotion\ was strong enough to compensate for the e}ect of parallel task but the enhancement produced by negative emotion was signi_cantly weaker[ Is it likely that opposite emotions change the P299 amplitude in the same direc! tion but with di}erent e}ectiveness< Chapman and co! authors found that positive words elicited larger ERP amplitudes than negative words for components in the 199 msÐ319 ms latency range ð2\ 3Ł[ More recently\ Pal!
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
omba and co!authors ð13Ł noticed that pleasant stimuli evoked slightly higher cortical positivity than unpleasant stimuli[ The di}erence did not reach the level of statistical signi_cance but the authors mentioned the e}ect as {evi! dent| ð13Ł[ ERPs associated with hitting the target with a photoelectric gun were higher than those recorded when the target was missed ð21Ł[ The di}erence was statistically signi_cant and was found within the P299 latency range but the experimental design di}ered from the typical P299 paradigm[ In all these studies\ the di}erences between positive and negative emotions were smaller than pre! sently reported[ Thus\ the acceptance of the hypothesis of the interacting e}ects of parallel task and emotion leads to the conclusion that manipulated feedback is e}ective in producing di}erent emotional states[ It seems reasonable to assume that feedback experi! ments\ in which subjects attempt to control the amplitude of their own brain potential\ produce higher arousal levels than the boring and aimless task of counting target stim! uli[ Available experimental data suggest that arousal\ triggered by both pleasant and unpleasant stimuli\ increases the P299 amplitude\ whereas relaxing stimuli seem to reduce it ð1\ 05\ 06\ 11\ 12\ 15Ł[ Arousal level probably in~uenced the results of the present experiment but it does not readily explain the di}erence between positive and negative context[ It seems unlikely that try! ing to achieve a good result and failing produces less arousal than being successful from the beginning[ That is\ unless the subject gives up and loses interest[ But can such reactions occur within the time scale of thirty repetitions of target stimuli< It seemed logical that experiments with {neutral| ran! dom feedback "49) {small| and 49) {big| feedback sig! nals# would provide the valuable reference[ Preliminary experiments in this laboratory have demonstrated that the amplitudes of P299 potentials\ recorded in such experiments\ did not di}er signi_cantly from amplitudes recorded in no!feedback conditions[ The reports of the subjects indicated that such {neutral| feedback sessions could be interpreted as both success or failure depending on the previous achievements[ If there were no previous experiences\ the feeling of success or failure was probably highly in~uenced by the personality of the subject or the current mood state[ Careful questioning and an increase in the number of feedback trials nearly uncovered the e}ect of the feedback manipulation[ Several subjects were lost to the experiment and the {neutral feedback| trials were abandoned[ Random feedback procedures are often used to pro! vide a reference condition in the real feedback experi! ments ð20Ł[ The present results indicate that such reference procedures should be treated with caution[ Random feedback processing may involve complex pro! cesses and produce unexpected results[ No!feedback ref! erence can also introduce a di}erence due to the removal of the parallel task e}ect[ Perhaps the best strategy is to
308
use both controls\ but even than the interpretation of the results can be complicated[ 4[ Conclusion Manipulated feedback is an interesting method of pro! ducing both positive and negative emotional states[ The analysis of the responses to neutral stimuli in di}erent\ tonic emotional states directly addresses every day con! ditions[ Tonic emotional states a}ect the P299 potential in a way that cannot be predicted from the responses to emotional stimuli[ The e}ects of positive and negative emotional states may not be complementary[ This asym! metry can be explained by the interference of other pro! cesses\ such as the drain on central processing resources\ but even then the additional assumption must be made that the e}ects of positive and negative emotions di}er in strength[ Some indirect indications of such di}erences can be found in the published literature[ Any attempt to use the present data for the interpret! ation of the real feedback e}ects would probably be premature[ But at least they indicate how complicated processes can be correlated with processing the feedback information[ Acknowledgement This research was supported by statutory grant of the Nencki Institute[
