Emotional categorization: Individual emotional differences and laterality effects in healthy and persons with multiple sclerosis

doi:10.4236/health.2010.212201

Paper Menu >>

Journal Menu >>

Vol.2, No.12, 1349-1365 (2010) Health

doi:10.4236/health.2010.212201

Emotional categorization: individual emotional

differences and laterality effects in healthy and

persons with multiple sclerosis

Farzaneh Pahlavan1*, Tudd Lubart2, Michelle Montreuil3, Stephen Jacob2, Anne-Yves

Jacquet2, Christelle Lemoine2, Hélène Petropoulou3, Franck Zenasni2

1Laboratoire de Psychologie Sociale, Université Paris Descartes, Boulogne, France;

*Corresponding Author: farzaneh.pahlavan@parisdescartes.fr;

2Laboratoire de Psychologie et Neurosciences Cognitives – CNRS, Université Paris Descartes, Boulogne, France;

3Equipe de Recherche en Psychologie Clinique, Université Paris 8 Saint-Denis, Paris, France.

Received 1 October 2010; revised 20 October 2010; accepted 25 October 2010

ABSTRACT

A study examining affective information

processing in persons with Multiple Sclerosis

and healthy adults was carried out. It was hy-

pothesized that individual characteristics could

modulate participants’ emotional categorization

and reaction times for categorization decisions.

For example, individuals w ith negative valenced

emotional profile (e.g. anxious) should choose

negative emotional alternatives faster and more

frequently. Participants consisted of two differ-

ent populations: 80 right-handed healthy

French-speakers, and 40 right-handed French-

speakers with multiple sclerosis. The results

showed a positive correlation between high-

level of negative emotional sensibility and emo-

tional categorization (decision and decision

speed) for affective information presented on

the right-side of the screen. For all participants

there were more fre quent emo tional choices a nd

faster decis ions for left-side prese nted em ot io na l

alternatives. It seems individuals’ emotional dif-

ferences in general and in MS populations mo-

dulate hemispheric asymmetry of processing

emotional judgments.

Keywords: Emotion; Categorization; Hemisphe ric

Asymmetry; Individual Emotional Differences;

Multiple Sclerosis

1. EMOTIONAL CATEGORIZATION:

INDIVIDUAL DIFFERENCES AND

LATERALITY EFFECTS

There are numerous ways in which emotions and af-

fective processes shape and organize cognitiv e activities.

Some studies demonstrated creative, explorative beha-

vior under positive mood and careful, error-avoidance

behavior for negative mood (e.g. [1-5]; see also [6]). It

has been also reported that emotional states can enhance

high-level cognitive control [7], focus attentional re-

sources, influencing encoding and organization of new

information [8], and facilitating access to information

previously acquired [9,10]. The influence of mental re-

presentation of affective reactions on the accessibility of

concept and knowledge has been also suggested [11,6].

Individual differences seem also be related to emo-

tional reactivity and processing information, specifically

emotional-relevant information (e.g. [12-14]). Our know-

ledge of the ways in which individual differences and

personality traits are associated with different affective

and cognitive processes has grown considerably in re-

cent years (e.g. [12-14] see also [6]). For example, it has

been shown that typical happy people perceive, categor-

ize, and retrieve pleasant information more readily, more

easily than typical unhappy people, and may even in-

terpret ambiguous stimuli more favorably [15,16]. In the

same manner, an anxiety-related responses bias [17] and

also an attention bias resulting in interpretation of am-

biguous stimuli as threatening ones have been noticed in

typical anxious individuals [18]. Therefore, researchers

have proposed hypotheses about the affective- con-

gruency [19] and trait-congruency [20].

