emotional arousal and gender

13 Pages·2016·0.87 MB·English

by  Vera Ferrari

Cognition and Emotion
ISSN: 0269-9931 (Print) 1464-0600 (Online) Journal homepage: http://www.tandfonline.com/loi/pcem20
Evaluative ratings and attention across the life
span: emotional arousal and gender
Vera Ferrari, Nicola Bruno, Rabih Chattat & Maurizio Codispoti
To cite this article: Vera Ferrari, Nicola Bruno, Rabih Chattat & Maurizio Codispoti (2016):
Evaluative ratings and attention across the life span: emotional arousal and gender, Cognition
and Emotion
To link to this article: http://dx.doi.org/10.1080/02699931.2016.1140020
Published online: 10 Feb 2016.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=pcem20
Download by: [Universita degli Studi Di Parma], [Nicola Bruno] Date: 11 February 2016, At: 00:53
COGNITIONANDEMOTION,2016
http://dx.doi.org/10.1080/02699931.2016.1140020
Evaluative ratings and attention across the life span: emotional arousal
and gender
Vera Ferraria , Nicola Brunoa , Rabih Chattatb and Maurizio Codispotib
6 aDepartmentofNeuroscience,UniversityofParma,Parma,Italy;bDepartmentofPsychology,UniversityofBologna,Bologna,Italy
1
0
2
y
ar ABSTRACT ARTICLEHISTORY
u Thisstudywasdesignedtoinvestigatetheevolutionofemotionalprocessingoverthe Received21May2015
r
eb wholeadultlifespanasafunctionofstimulusarousalandparticipants’gender.Tothis Revised19December2015
1 F end,self-reportedaffectiveevaluationandattentionalcapturepromptedbypleasant Accepted5January2016
1 and unpleasant pictures varying in arousal were measured in a large sample of
3 KEYWORDS
0:5 pfraormtic2ip0antots9(0ny=e2a1rs1.)Rbeaslualntscesdhobwyegdenagdeerdaifnfedreenqcueasllyonslpyrefoardaafcferoctsisveseevveanludaeticoandeosf Ageing;gender;evaluative
0 ratings;naturalscenes;
at pleasant stimuli, with opposite patterns depending on stimulus arousal. As age emotion
] increased, low-arousing pleasant cues (e.g. images of babies) were experienced as
o
n more pleasant and arousing by both males and females, whereas high-arousing
u
r stimuli (e.g. erotic images) were experienced as less pleasant only by females. In
B
a contrast, emotional pictures (both pleasant and unpleasant) were effective at
ol capturing attention in a similar way across participants, regardless of age and
Nic gender.Takentogether,thesefindingssuggestthatspecificemotionalcuesprompt
a], [ dinivffoelrveendt sinubajtetcetnivtieonreaslpeonngsaegseamcreonststdoiwffearrednstbaogtehgprloeauspasn,twahnildeubnapsilceamsaencthasntiimsmulsi
m
arepreservedinhealthyageing.
r
a
P
Di
di
Stu Arangeofstudiesoutlinethefactthatcognitive,affec- Todate,manystudieshavefocusedonthesubjec-
egli tjeivcetoarniedssoaccriaolsassptheectsliofefhsupmanan(Bliafeltehsa,veStdaiuffdeirnegnetrt,ra&- taitvieoni)mpaaccrotsosf eamgeotigornoaulpsst,imbuulit(i.fien. dafinfegcstivaereevanlout-
d
a Lindenberger,1999;Kensinger,2009).Indeed,onerel- consistentacrossstudies.Severalstudiesfailedtofind
rsit evant issue that has prompted intensive research in any age effect (e.g. Mikels, Larkin, Reuter-Lorenz, &
e
niv recent years concerns whether and to what extent Cartensen,2005;Neiss,Leigland,Carlson,&Janowsky,
U ageing affects emotional processing (Ebner & 2009; Ritchey, Bessette-Symons, Hayes, & Cabeza,
[
y Fischer, 2014; Isaacowitz & Blanchard-Fields, 2012; 2011; Smith, Hillman, & Duley, 2005; Söderholm,
b
d Scheibe&Carstensen,2010). Häyry, Laine, & Karrasch, 2013; Tsai, Levenson, & Car-
e
ad Emotionhasbeenwidelyinvestigatedviathepres- stensen,2000;Wieser,Muhlberger,Kenntner-Mabiala,
o
nl entationofaffectivepicturesthatareeffectivecuesin & Pauli, 2006; Wurm, Labouvie-Vief, Aycock, Rebucal,
w
evoking a broad range of emotional reactions, and & Koch, 2004), while others found a general decline
o
D
that involve both pleasant and unpleasant affects of the appetitive system activation with increasing
(Bradley, 2000; Ferrari et al., 2016; Lang, Greenwald, age, as all (both high and low in arousal) pleasant
Bradley, & Hamm, 1993). Several studies have shown stimuli were rated as less arousing and less pleasant
that valence and arousal are the most important in older compared to younger participants (Keil &
dimensions of emotional reactivity, and that they Freund,2009).Onthecontrary,otherstudiesreported
control most of the variance of subjective reports higherpleasantnessratingsforlow-arousingpleasant
(Bradley, Codispoti, Cuthbert, & Lang, 2001; Osgood, and/or neutral cues in older compared to younger
1969;Russell&Barrett,1999). adults(Backs,daSilva,&Han,2005;Grühn&Scheibe,
CONTACT VeraFerrari [email protected]
©2016Taylor&Francis
2 V.FERRARIETAL.
2008;Keil&Freund,2009;Mather&Knight,2005;Neiss between sexes have rarely been analysed (Gomez,
et al., 2009; Streubel & Kunzmann, 2011). In terms of von Gunten, & Danuser, 2013; Keil & Freund, 2009).