References ð0Ł Carretie L\ Inglesias J[ ERP in response to facial expressions of emotions[ Psychophysiology 0880^17]S05 "Abstract#[ ð1Ł Carretie L\ Inglesias J[ An ERP study on the speci_city of facial expression processing[ International Journal of Psychophysiology 0884^08]072Ð81[ ð2Ł Chapman RM[ Connotative meaning and averaged evoked poten! tials[ In] Begleiter H\ editor[ Evoked brain potentials and behav! iour[ New York] Plenum\ 0868]060Ð85[ ð3Ł Chapman RM\ McCrary JW\ Chapman JA\ Martin JK[ Behavioral and neural analyses of connotative meaning] word classes and rating scales[ Brain Language 0879^00]208Ð28[ ð4Ł Diamond SJ\ Farrington L[ Emotional response to _lms shown to the right and left hemisphere of the brain measured by heart rate[ Acta Psychologica 0867^30]144Ð59[ ð5Ł Donchin E\ Karis D\ Bashore TR\ Coles MGH\ Gratton G[ Cog! nitive psychophysiology and human information processing[ In] Coles MGH\ Donchin E\ Porges SW\ editors[ Psychophysiology] systems\ processes and applications[ New York] Guilford Press\ 0875]133Ð56 ð6Ł Donchin E\ Kramer A\ Wickens C[ Applications of brain event! related potentials to problems in engineering psychology[ In] Coles MGH\ Donchin E\ Porges SW\ editors[ Psychophysiology] systems\ processes and applications[ New York] Guilford Press\ 0875]691Ð07[ ð7Ł Eibl!Eibesfeldt I[ Human Ethology[ New York] Aldine de Gruyter\ 0878[
319
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
ð8Ł Gainotti G[ Emotional behavior and hemispheric side lesions[ Cortex 0861^7]30Ð4 ð09Ł Gardner H[ The shattered mind[ New York] Knopf\ 0864[ ð00Ł Goldstein K[ The Organism[ New York] American Book\ 0828[ ð01Ł Hillyard SA\ Picton TW[ Event!related brain potentials and selec! tive information processing in man[ In] Desmedt JE\ editor[ Cer! ebral evoked potentials in man[ Basel] Karger\ 0868]60Ð82[ ð02Ł Israel JB\ Chesney GL\ Wickens CD\ Donchin E[ P299 tracking di.culty] Evidence for multiple resources in dual!task perform! ance[ Psychophysiology 0879^06]148Ð62[ ð03Ł Israel JB\ Wickens CD\ Chesney GL\ Donchin E[ The event!related potential as an index of display monitoring workload[ Human Factors 0879^11]101Ð13[ ð04Ł Johnson R Jr[ The amplitude of the P299 component of the event! related potential] review and synthesis[ In] Ackles PK\ Jennings RR\ Coles MGH\ editors[ Advances in psychophysiology\ vol 2\ London] JAI Press\ 0876]58Ð026 ð05Ł Johnston VS\ Burleson M\ Miller DR[ Emotional value and late positive components of ERPs[ In] Johnston R Jr\ Rohrbaugh JW\ Parasuraman R\ editors[ Current trends in event!related potentials research[ Amsterdam] Elsevier\ Electroencephalography and Clini! cal Neurophysiology\ 0876^S39]078Ð192[ ð06Ł Johnston VS\ Miller DR\ Burleson MH[ Multiple P2s to emotional stimuli and their theoretical signi_cance[ Psychophysiology 0875^12]573Ð83[ ð07Ł Johnston VS\ Wang XT[ The relationship between menstrual phase and the P2 component of ERPs[ Psychophysiology 0880^17]399Ð 8[ ð08Ł Kramer A\ Sirevaag E\ Braune R[ A psychophysiological assess! ment of operator workload during simulated ~ight missions[ Human Factors 0876^18]034Ð59[ ð19Ł Laurian S\ Bader M\ Lanares J\ Oros L[ Topography of event! related potentials elicited by visual emotional stimuli[ International Journal of Psychophysiology 0880^09]120Ð7[ ð10Ł Michalski A[ The e}ect of tonic pain on processing the non!painful stimuli indexed by late components of event!related potentials[ Acta Neurobiologiae Experimentalis 0887^47]44Ð53[
ð11Ł Miltner W\ Johnson R Jr\ Braun C\ Larbig W[ Somatosensory event!related potentials to painful and non!painful stimuli] e}ects of attention[ Pain 0878^27]292Ð01[ ð12Ł Palomba D\ Angrilli A\ Bravi S[ Event!related potentials and heart rate changes to emotional pictorial stimuli[ Psychophysiology 0882^29]S38 "Abstract#[ ð13Ł Palomba D\ Angrilli A\ Mini A[ Visual evoked potentials\ heart rate responses and memory to emotional pictorial stimuli[ Inter! national Journal of Psychophysiology 0886^16]44Ð56[ ð14Ł Polich J[ P299 from a passive auditory paradigm[ Electro! encephalography and Clinical Neurophysiology 0878^62]201Ð19[ ð15Ł Polich J\ Kok A[ Cognitive and biological determinants of P299] an integrative review[ Biological Psychology 0884^30]092Ð35[ ð16Ł Radilova J\ Figar S\ Radil T[ Sexual arousal and visual perception[ Activitas Nervosa Superior 0872^14]057Ð69[ ð17Ł Radilova J\ Figar S\ Radil T[ Emotional states in~uence the visual evoked potentials[ Activitas Nervosa Superior 0873^15]048Ð59[ ð18Ł Radilova J[ P299 and the emotional states studied by psycho! physiological methods[ International Journal of Psychophysiology 0878^6]253Ð55[ ð29Ł Sackeim H\ Greenberg M\ Weiman A\ Gur R\ Hungerbuhler J\ Geschwind N[ Functional brain asymmetry in the expression of positive and negative emotions] lateralization of insult in cases of uncontrollable emotional outburst[ Archieves of Neurology 0871^08]109Ð7[ ð20Ł Sommer W\ Schweinberger S[ Operant conditioning of P299[ Bio! logical Psychology 0881^22]26Ð38[ ð21Ł Sobotka S\ Grabowska A\ Grodzicka J\ Wasilewski R\ Budohoska W[ Hemispheric asymmetry in event related potentials associated with positive and negative emotions[ Acta Neurobiologiae Exper! imentalis 0881^41]140Ð59[ ð22Ł Terzian H[ Behavioral and EEG e}ects of intracarotid sodium amytal injections[ Acta Neurochirurgica 0853^01]129Ð8[ ð23Ł Vanderploeg RD\ Brown WS\ Marsh JT[ Judgments of emotion in words and faces] ERP correlates[ International Journal of Psychophysiology 0876^3]082Ð194[