Nevertheless, less work has been devoted to uncover-

ing the basic factors of these influences, considering

neuropsychological evidence in analyzing individual dif-

ferences and affective-cognitive processes. In th e last few

years, however, this situation has begun to change. Con-

siderable evidence demonstrated that individual differ-

ences, for example, in temperament are associated with

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1350

differences in brain and peripheral physiological func-

tioning (see [13,7]). Findings from some studies suggest

that alexithymia (less ability to identify and communicate

feelings) could be investigated as a continuous personal-

ity trait associated with either the interhemispheric trans-

fer of information [21] or the relative development and

activation of the two hemispheres in non-clinical and

clinical populations ([22]; but see [23]) or as a syndrome

related to some psychiatric disturbance such as Multiple

Sclerosis. Therefore, individual differences in processing

emotional stimuli provide a potentially rich source of

information about the relationship between emotion and

cognition, in normal people but also in individuals suf-

fering from neurobiological impairment. To the extent

that we understand the functions of these systems, we

will be in a better position to develop more selective

therapies that are targeted for the specific brain networks

involved in regulation of the specific aspects of emotion-

al psycho-pathologi cal funct i ons.

The present study focuses on cognitive performance

in terms of emotional categorization in the general pop-

ulation, and also considers a clinical sample of patients

suffering from multiple sclerosis. We examine the extent

to which information processing depends on stable emo-

tional-related individual differences and also lateralized

nature of these processes. After reviewing some evi-

dence concerning categorization processes of affec-

tive-relevant information, we focus on hemispheric

asymmetries involved in processing of emotional cate-

gorization, and the potential role of emotion-related in-

dividual differences. Finally, we report the results of our

study designed to examine these relatio nships.

Categorization, identification and grouping of entities

into sets, is a basic cognitive process. Category learning

and using depend on selective attention to category- re-

levant stimulus features (shape, size, etc.) inherent to

objects that are perceptually similar [24,25] or share a

theory of cause and effect [26-28] that serve a common

goal (e.g. [29]). The idea that individual's subjective

experiences and motivations could be used for grouping

objects together has been also considered [30].

Distinguishing between cognitions about one’s affec-

tive reactions (its mental representations) and the affec-

tive reactions per se, it has been proposed that affect

increases the accessibility in memory of semantic con-

cepts and knowledge represen tatio n s of the same valence

([8,19]; see [6]) and also increases the likelihood of

thinking about information related to these concepts.

This increased attention might be reflected in (a) longer

time spent reading and thinking about the emotional

information than neutral ones, (b) better memory for the

information, and (c) therefore greater influence of the

information on judgments to which it is relevant [31].

However, in spite of association between positive affect

with holistic and negative affect with piecemeal

processing strategies, and greater automatic allocation of

attention towards negative information, it seems people

are motivated to maintain their happy mood and to di-

vest themselves of the negative events that give rise to

negative feelings [32]. Consequently, compared to nega-

tive information, time spent reading and thinking about

positive information could be longer.

Based on these assumptions, it has been proposed that

the objects and events that elicit the same emotion in a

given perceiver lead to a mental grouping of those ob-

jects and events, as instances of the same category. Re-

search by Niedenthal and her colleagues [10,16,30,33,

34] bear on this matter. Accordingly, emotions could

lead individuals to reorganize temporarily their concep-

tual space in function of the common evoked emotion.

Thus, a category of objects and events that have elicited

same emotional state (e.g. sadness) may be treated as

equivalent and categorized as the same sort of the things,

regardless of their perceptual, functional or theory dif-

ferences. Thereby, perceivers’ understanding of the mean-

ing of a particular category exemplar would be facili-

tated in terms of their own personal learning histories.

This reorganization then determines how people perce-

ive similarities and differences among objects and events.

In the same way, the autho rs suggested that those differ-

ences in processing emotional-relevant information

could be more noticeable in individuals with specific

emotional trait (e.g. anxious, depressives at clinical or

non-clinical level; [16 ]).