unpleasantemotions,severalstudiesdidnotobserve One might hypothesise that gender may mediate
relevant age-related changes in reported feelings differential responses to specific emotional cues
promptedbyunpleasantstimuli(e.g.Dolcos,Katsumi, across the life span and may thus contribute to the
& Dixon, 2014; Keil & Freund, 2009; Mather & Knight, variety of findings across studies. Sex differences in
2005; Wieser etal., 2006)but others found enhanced emotional response are well documented in young
subjective reactivity to stimuli inducing negative adult populations (Bradley, Codispoti, Sabatinelli, &
emotionsinoldercomparedtoyoungeradults(Beau- Lang, 2001; Murnen & Stockton, 1997), indicating
6 dreau, MacKay, & Storandt, 2009; Fajula, Bonin-Guil- that women are generally more defensively reactive,
1
0 laume, Jouve, & Blin, 2013; Grühn & Scheibe, 2008; whereas men are specifically more aroused by erotic
2
y Kunzmann&Grühn,2005;Kunzmann&Richter,2009; cues. Whether these affective preferences in men
r
a
u Streubel & Kunzmann, 2011), despite the lack of age and women remain stable or change over the
r
eb differencesintermsofautonomicreactivity(Gavazzeni, ageing process has yet to be clarified. Thus, in the
F
1 Wiens, & Fischer, 2008; Kunzmann & Grühn, 2005; present study, we used a picture-viewing paradigm
1
3 Kunzmann&Richter,2009). toassessemotionalreactivityinparticipantscovering
5
0: One explanation for these inconsistencies may a large segment of the adult life span (age 20–80+),
0
at have to do with differences or weaknesses in the grouped into seven age decades in which males and
] sampling criteriaacrossstudies.Thelargemajorityof femaleswereequallyrepresented.
o
n cross-sectionalstudiesinageingandemotionconsists Itiswellestablishedthatarousalisacriticalfactor
u
r
B of comparisons of samples of young and old adults inemotionalreactivityatthelevelofbothself-report
ola (i.e. extreme-groups design), however, this type of affectiveevaluations,aswellasphysiological(cortical,
Nic experimental design can be problematic for several subcortical and peripheral) changes (Bradley, Codis-
[ reasons. First, two groups that differ in age are also poti, Cuthbert et al., 2001; Codispoti, De Cesarei, &
],
a likely to differ in other respects, hence, it is always Ferrari, 2012; Codispoti, Ferrari, De Cesarei, & Cardi-
m
ar possible that the observed results are attributable to nale, 2006; Lang & Bradley, 2010; Kensinger, 2008;
P
Di some variable other than age. If data are available Streubel & Kunzmann, 2011). Moreover, several
di from adults across the complete age range (i.e. con- recentfindingssuggestthatageingaffectsemotional
u tinuous-groups design, Salthouse, 1999), then we are reactivitydifferentlyasafunctionofthelevelofstimu-
St
egli icnonafobuentdteinrgpovsaitriioanbletos stehpaatraatreethuenliekfefelyctstoofcoovthaeryr l2u0s1a1r)o. uThsaelre(Bfoarcek,sweet aasl.s,e2s0s0e5d;aSgtreeiunbgeelf&fecKtusnhzemreanbny,
d
a with age, or might partly covary within certain age examining emotional reactivity to high- and low-
sit ranges. Second, acontinuous group design allows us arousingpictureswithineachvalencecategory.
r
e
v to assess whether changes in emotional reactivity High-arousing emotional pictures are also known
Uni occur linearly during the ageing process, or whether to engage attentional resources naturally and reflex-
[
y they are more abrupt (nonlinear patterns), as they ively (Bradley, 2009; Ferrari, Codispoti, Cardinale, &
b
d maycoincidewithsignificantbiological,psychological Bradley, 2008; Lang, Bradley, & Cuthbert, 1997). A
e
d or sociocultural changes. Third, in many previous number of studies are consistent in showing greater
a
nlo studies, the “older group” comprised participants attention allocation to pleasant and unpleasant pic-
w with a wide range of ages (60–80+), and the mean tures compared with neutral contents. For example,
o
D agevariedacrossstudies,makingitdifficulttoevalu- interference effects caused by task-unrelated
ate the findings. Considering that people experience emotional pictures have been observed during a
very different conditions (e.g. the degree of mental variety of visual and acoustic tasks, suggesting that
and physical decline) as they grow older (e.g. Baltes motivationally relevant (pleasant and unpleasant)
etal.,1999;Hof&Mobbs,2009),amoreaccuratepor- stimuli draw more on attentional resources, leaving
trayalofsignificantlifechangescouldbeprovidedby them less available for task processing (Bradley,
subdividing the elderly population into separate Cuthbert, & Lang, 1999; Calvo & Nummenmaa, 2007;
decades, rather than grouping together all people De Cesarei & Codispoti, 2008; Mitchell et al., 2008;
over60. Okon-Singer, Tzelgov, & Henik, 2007; Weinberg &
Some of the previous studies did not compare Hajcak, 2011). In addition to self-reports of affective
gender-balanced groups; indeed, differences experience, in the present study we examined
COGNITIONANDEMOTION 3
attentionalcapturebythesameemotionalpicturesin characteristics (Table 1). The second session (held
thesameparticipants,tobetterclarifytherelationship aboutoneweeklater)requiredparticipantstositcom-
between the two components (self-report and atten- fortablyinfrontofalaptopandperformtwotasks:The
tion) of affective processing over the entire adult life emotional interference task, followed, after a short
span. Assuming that self-reports and attention to break,bytheaffectiveratingtask.
emotional cues reflect different facets of emotional Affective well-being measures: The questionnaire
processing, with the former being more vulnerable packet included: (1) The Positive and Negative Affect
to sociocultural factors as well as to subjective Schedule (PANAS); (2) The Affect Intensity Measure
control, we might expect different age-related (AIM); (3) The Emotion Regulation Questionnaire
6 effects across these two measures of emotional (ERQ): (4) The Beck Depression Inventory-II (BDI-II)
1
0 processing. and(5)TheState-TraitAnxietyInventory(STAI).