Neuropsychologia 26 "0888# 302Ð319
The e}ect of accomplishment and failure on P299 potentials evoked by neutral stimuli Andrzej Michalski Department of Neurophysiology\ Nencki Institute of Experimental Biology\ 2 Pasteur St\ 91!982 Warsaw\ Poland Received 04 September 0886^ accepted 06 March 0887
Abstract The e}ects of emotional states of {being successful| vs[ {being unsuccessful| were studied by measuring the P299 component of event!related potentials "ERPs#[ Nine subjects were instructed to reduce their P299 amplitude using feedback[ Feedback was random but the relative probability of di}erent signals created the situations of {being successful| or {being unsuccessful|[ The probability of {small| feedback was 9[6 in {successful| and 9[04 in {unsuccessful| trials[ ERPs recorded without the feedback were used as a reference[ Potentials\ evoked by light stimuli in a standard {odd!ball| procedure\ were recorded from Fz\ Cz and Pz scalp sites[ The amplitudes of P299 components were reduced in {unsuccessful| trials whereas in {successful| trials they did not di}er signi_cantly from responses recorded without the feedback[ There were no signi_cant di}erences in peak latencies[ These _ndings indicate that tonic emotional states a}ect the processing of neutral stimuli and that late components of ERPs can be useful indices in the analysis of these alterations[ The results also indicate that the e}ects of positive and negative emotional states are not always reciprocal[ Manipulated feedback is suggested as an useful model in the studies of emotions[ Data can also facilitate the interpretation of the real feedback e}ects[ Þ 0888 Elsevier Science Ltd[ All rights reserved[ Keywords] P299^ Success^ Failure
0[ Introduction There have been numerous attempts to study the e}ect of emotions on the P299 potential[ In a number of experi! ments facial expressions have been used\ since these are stimuli considered to be phylogenetically specialized for emotional communication ð7Ł[ The results were con! tradictory[ In some studies the neutral faces evoked lower amplitudes than emotional ones ð0\ 1\ 19Ł^ in others they evoked the highest responses ð23Ł[ Testing stimuli other than faces\ Radilova and co!workers ð16Ð18Ł showed that erotic stimuli produced more positive P299 waves than non!erotic pictures[ Johnston and co!workers ð05Ð07Ł showed that both pictures of nude models of opposite sex and pictures of dermatological diseases evoked grater P299 amplitudes than neutral pictures[ This _nding was con_rmed by Palomba and co!authors ð12\ 13Ł[ Apart from the di}erences in experimental methods\ one element of the procedure was common in all these studies] the same stimulus that evoked P299 potential also trig! gered the emotional response[
Corresponding author[ Tel[] ¦37 11 548 74 60^ fax] ¦37 11 11 42 31^ e!mail] michanÝnencki[gov[pl
To resolve the discrepancy between the results of experiments on hemispheric localization of emotions\ Sobotka and co!authors ð21Ł suggested di}erent mech! anisms for the brief reactions to emotional stimuli and the more prolonged emotional states that a}ect processing of the neutral stimuli[ The analysis of available data led these authors\ in agreement with the hypothesis originally developed by Gardner ð09Ł\ to conclude that patterns of brain functions associated with judgments of the emotional content of stimuli\ such as facial expressions\ indicated the dominance of right hemisphere for both positive and negative emotions[ On the other hand\ the charged emotional states associated with brain damage ð8\ 29Ł\ injections of sodium amytal ð22Ł or watching emotional _lms ð4Ł have been reported to produce pat! terns of activation that support an alternative hypothesis of Goldstein ð00Ł\ that the left hemisphere is dominant for positive emotions and right hemisphere for negative emotions[ Such di}erences suggest that event!related potentials\ especially the late components\ may also be di}erentially a}ected by prolonged emotional states and by brief responses to the emotional component of evok! ing stimuli[ The aim of the present experiment was to analyze the e}ects of tonic emotional states "being successful vs being
9917Ð2821:88:,*see front matter Þ 0888 Elsevier Science Ltd[ All rights reserved PII] S 9 9 1 7 Ð 2 8 2 1 " 8 7 # 9 9 0 9 5 Ð 6
303
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