Nevertheless, although there is a good deal of evi-

dence regarding affective-congruency of processing posi-

tive information, these congruency effects are not ob-

served for negative information [35,36]. In a study Nie-

denthal, Halberstadt and Innes-Ker [30] used a categori-

zation task (45 triads (a principal concept and two other

concepts) of happy (9), sad (9), and neutral words (27))

in which participants were induced or not (control con-

dition) to feel either happy or sad by having them

watching a film for 12-15 minutes prior to a categoriza-

tion task, so called triad task. Then while listening a mu-

sic of the same valence, they performed a triad task in

which a number of positively (related to happiness e.g.

puppy, celebration, etc.), negatively (related to sadness

e.g. cancer, divorce, etc.) valenced triads of words or

neutral ones were presented. The results showed that the

participants in emotion al states (happy or sad) compared

to those in a con trol condition, used more frequently the

happy concepts as basis for their categorization deci-

sions. But whatever the valence of their emotional state

(happy, sad or neutral), they all responded in the same

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1351

way to the sad concepts presented in the absence of

happy ones (no differences related to emotional states).

According to the authors, the sad concepts, when pre-

sented in the absence of the happy ones, (a) either sug-

gested sadness as a basis for categorization, (b) or al-

tered the emotional state of control participants. An

eventual (c) semantic priming effect has been suggested

too. In this sense, according to the authors the experi-

mental affective manipulation could prime generic

knowledge about emotions [8]. Therefore, it would be

the concepts or mental representations of emotion and

not experience of emotion which influence emotional

categorization of the negative information.

If generic knowledge about emotion is contextually

(emotion induction procedure) or chronically (typical

emotional reactivity) primed, it might be such a priming

and not the induced emotional state itself or emotional

experience which enhances the tendency for individuals

to group together concepts on the basis of their emotion-

al equivalence, specifically in the case of negative va-

lenced information. So, it is possible to propose that

negative informatio n captures au to matically th e atten tion

of perceivers [37], and is processed in a specific manner.

Consistent with this idea are information-processing

models of fear [involving activation of the amygdale,

[38] according to which the emotional significance of

incoming information is pre-consciously assessed by a

threat evaluation system [39,40]. Information judged to

be threatening is then treated in priority [41], and per-

haps without any retrieval of emotional experience.

Indeed, one of the most promising areas in research on

emotions is work on the underlying neural substrates

involved in processing emotional information. More

interesting, recent evidence revealed preferential use of

the left or right visual hemifield affecting everyday be-

havior (e.g. activities such as searching for food, agonis-

tic responses, or escape from predators) in natural envi-

ronment of a variety of species (from fish to mammals;

(see [42]). In human, in addition to well-established hy-

pothesis regarding speech production as a left hemis-

phere task, lateralization of emotion processing is an

ongoing debate. However, processing different aspects

of emotional information seems to involve differential

activation of the left or right hemispheres [43]. Based on

comparison between right brain damaged patients and

intact subjects, a number of investigators have found

right hemisphere specialization for the identification of

emotional words and sentences [44], especially negative

ones [45]. Consistent evidence is the faster reaction

times when participants were expo sed to negative words

presented in the left visual field. For positive words,

reaction times were faster when presented in the right

visual field ([46]; see also [47]).

Other studies provide some evidence for cerebral he-

misphere asymmetries in categorization processes [48].

It has been shown that the typicality of instances had a

large effect on categorization times in the left visual field,

suggesting that the right hemisphere relies strongly on a

prototypical-based comparison strategy [49]. The left

hemisphere seems to be able to categorize on the basis of

exemplar-based category knowledge [50,51].

Interaction between affect and cognition as two modes

of psychological functioning appears, therefore, to be

linked to hemispheric asymmetries of the processing of

emotional-relevant information, which is according to

Davidson [13] more functional than structural [13].

Nevertheless, the organization and localization of affec-

tive treatment remains to be firmly established. Within

this general area, the study of cerebral lateralization and

individuals’ affective-based differences is of particular

importance.