2
y ThePANASisa20-itemself-reportmeasureofposi-
r
a
u tiveaffect(PA)andnegativeaffect(NA)developedby
ebr Methods Watson, Clark, and Tellegen (1988). PA reflects the
F
1 Participants extent to which a person feels enthusiastic, active
3 1 and alert. NA reflects the extent to which a person
5 Atotal of 211 adults (108 female) participated in this
0: feelsNAstates,suchasbeingnervousandupset.Par-
0 study. Participants were distributed among seven
] at different age decades (from age 20 through 90), t(giceipnaenratsl winesrteruacstikoend; tsoeeratWeahtosowntheetyafle.,lt1i9n8g8e).nTehrael
o
n with a balanced number of males and females in AIM(Larsen&Diener,1987)isa40-itemquestionnaire
u
Br each age group. Most of the young adults (20–29 that measures trait levels of affective intensity. A
ola years old) were undergraduate psychology students sample item from the AIM is the following: “My
Nic from the University of Bologna; middle-aged and emotions tend to be more intense than those of
[ olderadultswererecruitedthroughvoluntaryorganis- most people.” The AIM has been shown to have
a], ations and social centres for elderly people. Prior to good internal consistency, test–test reliability and
m
ar participating, all subjects were informed concerning good discriminant validity (Larsen & Diener, 1987).
P the potentially disturbing content of some stimuli
udi Di awnitdhsitghneedetahiwcarilttsetnanindfaorrdmsedlaicdondsoewntnininactchoerd1a9n6c4e TeEhxripenrgEeRstwsQivo(eGfaSrcoutspospr&sr:eJCsosohigonnn,i2t(if0vo0eu3rR)ietcaeopmnpssri)as.tiTsshaoelf(ss1iixx0-iiittteeemmmssr)ecaaonpvd--
St DeclarationofHelsinkiandwiththerecommendations
egli of the local ethical committee of the Department of pinrdaiivsaidlusaulsbsrceagleulaatsesetshseeisr ethmeotfrioenqsuetnhcroyuwghiththwehuicshe
a d PsychologyofUniversityofBologna. of thought-change strategies. The four-item suppres-
sit Allparticipantshadnormalorcorrected-to-normal sion subscale measures the extent to which individ-
r
ve visualacuityandwerescreenedusingastandardised uals regulate their emotions through the use of
ni health interview prior to inclusion in the study. All
U suppression strategies. The ERQ has been shown to
y [ participants scored 27 or higher on the Mini-Mental have excellent psychometric properties. The BDI-II
d b State Examination (Folstein, Folstein, & McHugh, (Beck, Steer, & Brown, 1996) consists of 21 items to
de 1975), consistent with normal cognitive functioning. assess the intensity of depression in clinical and
a
o Education was lower with increasing age F(6,197)=
nl normal patients. Each item has a list of four or more
ow 35.8, p<.0001, h2p =.5, in line with the normative statements,arrangedaccordingtoincreasingseverity,
D increase in formal education over the last century in about a particular symptom of depression. The STAI
Italy.Thispatternwasnotmodulatedbyparticipant’s (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs,
gender (agexgender, F<1). Table 1 presents partici- 1983)isaself-reportassessmentdevicethatincludes
pants’ demographic and affective well-being infor- separate measures of state and trait anxiety. We
mationineachgroup. report only the trait form, composed of 20 items.
Trait anxiety denotes a general tendency to respond
withanxietytoperceivedthreatsintheenvironment.
Materials and procedure
Experimental stimuli: The stimuli were 72 coloured
The study consisted of two sessions. During the first pictures (800×600 pixel resolution) selected from
session,participantscompletedseveralquestionnaires the International Affective Picture System (IAPS,
assessing demographic and affective well-being Lang, Bradley, & Cuthbert, 2008), consisting of 24
4 V.FERRARIETAL.
Table1.Demographicandaffectivewell-beinginformationinmaleandfemaleparticipantsacrossdifferentagegroups.