unsuccessful# on P299 component of the ERPs evoked by non!emotional stimuli "red and yellow ~ashes of light#[ Since random feedback procedures were used to manipu! late the emotional states\ the results will also be relevant to the interpretation of the real feedback e}ects ð20Ł[ 1[ Methods 1[0[ Subjects Volunteers were informed that they were being asked to participate in standard feedback experiments and their consent was obtained[ Before the experiment\ each par! ticipant was asked to evaluate his:her level of anxiety and mood state using a ten!point scale[ The {normal| level of anxiety and the {neutral| mood corresponded to value 4 on the scale[ Data were collected from nine volunteers of both sexes "seven females and two males#\ aged 18Ð37 years[ They were all familiar with the laboratory and they all reported their anxiety level and mood state to range between 3 and 5 on the scale[ Two additional participants "at the end of the series# expressed their concern that {the experiment was strange|\ and they were excluded from the experiment[ 1[1[ Recording EEG signals were recorded with disc electrodes glued at Fz\ Cz and Pz positions\ referenced to linked mastoids and supplemented by vertical and horizontal EOG[ Sig! nals were sampled with 1937 Hz frequency\ 01 bit resol! ution\ digitally _ltered 9[05Ð29 Hz and reduced to 145 Hz by averaging the adjacent points "Elmiko Paperless EEG system#[ Data were stored in epochs containing 149 ms before and 0 s after the stimulus onset[ Epochs were rejected by computer program if EOG amplitude exceeded 39 mV[ Rejected epochs were replaced with the new ones[ Flashes of spatially overlapped arrays of red and yel! low LED diodes "1> 1> of the visual angle\ 09 cd:m1 luminosity on 0 cd:m1 background\ 099 ms duration# were used as stimuli in a standard odd!ball procedure "stimulus rate*1[4 s^ target probability*9[01^ minimum number of good target stimulus repetitions*29#[ Sub! jects were asked to mentally count the target stimuli and report the number seen at the end of the recording[ Data were rejected if the error in counting was bigger than 19)[ 1[2[ Procedures To reduce the number of variables\ the yellow LED diode was always used as a target and the task instruction was always to reduce the amplitude of P299 potentials "downtraining#[ Downtraining was selected because it
had produced bigger e}ects in an earlier experimental attempt to use the real feedback paradigm "unpublished data from the same laboratory#[ Subjects were asked to reduce the P299 amplitude using feedback provided by the spoken words {small|\ {medium| or {large|\ pro! nounced by the experimenter 0 s after the onset of each target stimulus[ The words were taken from tables which ensured their random distribution\ regardless of the amplitude of real potentials[ In the {successful| trials the probability of {small| was 9[6\ {medium|\ 9[04 and {large|\ 9[04[ In the {unsuccessful|\ trials the probability of {small| was 9[04\ {medium|\ 9[04 and {large|\ 9[6[ The exper! imental session always contained a sequence of _ve rec! ordings] 0[ recording without the feedback\ 1[ randomly selected {successful| or {unsuccessful| feed! back recording\ 2[ recording without the feedback\ 3[ complementary {successful| or {unsuccessful| feedback recording\ 4[ recording without the feedback[ 1[3[ Data analysis Epochs of EEG were averaged\ stored and printed using Elmiko software package[ Di}erence curves were computed for each pair of target and non!target ERPs[ Grand!averaged waveforms were computed by averaging data for all subjects in each experimental condition[ Stat! istical analysis followed two basic strategies] 0[ Amplitudes and latencies of peaks were measured using cursors on the computer screen and the data were transferred to SYSTAT program for multi!factor analysis of variance "ANOVA#[ 1[ Grand!averaged curves\ computed for each exper! imental condition\ were compared[ Within the regions of visible di}erences the corresponding points of grand!averaged curves were expanded into sets of paired data "points of individual di}erence curves# and compared with Student{s t!test[ Direct comparison of the curves avoided the arbitrary peak identi_cation\ on the other hand\ it did not allow the {latency cor! rections| "the scatter of peak latencies could a}ect the grand averaged waveshapes#[ 2[ Results The responses to target stimuli\ grand!averaged over all subjects\ are shown in Fig[ 0[ Three sets of curves represent the data recorded from Fz "top#\ Cz "middle# and Pz "bottom# electrodes[ Dotted lines show the wav! eforms obtained without the feedback[ Dashed lines show the potentials recorded with the dummy feedback that contained 69) of {small| messages[ Since the instruction
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
304
Fig[ 0[ Grand!averaged ERPs evoked by target stimuli[ Potentials were recorded with Fz "top#\ Cz "middle# and Pz "bottom# electrodes[ Dotted lines show the ERPs recorded without the feedback[ Dashed lines show the ERPs recorded with the feedback and the instruction to reduce the P299 amplitude[ Feedback was random but it contained 69) of {small| signals[ The subjects were {successful|[ Solid lines show the ERPs recorded with the feedback and the instruction to reduce the P299 amplitude[ Feedback was random but contained only 04) of {small| signals[ The subjects were {unsuccessful|[
was to reduce the P299 amplitude\ the subject was {suc! cessful|[ This situation will be referred to as a {positive context|[ Solid lines show the recordings obtained with the dummy feedback that contained 04) of {small| mess! ages[ Since the instruction was still to reduce the P299