Some of the inconsistencies and controversies that

mark the literature regarding asymmetrical processing of

emotional information could be explained in terms of

individual differences. Individuals’ affective-based dif-

ferences may influence asymmetric processing of infor-

mation in different ways. For example, association of

different temperamental styles with asymmetric activity

of frontal cortex has been suggested (see [43]). Com-

pared with individuals displaying higher left prefrontal

activation, those with higher activation in right prefron-

tal regions report more dispositional negative affect.

Some studies demonstrated that depression and impul-

sivity were associated with a decreased activation in the

left-prefrontal region [52,53], whereas anxiety was asso-

ciated with an increase in the right prefrontal region [54].

Experimental studies have shown that alexithymia was

common among patients with a right-hemisphere lesion

[55,56].

Although, alexithymia was initially identified as a

syndrome for clinical patients, recent works revealed

that its features might be observed in nonclinical popula-

tions, too. And in spite of increasing number of perso-

nality traits identified as being involved in processing

emotional information [55,57-61 ], alexithymia offers the

most interesting condition to analyze relation between

emotional aspects of personality and asymmetrical

processing of emotional information in terms of psycho-

logical functions in clinical as well as nonclinical popu-

lations.

Recent works focused on alexithymia have proposed

different hypotheses to explain it in structural and func-

tional terms (see [23]). In structural terms, one neural

hypothesis suggests that alexithymia is related to poor

interhemispheric transfer, in normal as well as clinical

populations [21,23,62]. Another hypothesis associates it

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1352

with a poor right hemisphere activity [22]. Actually, stu-

dies in psychiatric field have revealed, in patients and

some normal subjects, an interhemispheric cerebral tran s-

fer deficit [63-65]. However, neuropsychological studies

have shown that patients with brain damage in the right

hemisphere perform worse than patients with brain dam-

age in the left hemisphere when processing and organiz-

ing emotional experience [66,67]. In functional terms,

alexithymia represents a unique personality trait that

might predispose individuals to somatic and mental dis-

orders [68]. The psychological studies have highlighted

that alexithymia is correlated with both anxiety and de-

pression [68,69]. It has also been found that alexthymic

individuals use more emotional words, specifically the

negative ones. These underline the possibility th at alexi-

thymia might be a cop ing or defense mechanism utilized

for life-threathening situations [68], and thus invo lved in

processing of negative information.

However, alexithymic features are also associated

with psychopathology and certain neurological syndr o mes ,

such as multiple sclerosis (MS; demyelinating disease of

the central nervous system disrupting neural transmis-

sion). The case of MS is of particular interest to re-

searchers because of its close neurological and psy-

chopathological correspondence to those resulted from

alexithymia in general. Alexithymia is more frequently

reported in MS patients (50%) compared to general pop-

ulation (8%) [70]. Cognitive and emotional disturbances

in the early stages of MS have been frequently described

in terms of high frequency of affective incontinence and

lability. Although, these abnormalities are not directly

related to depressive states [71], depression and anxiety

are frequently reported in the case of MS (The Goldman

Consensus Group, [72,73]). Explosive laughing or cry-

ing, and euphoria are also observed in individuals with

MS [74]. Such variation provides a potentially rich and

relatively untapped source of information enabling

comparison of clinical and non-clinical populations

based on relationships between emotional aspects of

personality and asymmetrical processing of emotional

information.

2. PRESENT STUDY

The primary aim of the present study was to examine

lateralized cerebral processing in a triad task as a func-

tion of individuals’ affective-based characteristics. Our

second objective was to evaluate the exten t to which the

pathological deficits in processing emotional informa-

tion might interfere in lateralized cerebral processing of

emotional-relevant information. Because of their partic-

ular emotional and neurobiological characteristics, the

patients suffering from multiple sclerosis were chosen

for our comparative study of emotional information

processing.