Agegroup N Age Education(year) MMSE AIM STAI PA NA BDI ERQs ERQr
20–29 F=20 25(2.1) 12(2.1) 30(0.5) 166(24) 43(5.3) 31(6.4) 24(5.3) 8(4.1) 11(5.1) 30(6.1)
M=20 25(2.5) 11(2.3) 30(0.6) 143(18) 40(8.7) 28(4.6) 23(4.3) 7(4.6) 12(3.7) 27(6.9)
30–39 F=15 34(2.8) 16(3.5) 30(0.2) 154(25) 39(8.3) 31(4.8) 22(4.9) 7(5.6) 11(5.0) 29(7.7)
M=15 34(2.8) 15(4.1) 30(0.1) 144(20) 35(5.9) 33(4.4) 21(4.4) 5(4.1) 14(4.1) 28(7.0)
40–49 F=15 44(3.1) 14(3.3) 30(0.4) 154(16) 38(9.6) 30(4.4) 22(4.5) 6(4.2) 13(5.9) 30(6.7)
M=15 45(2.9) 13(4.3) 30(0.1) 139(23) 38(7.6) 32(4.3) 22(5.4) 7(6.2) 14(5.3) 30(5.3)
50–59 F=15 54(2.7) 12(2.7) 29(0.5) 154(15) 46(7.4) 31(4.1) 24(6.2) 8(5.6) 15(4.4) 29(5.5)
M=14 54(2.8) 11(3.4) 30(0.3) 142(23) 36(7.4) 32(3.4) 22(5.2) 5(3.5) 13(3.2) 31(6.9)
60–69 F=14 65(2.6) 8(3.3) 30(0.8) 157(15) 38(9.4) 32(4.4) 25(6.9) 10(6.7) 15(6.5) 32(8.4)
6 M=12 67(2.2) 8(3.3) 29(0.8) 145(24) 39(9.2) 31(5.6) 24(5.6) 8(5.1) 18(4.8) 34(5.4)
1 70–79 F=13 73(2.8) 6(1.5) 30(1.0) 149(23) 45(6.3) 32(5.3) 27(4.6) 17(9.2) 19(7.4) 35(6.6)
0
2 M=14 74(2.5) 7(3.4) 28(1.8) 153(25) 38(6.1) 30(4.1) 24(6.2) 11(6.4) 16(6.4) 29(9.7)
y >80 F=17 83(2.9) 6(2.1) 28(1.5) 153(21) 39(9.1) 32(4.6) 25(4.7) 10(5.6) 19(7.6) 35(4.7)
r
a M=13 84(2.7) 7(2.9) 28(1.7) 155(21) 40(7.7) 31(4.4) 24(4.9) 11(7.6) 19(5.9) 34(6.4)
u
r
b Note.MMSE,Mini-MentalStateexamination(Folsteinetal.,1975);AIM,AffectiveIntensityMeasure(Larsen&Diener,1987);STAI:State-Trait
e
F Anxiety Inventory (Spielberger et al., 1983; Italian version: Pedrabissi & Santinello, 1989); Positive (PA) and Negative (NA) affect scale
1 (PANAS, Watson etal., 1988); BDI, BeckDepressionInventory-II(Beck, Steer,&Brown,1996); ERQsuppressionscale andERQreappraisal
1
3 scaleoftheEmotionRegulationQuestionnaire(Gross&John,2003).Allvaluesrepresentrawmean(±SEM)scores.
5
0:
0
at pleasant,24unpleasantand24neutralpictures.Based “m”). For half of the participants, the probe-digit
]
no on normative ratings acquired from college students mapping was reversed. Participant response was
u
r (Langetal.,2008),bothpleasantandunpleasantpic- allowed during the target presentation, as well as
B
a tureswereselectedinsuchawaythathalfweretypi- during the inter-trial interval (for a total of 6s). Two
col cally rated as high-arousing – pictures of erotic presentation orders were constructed that varied,
Ni
couples and pictures of injured bodies and the other across participants, the order of picture presentation,
[
], half as low-arousing stimuli – pictures of babies/ as well as the association of a specific picture with
a
m
families and pictures of accidents, contamination, the target (square or circle). In this task, there was a
r
a
P illness and loss. Neutral pictures depicted non-arous- totalof72trials,correspondingtothe72picturesorig-
Di ing scenes, such as people and objects. Therefore, inally selected for the study, so that each picture
di picturestimuliwerearrangedasafunctionofstandar- exemplar was presented only once throughout the
u
St dised pleasantness (p) and arousal (a) ratings (Lang task.
gli et al., 2008) in five picture content categories: high- Intheaffectiverating,thesame72picturesusedin
e
d arousingpleasant(p=6.5;a=6.4)andunpleasantpic- the interference task were then presented again and
a
sit tures (p=1.9; a=6.3), low-arousing pleasant (p=6.6; subjects viewed each picture for as long as they
r
e a=4.5) and unpleasant pictures(p=2.8; a=5.1), and liked,terminatingexposurewithakeypress.Afterter-
v
ni neutralpictures(p=5.3;a=3.4). mination, the subjects were required to rate first the
U
[ In the emotional interference task (Mitchell et al., pleasantness (i.e. valence) and then the intensity (i.e.
y
b 2008; Weinberg & Hajcak, 2011), participants were arousal) of the experienced emotional state on a 9-
d
e instructed to respond with their left and right index point scale using the Self-Assessment Manikin (SAM;
d
oa fingers in a simple discrimination task (circle/square Lang, 1980). A computer version of the SAM scales
wnl stimuli, 3° x 3°of visual angle, white filled on a black was displayed and remained on the screen until par-
Do background).Thedisplayshowingthetargetstimulus ticipants provided their ratings using the 1–9 keys
(circle/square) was always preceded by the presen- onthelaptopkeyboard.