amplitude\ the subject was {unsuccessful|[ This situation will be referred to as a {negative context|[ In addition to the early waves\ all the recordings show the well developed late components within the P199 and P299 regions[ It is clear that the experimental procedures
305
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
a}ected the P299 waves much more than the earlier com! ponents[ Both types of feedback reduced the P299 ampli! tude but in the case of negative context the reduction was greater[ The P299 peak was de_ned as the positive value of greatest magnitude between 149 msÐ399 ms following the stimulus onset\ and the P199 peak was de_ned as the largest positive value recorded between 049 msÐ189 ms after the stimulus onset[ Amplitudes\ measured in indi! vidual ERPs relative to 149 ms pre!stimulus baseline\ were analyzed with two way ANOVA "electrode context#[ Context variable included the positive context\ negative context and no feedback condition[ Di}erences
in both P299 and P199 peak amplitudes due to the context type were insigni_cant[ Latencies of both peaks\ analyzed with the identical two way ANOVA\ also did not show any signi_cant e}ects of the context[ The grand!averaged potentials evoked by non!target stimuli are illustrated in Fig[ 1 "_gure conventions are the same as in Fig[ 0#[ Besides the early components\ recordings from all electrodes showed well!developed P199 waves and only greatly reduced P299 components[ The comparison of Figs 1 and 0 indicates that di}erences\ due to the type of the context\ were much greater in the target responses[ In the individual non!target recordings\ the very small
Fig[ 1[ Grand!averaged ERPs evoked by non!target stimuli[ Figure conventions are the same as in Fig[ 0[
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
size of the P299 components made it impossible to ident! ify and measure this component reliably[ Thus\ only P199 waves were analyzed[ Two way ANOVA "electrode context# showed that context type clearly did not produce signi_cant di}erences\ either in amplitudes or in latencies of P199 waves in non!target responses[ However\ in spite of their statistical insigni_cance\ the di}erences associated with the negative context in target
306
responses looked promising[ The concentration of di}erences within the P299 region indicated that they were not entirely due to the random scatter[ In an attempt to remove some of the sources of uncontrolled amplitude ~uctuations\ di}erence curves were computed by sub! tracting ERPs evoked by non!target stimuli from those evoked by target stimuli[ These di}erence curves\ grand! averaged for all subjects\ are shown in Fig[ 2 "_gure
Fig[ 2[ Grand!averaged di}erence curves computed by subtracting ERPs evoked by non!target stimuli from ERPs evoked by target stimuli[ Figure conventions are the same as in Fig[ 0[ Black diamonds indicate the points of signi_cant di}erences between the curves recorded in {unsuccessful| trials "solid lines# and the control\ no!feedback trials "dotted lines#[
307
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
conventions are the same as in Figs 0 and 1#[ All di}erence curves show the distinct peaks of P199 and P299 com! ponents[ P199 amplitudes are clearly smaller in the di}erence curves than in target responses\ whereas P299 amplitudes are almost the same[ It is apparent that\ like in target ERPs\ dummy feedback a}ects predominantly the P299 component[ The asymmetry between the e}ects of positive and negative context seems to be even more pronounced in the di}erence curves] within the P299 region the dashed lines "positive context# and the dotted lines "no feedback# are closer to each other than in target ERPs "Fig[ 0#[ In contrast\ the solid lines "negative con! text# show the marked di}erences[ Within the P199 region the di}erence curves show the e}ect that could not be seen in target ERPs] both types of dummy feedback "solid and dashed lines# reduce this component with respect to the no feedback condition "dotted lines#[ The amplitudes and latencies of P299 and P199 com! ponents\ measured in the di}erence waves\ are presented in Table 0[ A two!way ANOVA "electrode context# showed a signi_cant e}ect of the type of context on the P299 amplitude "df 1\ F 2[00\ P ³ 9[94#[ Post!hoc tests showed that the amplitudes of P299 waves recorded with negative context were signi_cantly smaller than the amplitudes recorded without the feedback "Bonferroni\ P ³ 9[933#[ Neither the di}erences in P299 peak latencies nor the di}erences in P199 amplitudes and latencies were signi_cant[ Black diamonds beneath the curves in Fig[ 2 indicate the signi_cant di}erences between the corresponding points of negative context and no!feedback curves[ At every 13!ms\ within the range of 091 msÐ475 ms\ the corresponding points of grand!averaged curves were expanded into sets of paired data "points of individual di}erence curves# and compared with the Student{s t!test[ All diamonds are clearly located within the P299 region[ When the positive and negative context curves were com!