Many studies reported for these patients an emotional

profile marked by affective lability, with high prevalen ce

of depression (60%) [75,76] and anxiety (37%) [77]. As

mentioned before, compared to general population, ale-

xithymia is 6 times more frequent in MS patients. Some

studies demonstrated dysfunction of interhemispheric

transfer (callosal dysfunction) in MS patients, too [70].

In general, this kind of functional impairment is corre-

lated to the degree of corpus callosal atrophy and the

severity of diffusion of white matter changes identified

by magnetic resonance imaging (MRI). Few studies

suggest the potential clinical value of callosal involve-

ment and alexithymia and its interest as a model to study

interhemispheric disconnection (e.g. [23,70]), specifi-

cally in MS patients.

Usually, impaired information processing in individu-

als diagnosed with multiple s clerosis is mild or moderate,

and mainly affects working memory and processing time,

through which inhibition and interpretation capacities

[78]. The causal mechanisms of these affective-cognitive

i mpaired relationships are not w ell known. Neuroan a to mi c

account suggests that both gray and white cerebral mat-

ter atrophies contribute to neuropsycho logical deficits in

MS [79]. The role of atrophied corpus callosum leading

to intra- and inter-hemispheric disconnection has been

also suggested. Considering the studies published to date,

which have reported specific affective and cognitive

psychological impaired processes associated with this

disease, and the importance of these processes in emo-

tional categorization involving working memory, we

decided to assess and compare the performance of indi-

viduals with multiple sclerosis with a non-clinical popu-

lation. We believe that comparative stud ies of such cases

might provide new perspectives on the relation between

cognitive neuropsychology and social-personality psy-

chology.

In the case of the present study we address the ques-

tion: How individuals’ emotional characteristics could

influence their emotional response categorization (group-

ing together stimuli, e.g. words, related to the emotion of

the same valence), and if these influences could be asso-

ciated with hemispheric asymmetries. More precisely we

predicted an association between higher frequencies of

emotional choices and shorter time decisions for left

sided presentation of negative emotional words for indi-

viduals with higher scores on anxiety, depression, and

alexithymia scales. Our second hypothesis was focused

on comparison of the data from non-clinical individuals

to those of MS patients. Because of their psychological

and neuropsychological profiles, and also their impair-

ment in processing emotional-relevant information, we

predicted higher frequencies of choices and shorter time

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1353

decisions for right sided presentation of neutral words

considered as emotional in MS patients compared with

non-clinical individuals. However, regarding emotional

words, lower frequencies of emotional choices and faster

time decisions were predicted for MS patients, compared

with non-clinical individuals specifically in the case of

left sided presentation of negative words.

2.1. Method

2.1.1. Participant s

2.1.1.1. Non-Clinical Population

Eighty right-handed (male n = 34 and female n = 46)

native French-speakers aged between 20 to 59 years (M

= 33.43 years, SD = 9.67 years) were recruited by an-

nouncements published in local newspapers to partici-

pate in a paid study.

Upon arrival at the laboratory, a brief description of

the experimental procedure was given. The subject was

then asked to complete a consent form and questionnaire

concerning medications and drug use. The subjects were

excluded if they were left-handed, had used drugs, or

psychoactive medi cati on wi t hin the past six m ont hs.

2.1.1.2. Clinical Population

Forty right-handed native French-speakers (male n =

15 and female n = 25) aged between 21 to 55 years (M =

37.84 years, SD = 8.38 years) with clinically relaps-

ing-remitting multiple sclerosis less than 5 years dura-

tion (M = 3 years, SD = 1.54) were selected. They parti-

cipated after providing their informed consent approved

by the ethical committee. Multiple Sclerosis was diag-

nosed by a physician, applying the New McDonald Cri-

teria. New diagnostic criteria from the International Pan-

el of McDonald and colleagues incorporate MRI evi-

dence of dissemination in time and space to allow a di-

agnosis of MS in patients with clinically definite and

remitting syndromes with or without inflammatory dys-

function in the central nervous system (CNS). For five

patients the functional scores on the Kurtzke expanded

disability status scale ([80]; EDSS) were between 4.5-

6.0, without any aggravation (mean follow-up 3 months).