tation of an IAPS picture (either emotional or For all participants, the interference task always
neutral), which was irrelevant for the task. The com- went first, so that the amount of attentional capture
ponentsofeachtrialwerepresentedserially.Firstafix- was not affected by picture repetition (Ferrari,
ationcrosswasshown(1s),thenanIAPSpicture(2s), Bradley, Codispoti, & Lang, 2015). Stimulus presen-
which was immediately followed (no blank in tation and response recording were controlled by E-
between) by the onset of the target (2s). Subjects Prime (Version 1.2) experimental software. Both tasks
were required to respond as fast and accurately as were presented on a PC laptop, situated approxi-
possibleastowhethertheysawasquareorcircleby mately 0.5m from the participant. For both tasks,
pressing one of two digits on the keyboard (“x” or instructions were given both orally by the
COGNITIONANDEMOTION 5
experimenter and visually on the laptop screen, and p<.0001,h2 .1,indicatedthatchangesinpleasant-
p =
before beginning the actual task, participants per- nessratingofspecificpicturecontentsweredifferent
formed several practice trials to ensure compliance formenandwomenacrossagegroups.Thisthee-way
with task instructions. Each task consisted of a total interaction was primarily due to high-arousing plea-
of72trials.Eachsessiontookabout1hour. sant pictures, mostly depicting erotic contents (see
Inaccordancewiththeguidelinesofthejournal,we Figure 1, lower-left panel): whereas men rated high-
reportedhowwedeterminedoursamplesize,alldata arousing pleasant pictures as highly pleasant, com-
exclusions, all manipulations and all measures in the pared to neutral contents, F(1,96)=392, p<.0001,
study. h2 .78, and this pattern remained stable across
p =
6 the whole life span (main effect of age group for
1
0 high-arousing pictures in men, F<1), women rated
2
y Data analysis this picture content as less pleasant with advancing
r
ebrua AagmeuGlrtoivuapria(7te:Aregpee2a0te–d29m,3e0a–s3u9re,4A0N–O49V,A5,0w–5it9h,f6a0c–to6r9s, a<g.0e0,0F(16.,1M01o)r=e8s.5p,epci<fic.0a0ll0y1,,who2pm=e.n33;inlinyeoaruntrgendanpd
F
1 70–79,Over80)xGender(2)xpictureContent(5:Plea- middle-aged adulthood (age group 1–4; age < 60)
1
3 sant High-Arousing, Pleasant Low-Arousing; Neutral, rated high-arousing pleasant pictures as more plea-
0:5 UnpleasantHigh-Arousing,UnpleasantLow-Arousing) sant then neutral pictures (Fs>9, h2p .4) with no
0
at wasconductedonpleasantnessandarousalratingsas difference across groups; the pattern was then
o] wellasonreactiontimes(RTs)intheemotionalinter- reversed in older women (age > 60), who rated this
un ference task.1 All post-hoc tests were corrected for picture content as less pleasant than younger
Br women(ps<.05),andevensignificantlylesspleasant
multiple comparisons with the Bonferroni procedure.
ola The partial eta squared statistic (h2), indicating the than neutral pictures (age group 60–69, F(1,13)=9.8;
Nic proportion of the variance explainepd by one exper- p<.01,h2p =.4 ; 70–79, F(1,12)=7.7; p<.05,h2p =.5;
[ over80,F(1,16)=37;p<.0001,h2 .7).
], imental factor and the total variance, was calculated p =
a Ontheotherhand,bothlow-arousingpleasantand
m andisreported.
ar Analysis of RTs was performed only for accurate neutral pictures were rated as more pleasant with
P
Di response trials. In addition, RTs that were shorter or increasingage(maineffect ofgroup,Fs(6,204)>12.4,
udi longer than the mean (within each subject and con- pslsig<h.t0b0u0t1,sighn2pifi.ca.n3;tslitneeeaprertrinecnrdeapses<fo.r0l0o0w1-)a,rwouitshinga
St dition)±3 times the standard deviation of the mean
degli werAenstoattiisntcicluadlepdoiwnetrheananalaylsyissis.was performed for p(6le,1a9s7a)n=t,2c.o8m, ppa<re.0d5,tohn2pe.utr.8al),.pTihctisuriensc(rleinaesearitnrepnlde,aF-
a sample size estimation (G*Power 3; Faul, Erdfelder, santness for low-arousing pleasant and neutral pic-
rsit Lang, & Buchner, 2007). Assuming a medium effect tures with increasing age was similar in men, Fs
e
niv size (Cohen’s f=0.25; Cohen, 1988) at 95% power, (6,101)>5.7, ps<.0001, h2p ..25, linear trend, ps
U the current sample (n=211) was considered reason- >.0001, and women, Fs(6,96)>4.5, ps<.0001,
[
by able to detect a statistically significant interaction h2p ..22; linear trend, ps>.0001. Therefore, whereas
d between our factors (age group, gender and picture young and middle-aged men generally rated low-
e
d arousing pleasant pictures as less pleasant than
a content)aswellasthemeancomparisonsofinterest.
nlo high-arousing pleasant pictures (group 20–29;30–39;
w 40–49 Fs>52, 20, 9, respectively), in older men (>50
o
D Results years old), pleasant low-and high-arousing pictures
didnotdifferinpleasantnessratings.
Affective space
It is worth noting that pleasantness ratings of
Figure1presentseachofthe72picturesusedinthis unpleasant pictures, either high or low in arousal,
study (solid symbols) in the two dimensional space did not change across the life span (Fs<1). High-
formedbyplottingeachpicturebyitsmeanpleasant- arousing unpleasant pictures were rated as more
ness (i.e. valence) and arousal rating separately for unpleasant than low-arousing pictures in all age
women (upper panel)and men(lower panel)in each groups (Fs>314), and compared to men, women
agegroup. rated both low- and high-arousing unpleasant
Pleasantness rating: A significant age Groupx pictures as generallymore unpleasant, Fs(1,197)>21,
Genderxpicture Content interaction, F(24, 677)=2.3, ps<.0001,h2 ..1.
p
6 V.FERRARIETAL.
6
1
0
2
y
r
a
u
r
b
e
F
1
1
3
5
0:
0
at
]
o
n
u
r
B
a
ol
c
Ni
[
],
a
m
r
a
P
Di
di Figure1.(a)Eachofthe72picturespresentedinthisexperimentisplottedintheaffectivespaceformedbyitsmeanpleasureandarousalrating
u
St scale,averagedacrossparticipantsseparatelyforthesevenagedecadesandforwomen(upperrow)andmen(lowerrow).Coloursrepresent
gli phiigcthu-raerocuosnitnegn)t.((bb)luEea:cphlecairscalentrehpigrhes-aernotusstihneg;mliegahntbvaluluee:p(±leaSsEaMnt)loofwa-raoruosuaslinragt;inggresefno:rnneeuuttrraall;(pgurrepelne:suynmpbleoalss)aanntdlolwow-a-raoruosuinsign;grepdle:ausnapnltea(lsigahntt
e
d bluesymbols)picturesineachagegroup.(c)Eachcirclerepresentsthemeanvalue(±SEM)ofpleasantnessratingsforhigh-arousingpleasant
a picturesineachagegroup,separatelyforwomen(solidsymbols)andmen(opensymbols).