Table 0 The e}ect of positive and negative context on the parameters of P299 and P199 waves\ computed as a di}erence between potentials evoked by target and non!target stimuli[ Standard deviations are shown in parentheses beneath the mean values[ P299
P199
Amplitude ðmVŁ
Latency ðmsŁ
Amplitude ðmVŁ
Latency ðmsŁ
Positive context
08[5 "09[5#
270[4 "16[4#
8[1 "2[1#
199[5 "05[2#
Negative context
04[8 "7[1#
275[2 "23[8#
8[8 "1[7#
193[9 "05[5#
No!feedback recordings
11[0 "8[6#
274[1 "07[9#
09[9 "4[1#
190[2 "05[2#
pared\ only one point of signi_cant di}erence was found in the position of the second diamond in Pz recording[ All the di}erences between the positive context curves and the control no!feedback recordings failed to reach signi_cance[
3[ Discussion The present results demonstrate that being unsuc! cessful in the test "negative context# reduced the ampli! tude of the P299 potential[ Being successful "positive context# did not result in signi_cant di}erences[ It is unlikely that the enlargement caused by positive context\ only failed to reach statistical signi_cance\ because the P299 amplitudes were slightly reduced rather than enlarged in these trials[ Cognitive variables that are known to a}ect P299 amplitude\ such as target stimulus probability or stimulus repetition rate ð01\ 04Ł\ were constant in the experiment[ The variables that could a}ect this amplitude were] the e}ect of the parallel task\ arousal level and emotion[ It was shown that the P299 amplitude was reduced by any parallel activity that competed for brain processing resources with the main experimental task ð5\ 6\ 02\ 03\ 08\ 10\ 14Ł[ Interviews with the subjects taking part in this experiment and the other experiments with real feedback indicated that the process of using feedback information included activities such as] the analysis of the internal states of the brain during the last repetition of the target stimulus\ the attempts to correlate them with the feedback information\ the elaboration of the decision about what to do next\ repetitive veri_cation and modi_cation of this decision etc[ These are all highly complex cognitive processes that most likely recruit a large amount of pro! cessing resources and thereby limit the resources available for the main task of mental counting of target stimuli[ Thus\ the general suppression of P299 amplitude could be expected[ The lack of suppression in positive context trials indicates the presence of other\ superimposed pro! cesses[ The enhancement of P299 amplitude by both positive and negative emotions has been reported in a number of other studies ð0\ 1\ 05Ð07\ 19\ 12\ 13\ 16Ð18Ł[ In all these experiments emotions were triggered directly by target stimuli and there were no parallel activities[ The expla! nation of the present results can be based on the assump! tion that the enhancement\ produced by positive emotion\ was strong enough to compensate for the e}ect of parallel task but the enhancement produced by negative emotion was signi_cantly weaker[ Is it likely that opposite emotions change the P299 amplitude in the same direc! tion but with di}erent e}ectiveness< Chapman and co! authors found that positive words elicited larger ERP amplitudes than negative words for components in the 199 msÐ319 ms latency range ð2\ 3Ł[ More recently\ Pal!