Patients with moderate and severe cognitive impairment

were excluded after assessment with the BNI [81]. The

selected MS participants were then compared with 40

participants (male n = 14 and female n = 26) selected

among our non-clinical sample based on their autobio-

graphical profile (gender, age, education, and so-

cio-economical background).

2.1.2. Stimuli

The stimuli were triads (see Annex) of concepts com-

posed of a target and two comparison concepts. In each

triad, one of the two comparison concepts was related to

the target concept through an emotional (positive or

negative) association, whereas the other one was related

to the target concept through a non-emotional, taxono m-

ic association. Three categories of triads (neg ative, posi-

tive, and neutral), with nine triads per category were

used. Some triads were extracted from material used by

Niedenthal, Halberstadt, and Innes-Ker [30]. Other triads

were created for this study following the procedure de-

scribed by Niedenthal and her colleagues [30].

2.1.3. Material Check

An initial pilot test (n = 30 participants who did not

participate in the main study) on a large set of potential

triads (88 concepts) was conducted to ensure that partic-

ipants chose a variety of responses for each triad (at least

20% and no more than 80% of pilot participants select-

ing each possible response).

2.1.4. Measure of Individual Characteristics:

Questionnaires and Scales

In the present study, we used a battery of self-report

questionnaires that measure emotion-related personality

characteristics. As moderator variables, anxiety, depres-

sion, and Alexithymia have been known to be emotion-

related personality traits relevant to cognitive perfor-

mance, but also as characteristics of individuals with

multiple sclerosis. Therefore, all participan ts were admi-

nistered the Trait-Anxiety Inventory [82], Beck Depres-

sion Inventory [83], and Toronto Alexithymia Scale [84].

All three are self-report instruments intended to assess

the existence, intensity or severity of the felt anxiety,

symptoms of depression, and difficulties in identifying,

describing, and communicating emotional feelings in clin-

ical and non-clinical populations. High scores on theses

scales are indicatives of anxietous, depressive and Alex-

ithymic tendencies.

The standardized version of the Trait-Anxiety Scale

contains 15 items. For each item, the respondent was

asked to indicate on a 4-point scale from “never” to “al-

ways” how frequently s/he would feel in that way. The

standardized version of the Beck Depression Inventory

is a 21-item self-report instrument intended to assess the

existence and severity of symptoms of depression. Each

of the 21 items corresponding to a symptom of depres-

sion is summed to give a single score. There is a four-

point scale for each item ranging from 0 to 3. On two

items (16 and 18) there are seven options to indicate

either an increase or decrease of appetite and sleep. The

revised-version of the Toronto Alexithymia Scale con-

sists of 20 items, and a high score on this scale corres-

ponds to difficulty to distinguish between (identification),

to describe (description), and to express (ex teriorization)

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1354

one’s emotions. It was constructed after a literature re-

view revealing 5 main content areas thought to reflect

the construct. However, factorial analysis of the finalized

20-item version of the TAS suggested a three- factor

solution. First factor consisted of items that refer to the

ability to identify and describe feelings and to distin-

guish between bodily sensations, secondary factor re-

flected the ability to communicate feelings to others, and

third factor represented the tend ency to focus on extern al

events over inner experiences. For each item, the res-

pondent was asked to indicate on a 5-point scale ranging

from “strongly disagree” to “strongly agree” the extent

to which s/he was agreed with statement. The TAS has

shown adequate internal consistency, good test- retest

reliability, and good convergent and discriminant validi-

ty [85]. The TAS and TAS-20 are now the most widely

used measures of alexithymia [86].

2.1.5. Apparatus and Response Measurement

for the Experimental Task

The instructions, presentation of the triads, collection

of subject's responses and response times were con-

trolled by a Pentium IBM compatible computer with an

IIyama 15 inch vision master 404, 50/60 hz monitor. The

time between presentation of each triad and the lexical

decision by subjects was recorded as decision time.