sit
r
ve Arousal rating: A significant interaction of picture h2 .41 linear trend, p<.0001, and women, F
Uni Contentxage Group, F(24,678)=6.9, p<.0001, (6p,1=01)=7,p<.0001,h2 .29lineartrend,p<.0001.
by [ h2p =.17, indicated that an age effect for arousal As expected, men apnd=women differed in arousal
d ratingswasevidentonlyforlow-arousingpleasantpic- ratings in all picture contents except for neutral
e
d tures, main effect of Group, F(6,197)=18, p<.0001, images, genderxcontent F(4,194)=19, p<.0001,
a
nlo h2p =.35, which were rated as increasingly more h2p =.28.Comparedtomen,womenratedunpleasant
w arousing with advancing age (linear trend, p<.0001). pictures,bothhighandlowinarousal,asmorearous-
o
D More specifically (see Figure 1, lower-right panel), ing,Fs(1,197)>11,ps<.005,h2 ..1;conversely,men
p
young adults (20–29) rated these pictures as less rated high-arousing pleasant pictures as more arous-
arousing than did older people starting from 40 ing, F(1,197)=26, p<.0001,h2 .16, and low-arous-
p =
yearsofage(allcomparisons,ps<.005).Also,arousal ing pleasant pictures as less arousing F(1,197)=8, p
ratings of middle-aged adults (40–49; 50–59 years) <.01, h2 .05, than women. These gender effects
p =
were significantly lower compared to older people wereconsistentacrossagegroup.
(from 60 to over 80 years; ps<.05). No further differ-
ences were found between older age groups (60–69
Emotional interference task
vs. 70–79 vs. 80–89, Fs<1). Moreover, this linear
increase in arousal ratings for low-arousing pleasant Reaction times to the visual target (either the square
pictures was similar in men, F(6,96)=11, p<.0001, or the circle) were significantly modulated by picture
COGNITIONANDEMOTION 7
content, F(4,792)=25, p<.0001, h2 .11; quadratic worse thanall theotheragegroups,F(6,197)=7.1, p
p =
trend, F(1,198)=68, p<.0001;h2 .26, see Figure 2: <.0001,h2 .18.
p = p =
high-arousingpictures,bothpleasantandunpleasant,
promptedlongerRTs,comparedtoneutral,Fs(1,197)
Individual differences and ageing
>51, ps<.0001,h2 ..21, and low-arousing pictures,
p
pleasantlow,Fs>38,ps<.0001,h2 ..16;unpleasant AmultivariateANOVA,withfactorsageGroup(7:Age
p
low,Fs>9,ps<.005,h2 ..04.High-arousingpleasant 20–29,30–39,40–49,50–59,60–69,70–79,Over80)x
p
and unpleasant pictures did not differ from each Genderwasconductedonallquestionnairescores.All
another.RTsafterviewingoflow-arousingunpleasant post-hoc tests were corrected for multiple compari-
6 pictures were also slower compared to neutral and sons with the Bonferroni procedure. Table 1 reports
1
0 low-arousing pleasant pictures, F(1,197)=28, p mean scores (±SEM) in each age group, separately
2
y <.0001, h2 .13; F(1,197)=16, p<.0001, h2 .08, formalesandfemales.
ar p = p =
u respectively, whereas low-arousing pleasant and AgeeffectswerefoundfortheBDI,F(6,197)=7.4,p
r
Feb neutralpicturespromptedsimilarRTs. <.0001,h2p =.18, as well as for the ERQ, both in the
53 11 witMhoagreeiinngte,rFe(s6t,i1n9g7l)y,=d2e0s,ppit<e.a00g0e1n,ehra2pl=RT.3s8lo,wpdicotuwrne rperaepsspioranissacl,alFe(,6,F1(967,1)9=7)3=, p7.<7,.0p0<1,.0h02p01=,h.02p8=, a.n1d9.sTuhpe-
0: contentsimilarlymodulatedRTsinallagegroups,(age highest BDI scores were observed in the age group
at 0 Groupxpicture Content, F(24,678)=1, p=.4; quadra- 70–79, being significantly higher compared to the
] tic trends in each age group, Fs>6, ps<.05, groups ranging from 20 to 60 years old, Fs>.005.
o
n h2 ..17, Table 2). Aside from a small Gender effect, The oldest group (>80) only differed from age
u p
Br F(1,197)=6.5, p<.05, h2 .03, due to slower RTs in decade 30–39, F(1,56)=10, p<.005, h2 .16, and
ola womencomparedtompen=,noother significanteffect the group 60–69 was similar to all agepg=roups. For
Nic involvingGenderwasfound. the ERQ-reappraisal scale, post-hoc comparisons
[ Discrimination accuracy was overall good (> 97%) revealed that only the oldest group (>80) scored
],
a across participants, and was not affected by picture higher compared to the two youngest groups, ps
m
ar content. Women were slightly less accurate than <.05.SimilarageeffectswerefoundfortheERQ-sup-
P
Di men, F(1,6)=6,13, p<.05, h2p =.03, means=97, pression scale, where the highest scores were found
udi 9880%y,eraersspeocltdiv,emlye,aann=d9t5h%e)opldeersfotrpmaerdticispiganntifisc(aonvtelyr ayoguainngeinr tghreouapgse2g0r–o5u9p,>p8s0<, .t0h5a.tTdhieffegrerdoufprom70–th79e
egli St sdceocraeddesh(2ig0h–e2r9,c3o0m–3p9a,rped<.0to001th;pe<fi.0rs5t,rtewspoecytiovuenlyg)
d
a and the group 60–69 only differed from the very
sit young group (20–29, p<.01).The three old groups
r
ve didnotsignificantlydifferfromeachother.