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
omba and co!authors ð13Ł noticed that pleasant stimuli evoked slightly higher cortical positivity than unpleasant stimuli[ The di}erence did not reach the level of statistical signi_cance but the authors mentioned the e}ect as {evi! dent| ð13Ł[ ERPs associated with hitting the target with a photoelectric gun were higher than those recorded when the target was missed ð21Ł[ The di}erence was statistically signi_cant and was found within the P299 latency range but the experimental design di}ered from the typical P299 paradigm[ In all these studies\ the di}erences between positive and negative emotions were smaller than pre! sently reported[ Thus\ the acceptance of the hypothesis of the interacting e}ects of parallel task and emotion leads to the conclusion that manipulated feedback is e}ective in producing di}erent emotional states[ It seems reasonable to assume that feedback experi! ments\ in which subjects attempt to control the amplitude of their own brain potential\ produce higher arousal levels than the boring and aimless task of counting target stim! uli[ Available experimental data suggest that arousal\ triggered by both pleasant and unpleasant stimuli\ increases the P299 amplitude\ whereas relaxing stimuli seem to reduce it ð1\ 05\ 06\ 11\ 12\ 15Ł[ Arousal level probably in~uenced the results of the present experiment but it does not readily explain the di}erence between positive and negative context[ It seems unlikely that try! ing to achieve a good result and failing produces less arousal than being successful from the beginning[ That is\ unless the subject gives up and loses interest[ But can such reactions occur within the time scale of thirty repetitions of target stimuli< It seemed logical that experiments with {neutral| ran! dom feedback "49) {small| and 49) {big| feedback sig! nals# would provide the valuable reference[ Preliminary experiments in this laboratory have demonstrated that the amplitudes of P299 potentials\ recorded in such experiments\ did not di}er signi_cantly from amplitudes recorded in no!feedback conditions[ The reports of the subjects indicated that such {neutral| feedback sessions could be interpreted as both success or failure depending on the previous achievements[ If there were no previous experiences\ the feeling of success or failure was probably highly in~uenced by the personality of the subject or the current mood state[ Careful questioning and an increase in the number of feedback trials nearly uncovered the e}ect of the feedback manipulation[ Several subjects were lost to the experiment and the {neutral feedback| trials were abandoned[ Random feedback procedures are often used to pro! vide a reference condition in the real feedback experi! ments ð20Ł[ The present results indicate that such reference procedures should be treated with caution[ Random feedback processing may involve complex pro! cesses and produce unexpected results[ No!feedback ref! erence can also introduce a di}erence due to the removal of the parallel task e}ect[ Perhaps the best strategy is to
308
use both controls\ but even than the interpretation of the results can be complicated[ 4[ Conclusion Manipulated feedback is an interesting method of pro! ducing both positive and negative emotional states[ The analysis of the responses to neutral stimuli in di}erent\ tonic emotional states directly addresses every day con! ditions[ Tonic emotional states a}ect the P299 potential in a way that cannot be predicted from the responses to emotional stimuli[ The e}ects of positive and negative emotional states may not be complementary[ This asym! metry can be explained by the interference of other pro! cesses\ such as the drain on central processing resources\ but even then the additional assumption must be made that the e}ects of positive and negative emotions di}er in strength[ Some indirect indications of such di}erences can be found in the published literature[ Any attempt to use the present data for the interpret! ation of the real feedback e}ects would probably be premature[ But at least they indicate how complicated processes can be correlated with processing the feedback information[ Acknowledgement This research was supported by statutory grant of the Nencki Institute[