2.1.6. Experiment al Task

Each trial consisted of simultaneous presentation of

three concepts in an equilateral triangular shape. The

centered target word was presented 8 cm from the top of

the screen, with two alternative response words dis-

played 8 cm below the target. Two comparison concepts

were presented 14.5 cm apart from each other. The con-

cepts were presented in 24 p ixel MS serif small letter. A

fixed, random order of triads (positiv e, negative, n eutral)

was used. The left or right sided presentation of emo-

tional terms was counterbalanced across trials for each

category of triad s.

After three practice trials, all the participants received

27 test trials. Each triad was displayed by the computer

until the subj ect made a d ecision. The particip an ts used a

9 cm × 20 cm × 4.20 cm response box with two keys

placed 5.5 cm apart on the board corresponding respec-

tively to the position of the index of right an d left hands

on an outstretched bras. They were seated 70 cm away

from the screen, and 30 cm away from the response box.

They were encouraged to keep the palm of the hand on

the table, making discrete index finger movements only.

The participants were asked to fixate on a central point

presented at the center of the screen when each triad was

removed. For each trial, the computer began to display a

fixation point at the center of the screen, and after 500

ms a triad was displayed. The participants were in-

structed to press with their left index finger on the left

key if they thought the target concept was most similar

to the concept displayed on the left and with their right

index finger on the right key in the other case. The next

trial was initiated only after the subject responded. Res-

ponses and response times were record ed an d encod ed in

terms of the valence (positive, negative, neutral) and

presentation-side (left/right) of the emotional informa-

tion. Then, mean frequencies (number of valenced emo-

tional decisions for each presentation-side divided by

total number of negative or positive triads per presenta-

tion-side) and mean time of decisions (sum of valenced

emotional decision times for each presentation-side di-

vided by number of valenced emotional decisions for

each presentation-side) were computed and analyzed.

2.1.7. Procedure

Participants were told that the purpose of the experi-

ment was to assess their decisions about different con-

cepts. After entering a sound-attenuated room, the par-

ticipants were invited, by experimenters, to sit on a chair

facing a computer monitor on which a uniformly gray

image was presented and to complete a questionnaire. If

the subjects indicated no recent psychoactive medica-

tions they would complete a consent form and read the

instructions displayed on the computer. The instructions

indicated that the experiment concerned participant's

perception of lexical similarity. Participants were in-

structed to view each triad on the screen and to use the

response box by pressing the appropriate key to transmit

their choice. Participants were also instructed that an

initial block of three trials was designed to aid habitua-

tion to the experimental task and experimental process.

After presentation of all the triads, participants com-

pleted a number of self-report scales intended to measure

emotion-related individual differences (Trait-Anxiety

Scale, Beck Depression Inventory, Toronto Alexithymia

Scale, and autobiographic questionnaire). Finally, par-

ticipants were debriefed, thanked, and paid.

3. STATISTICAL ANALYSIS

For both samples, descriptive statistics were calculated

for each of the variables and Pearson’s correlation coeffi-

cients were used to investigate the relationship between

the individuals’ differences and cognitive performances.

To test the hypothesis that processing affective informa-

tion could be related to hemispheric asymmetries, we

conducted analyses of variance in which we examined

the relationships between valence and presentation-side

of the emotional information with frequency and time of

categorical decisions for both non-clinical and clinical

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1355

populations. Regression and variance analyses were used

in order to examine the effects of age and gender on the

emotional scores and cognitive performances. All data

were analyzed with Statistica and SPSS for Windows.