Uni An overall significant Gender effect was found for
[
y NA, AIM, STAI and BDI, with females showing higher
b
d scores compared to males, F(1,197)=4.7, p<.05,
e
d h2 .02; F(1,197)=10, p<.005, h2 .05; F(1,197)=
nloa 8.p4,= p<.005, h2p =.04; F(1,19p7=)=4.4, p<.05,
w h2 .02, respectively. No significant age Groupx
o p =
D Gender interaction was found in any of the
questionnaires.
Thispatternofresultswasconfirmedbyamultiple
linear regression analysis, in which participant’s age
was treated as continuous variable. The BDI and
boththeERQsubscales(suppressionandreappraisal)
showedsignificantregressioncoefficients,Fs(2,209)>
9,p<.0001,R2s>.08,whereagewastheonlysignifi-
Figure2.OverallmeanRTs(±SEM)toavisualtargetdiscrimination cant predictor (βs>.27, ps<.0001). A significant
(circle/square) after viewing of a pleasant, neutral or unpleasant
regression equation was found, Fs(2,209)>4, p<.05,
picture. Pleasant and unpleasant pictures were further subdivided
accordingtoarousalintensity(highorlow). R2s>.03, for the STAI, NA and AIM, where only
8 V.FERRARIETAL.
Table2.Rawmeans(±SEM)ofRTs(ms)tothevisualtarget(circle/square)foreachagegroupandpicturecontent.
20–29 30–39 40–49 50–59 60–69 70–79 >80
Pleasanthigh 638(26) 718(25) 643(30) 689(42) 812(57) 1045(66) 1143(57)
Pleasantlow 607(25) 676(35) 611(26) 664(34) 750(51) 979(63) 1112(68)
Neutral 615(24) 677(42) 609(24) 672(34) 731(46) 967(65) 1109(64)
Unpleasantlow 639(24) 705(39) 637(26) 690(35) 772(53) 1008(64) 1123(61)
Unpleasanthigh 670(27) 747(35) 684(29) 697(34) 803(67) 1011(66) 1142(69)
genderwasasignificantpredictor,βs>−.19,ps<.01, inolderpeople,forexampleprovidingcuestoretrieve
suggesting that females, compared to men, showed autobiographical past memories related to relatively
6 higher scores in all these questionnaires. PA scores happy times, when they were younger and actively
01 did not show any significant relationship with age or engagedinchildrearing(Charles,Mather,&Carstensen,
2
y gender. 2003). Whether this selective difference in emotional
r
ua preferencesreflectsanexplicitandconsciousstrategy,
r
b or rather, a more inherent biological need is hard to
e
1 F Discussion establishhere,buteitherwayitmightreflectanadap-
3 1 The present study provides clear evidence for age- tivebehaviourforthebenefitofmomentarymoodand
0:5 related changes of subjective emotional experience presumably long-term well-being (Freund & Baltes,
at 0 to specific pleasant cues, whereas unpleasant cues 2007;Scheibe&Carstensen,2010).
] seem to prompt emotional feelings that remain A different age-related effect mediated by gender
o
n stableacrossthelifespan.Moreover,ageingaffected differences was found in the subjective ratings of
u
Br emotional reactivity differently as a function of the high-arousing pleasant pictures, mostly depicting
ola arousallevelofthestimuli.Low-arousingpleasantpic- erotic couples. Sexualcues represent primaryreinfor-
Nic tures, which mostly included pictures of families and cers that directly engage the appetitive system, and
[ babies,promptedreportsofmorepleasantandarous- several studies have demonstrated that, within plea-
],
a ing affective experiences in elderly people, and this santpicturecategories,sexualcuesarethemosteffec-
m
ar was found for all participants regardless of gender. tive in eliciting autonomic, cortical and subcortical
P
Di An opposite pattern, that is, a decrease in pleasant- changes (Bradley, Codispoti, Sabatinelli et al., 2001;
di ness,wasfound,however,forhigh-arousingpleasant Codispoti & De Cesarei, 2007; Sabatinelli, Bradley,
u pictures, which seem to vary in affective meaning Lang, Costa, & Versace, 2007; Schupp, Junghöfer,
St
egli annodtinprmomenp,todveifrfetrheentaigaleirnegsppornodceinsgs. in women, but Wpreeiskeen,t&stuHdaymsmho,w2e0d04t)h.aCtoynosuisntgenatndwimthidtdhleis-,agtehde
d
a Severaltheoriesonemotionalageing(forareview adults, both men and women, reportedpleasant and
sit seeScheibe&Carstensen,2010)holdthattheageing intense subjective reactions after viewing pictures
r
e
v process may prompt changes in emotional prefer- depicting sexual contents, and consistent with pre-
ni
U ences and strategies, enhancing emotionally gratify- vious studies, ratings were less extreme in women
[
y ing experiences and minimising negative ones. (Bradley, Codispoti, Sabatinelli et al., 2001; Codispoti,
b
d Consistent with these theories, in the present study, Surcinelli, & Baldaro, 2008; Murnen & Stockton,
e
d pleasant emotions prompted by low-arousing plea- 1997).Withincreasedage,mencontinuedtoevaluate
a
nlo sant cues are reported to be more pleasant and eroticscenesashighlypleasantandarousing;women,
w more intense with increasing age. Interestingly, this instead, rated this picture content as unpleasant as
o
D age effect specific for pictures of babies and smiling they grew older (60+). A decrease in pleasantness
families seems to develop linearly over age decades, ratings with ageing was previously observed (Backs
starting from middle adulthood (> 40 years old) and etal.,2005;Keil&Freund,2009),withpleasantpictures
becoming even more evident across old age; more- prompting lower pleasantness in older, compared to
over,itappliesequallytomenandwomen,suggesting younger, adults, but no sex modulatory effects
thatitdoesnotnecessarilydependonmotherhoodor emerged in these studies, probably because of the
genderroles,butratherreflectsageneraldrivetoward relative lownumber ofmalesinthesamplegroup or