References ð0Ł Carretie L\ Inglesias J[ ERP in response to facial expressions of emotions[ Psychophysiology 0880^17]S05 "Abstract#[ ð1Ł Carretie L\ Inglesias J[ An ERP study on the speci_city of facial expression processing[ International Journal of Psychophysiology 0884^08]072Ð81[ ð2Ł Chapman RM[ Connotative meaning and averaged evoked poten! tials[ In] Begleiter H\ editor[ Evoked brain potentials and behav! iour[ New York] Plenum\ 0868]060Ð85[ ð3Ł Chapman RM\ McCrary JW\ Chapman JA\ Martin JK[ Behavioral and neural analyses of connotative meaning] word classes and rating scales[ Brain Language 0879^00]208Ð28[ ð4Ł Diamond SJ\ Farrington L[ Emotional response to _lms shown to the right and left hemisphere of the brain measured by heart rate[ Acta Psychologica 0867^30]144Ð59[ ð5Ł Donchin E\ Karis D\ Bashore TR\ Coles MGH\ Gratton G[ Cog! nitive psychophysiology and human information processing[ In] Coles MGH\ Donchin E\ Porges SW\ editors[ Psychophysiology] systems\ processes and applications[ New York] Guilford Press\ 0875]133Ð56 ð6Ł Donchin E\ Kramer A\ Wickens C[ Applications of brain event! related potentials to problems in engineering psychology[ In] Coles MGH\ Donchin E\ Porges SW\ editors[ Psychophysiology] systems\ processes and applications[ New York] Guilford Press\ 0875]691Ð07[ ð7Ł Eibl!Eibesfeldt I[ Human Ethology[ New York] Aldine de Gruyter\ 0878[
319
A[ Michalski:Neuropsycholo`ia 26 "0888# 302Ð319
ð8Ł Gainotti G[ Emotional behavior and hemispheric side lesions[ Cortex 0861^7]30Ð4 ð09Ł Gardner H[ The shattered mind[ New York] Knopf\ 0864[ ð00Ł Goldstein K[ The Organism[ New York] American Book\ 0828[ ð01Ł Hillyard SA\ Picton TW[ Event!related brain potentials and selec! tive information processing in man[ In] Desmedt JE\ editor[ Cer! ebral evoked potentials in man[ Basel] Karger\ 0868]60Ð82[ ð02Ł Israel JB\ Chesney GL\ Wickens CD\ Donchin E[ P299 tracking di.culty] Evidence for multiple resources in dual!task perform! ance[ Psychophysiology 0879^06]148Ð62[ ð03Ł Israel JB\ Wickens CD\ Chesney GL\ Donchin E[ The event!related potential as an index of display monitoring workload[ Human Factors 0879^11]101Ð13[ ð04Ł Johnson R Jr[ The amplitude of the P299 component of the event! related potential] review and synthesis[ In] Ackles PK\ Jennings RR\ Coles MGH\ editors[ Advances in psychophysiology\ vol 2\ London] JAI Press\ 0876]58Ð026 ð05Ł Johnston VS\ Burleson M\ Miller DR[ Emotional value and late positive components of ERPs[ In] Johnston R Jr\ Rohrbaugh JW\ Parasuraman R\ editors[ Current trends in event!related potentials research[ Amsterdam] Elsevier\ Electroencephalography and Clini! cal Neurophysiology\ 0876^S39]078Ð192[ ð06Ł Johnston VS\ Miller DR\ Burleson MH[ Multiple P2s to emotional stimuli and their theoretical signi_cance[ Psychophysiology 0875^12]573Ð83[ ð07Ł Johnston VS\ Wang XT[ The relationship between menstrual phase and the P2 component of ERPs[ Psychophysiology 0880^17]399Ð 8[ ð08Ł Kramer A\ Sirevaag E\ Braune R[ A psychophysiological assess! ment of operator workload during simulated ~ight missions[ Human Factors 0876^18]034Ð59[ ð19Ł Laurian S\ Bader M\ Lanares J\ Oros L[ Topography of event! related potentials elicited by visual emotional stimuli[ International Journal of Psychophysiology 0880^09]120Ð7[ ð10Ł Michalski A[ The e}ect of tonic pain on processing the non!painful stimuli indexed by late components of event!related potentials[ Acta Neurobiologiae Experimentalis 0887^47]44Ð53[
ð11Ł Miltner W\ Johnson R Jr\ Braun C\ Larbig W[ Somatosensory event!related potentials to painful and non!painful stimuli] e}ects of attention[ Pain 0878^27]292Ð01[ ð12Ł Palomba D\ Angrilli A\ Bravi S[ Event!related potentials and heart rate changes to emotional pictorial stimuli[ Psychophysiology 0882^29]S38 "Abstract#[ ð13Ł Palomba D\ Angrilli A\ Mini A[ Visual evoked potentials\ heart rate responses and memory to emotional pictorial stimuli[ Inter! national Journal of Psychophysiology 0886^16]44Ð56[ ð14Ł Polich J[ P299 from a passive auditory paradigm[ Electro! encephalography and Clinical Neurophysiology 0878^62]201Ð19[ ð15Ł Polich J\ Kok A[ Cognitive and biological determinants of P299] an integrative review[ Biological Psychology 0884^30]092Ð35[ ð16Ł Radilova J\ Figar S\ Radil T[ Sexual arousal and visual perception[ Activitas Nervosa Superior 0872^14]057Ð69[ ð17Ł Radilova J\ Figar S\ Radil T[ Emotional states in~uence the visual evoked potentials[ Activitas Nervosa Superior 0873^15]048Ð59[ ð18Ł Radilova J[ P299 and the emotional states studied by psycho! physiological methods[ International Journal of Psychophysiology 0878^6]253Ð55[ ð29Ł Sackeim H\ Greenberg M\ Weiman A\ Gur R\ Hungerbuhler J\ Geschwind N[ Functional brain asymmetry in the expression of positive and negative emotions] lateralization of insult in cases of uncontrollable emotional outburst[ Archieves of Neurology 0871^08]109Ð7[ ð20Ł Sommer W\ Schweinberger S[ Operant conditioning of P299[ Bio! logical Psychology 0881^22]26Ð38[ ð21Ł Sobotka S\ Grabowska A\ Grodzicka J\ Wasilewski R\ Budohoska W[ Hemispheric asymmetry in event related potentials associated with positive and negative emotions[ Acta Neurobiologiae Exper! imentalis 0881^41]140Ð59[ ð22Ł Terzian H[ Behavioral and EEG e}ects of intracarotid sodium amytal injections[ Acta Neurochirurgica 0853^01]129Ð8[ ð23Ł Vanderploeg RD\ Brown WS\ Marsh JT[ Judgments of emotion in words and faces] ERP correlates[ International Journal of Psychophysiology 0876^3]082Ð194[