these affective cues. One possible speculation could thespecificpicturecontentsusedinthehigh-arousing
be that the increased affective engagement for condition.Thus,thepresentfindingsindicatethatthe
these specific contents reflects the need to age-related appetitive impairment for high-arousing
somehowdealwiththelimitedfuturetimeperspective contents depicting sexual cues more likely affects
COGNITIONANDEMOTION 9
the female population, rather than being a general group differences in emotional processing simply
effectofageing. reflect sociocultural or personality traits (e.g. PA and
We have to be cautious in interpreting these NA) that differ across groups. Data derived from the
results, though, given the nature of the subjective affective well-being questionnaires did not reveal
ratings,whichareclearlyatypeofmeasureofaffective any reliable difference between age groups, except
reactivity that can be highly vulnerable to sociocul- for a slight increase of the BDI scores in the age
tural factors as well as to subjective control. In our decades ranging from 70 to 79 years, compared to
study, thus, ratings of sexual contents could be younger participants. Consistent with recent findings
biased by social desirability/appropriateness, in that highlight the role of emotion regulation in suc-
6 addition to (or instead of) subjective feelings about cessful ageing (Suri & Gross, 2012), increasing age
1
0 the stimulus. Due to possible cohort effects, the wasassociatedwithbothtypesofemotionregulation
2
y groupofolderwomencouldbemoresensitive,com- strategies (ERQ-suppression and -reappraisal).
r
a
u pared to older men or younger participants, to the However, these age differences in the questionnaire
r
eb socialcontextoftheexperimentandfeeluncomforta- scoresdonotseemtohelpinexplaininggroupdiffer-
F
1 ble watching pictures of erotica in such context, ences observed in evaluative ratings. Future studies
1
3 prompting therefore a negative judgment towards with a larger sample size are needed to explore
0:5 these specific picture content. Various sociocultural more closely the role of sociocultural (e.g. religiosity
at 0 factors,suchascultureandreligion,haveaninfluence beliefs) factors as well as personality traits in age-
] on sexual attitudes and behaviour (Rupp & Wallen, relateddifferencesofemotionalreactivity.
o
n 2008), and we cannot exclude that our sample of Theperceptionofemotionalcuesisalsoassociated
u
r
B olderfemaleparticipantshadahighsenseofreligios- with attentional engagement (i.e. natural selective
ola ity. Thus, rather than reflecting a real age-related attention, Bradley, 2009), here measured through an
Nic decline in appetitive engagement towards erotic emotionalinterferenceparadigm.Overall,participants
[ cues, these findings may reflect group differences in showed the typical RT slowdown to the target after
],
a a more general social desiderability bias, triggered viewing a high-arousing emotional (either pleasant
m
ar byerotic cues. Distinguishing betweenthese alterna- or unpleasant) picture, compared to when the
P
Di tivesreliesonfutureresearchinwhichtherelationship picture depicted a neutral content, consistent with
di between age, gender and sociocultural factors in previous findings with different variants of the para-
u affectiveratingswillbecarefullyassessed. digm (Bradley et al., 1999; De Cesarei & Codispoti,
St
egli encTehseaplroensgentadceosingtninaululomw,edshuoswtiongtratchkatagtehedifafegre- 22000184;; MFeitrcrahreilleettaall..,, 22000088;;WFeerirnabrie,rgMa&stHriaaj,ca&k,B2r0u1n1o),.
d
a effect specific for low-arousing pleasant pictures Low-arousing unpleasant pictures were also effective
sit appears gradually over age decades, already starting inpromptingaslightbutsignificantRTslowdownrela-
r
e
v from middle (>40 years) adulthood, whereas the age tivetoneutralandlow-arousingpleasantpictures.Itis
ni
U effectforhigh-arousingpleasantpicturesonlyoccurs morerelevanttothecurrentstudytonotethatthisRT
[
y in late (> 60 years) adulthood; both effects, however, affective modulation was similar across age groups:
b
d are more evident starting from the group of sixty, both high-arousing pleasant and unpleasant pictures
de possibly reflecting the impact of relevant events of were equally effective in capturing attention in
a
o
nl social life (e.g. retirement, becoming grandparents) younger as well as older people, suggesting that the
w that somehow mark the official transition into the basicmechanisminvolvedinattentionalengagement
o
D period of old age. Although future studies may con- prompted by motivationally relevant cues, which is
sider to increase the sample size (n>20) per group clearly driven by emotional arousal and not valence
to ensure a more adequate statistical power, the fact (Wangelin,Bradley,Kastner,&Lang,2012)ispreserved
that these effects of ageing in emotional reactivity from ageing (Knight et al., 2007; Leclerc & Kensinger,
remain pretty much stable across the three groups 2008; Murphy & Isaacowitz, 2008; Steinmetz, Musca-
of older people (60–69; 70–79; > 80 years) makes tell,&Kensinger,2010).
ourfindingsratherconvincing,1asitavoidsconfound-
ing issues that may arise when comparing two
Summary and conclusions
extreme groups (“extreme-groups design”, see intro-
duction). When ageing effects in emotional reactivity Evaluative ratings of pleasant emotions were clearly
are observed, a critical question is whether the age affected by the ageing process, showing, on one