4. RESULTS

4.1. Individual Emotional Differences

Correlations computed between different measures of

individuals’ emotional characteristics showed that each

characteristic preserved it’s specificity, at least partly,

reflecting the interest of taking into account these meas-

ures for evaluation of their impact and/or relationships

with cognitive performances. The results showed that the

level of alexithymia was significantly and positively

correlated with the level of anxiety and depression (r

anxiety = 0.34; r BDI = 0.28, p < 0.05). These two last

emotional characteristics were also significantly and

positively correlated (r = 0.71, p < 0.05). Although, in-

tercorrelations between global level of alexithymia and

its components were strong (r identification = 0.77; r

description = 0.78; r externalization = 0.58, p < 0.05),

only its identification component was correlated signifi-

cantly with the scores on anxiety and depression (r an-

xiety = 0.48; r BDI = 0.40, p < 0.05).

4.1.1. Emotional Characteristics and Cognitiv e

Performance

Correlations were computed between individuals’

emotional characteristics and categorization decision

scores (frequency and reaction times). The correlations

suggest relations between emotional individual differ-

ences and triad task performance. These relations vary,

however, depending on the side of presentation of emo-

tional terms of each triad suggesting a role of hemis-

pheric organization of processing of emotional-relevant

information. In our sample, the more anxious and de-

pressive participants chose preferentially negative emo-

tional information presented in their left visual space (r

anxiety = 0.29, p < 0.05; r BDI = 0.24, p < 0.05) indi-

cating the right hemisphere implication.

For the neutral triads, participants with the higher

scores of alexithymia chose more frequently the left side

concept. Correlations between measures of alexithymia

and frequency of the left-sided (r TAS20 = 0.26, p <

0.05), and right-sided choices (r TAS20 = –0.26, p < 0.05)

for neutral triads could be accounted for a more left-sided

search than right-sided when individuals with difficulties

to identify and name emotional information try to make

an emotional categorical decision, corresponding to so

called leftward bias in literature [87]. There were no sig-

nificant correlation between individuals’ emotional cha-

racteristics and categorization decision t imes.

Simultaneous multiple regression (see [88], for details

about the regression procedure) of the emotional catego-

rization (frequencies and decision times) on the predic-

tors (age, gender: male = 0, female = 1, anxiety, depres-

sion, and components of alexithymia) revealed only sig-

nificant contributions of the global alexithymia scores

and its identification component for frequencies related

to neutral triads (R² = 0.12, F(5,74) = 1.92, p < 0.10;



–0.30, t = –2.51, p < 0.02) and for decision times rela-

tive to left-sided neutral triads (R² = 0.15, F(7,72) = 1.79,

p < 0.10;



= –0.42, t = –3.18, p < 0.003).

In sum, frequencies of the categorical decisions for

neutral triads were lesser and also decision times for

left-sided neutral triads were longer when participants

with difficulties to processing emotional information,

specifically identifying them, tried to make a categorical

decision. The lengthened decision times for left-sided

neutral triads corresponds to the right hemisphere hy-

pothesized-error detection function predicted by Smith

and colleagues [89].

Although non-significant, the contribution of the age

and gender variables on variance of the categorical deci-

sions found to be dependent on the valence of the triads.

In general, men’s decision times for emotional triads

were longer whatever their hand-side, but they were cho-

sen more frequently by women when they were presented

on the right side. However, negative left-sided triads were

chosen by younger men, and positive ones by younger

women. Regarding neutral triads, the decision times for

left-sided triads were longest for older women, and

shortest for men’s right-sided decision. Frequencies of

left-sided neutral triads cho sen as emotional were h ighest

in younger women, but frequencies of right-sided choic es

were highest in older men.

4.2. Non-Clinical Population

4.2.1. Emotional Characteristics

Separate analyses of individual differences for anxiety,

depression, and alexithymic scores for the non-clinical

sample were computed. The results showed no significant

effects except for externalization involving sex of the

subjects, F(1,78) = 4.42, p < 0.04; f = 0.48. Consistent

with other published data (e.g. [90]), compared with

women, men demonstrated more difficulties for express-

ing their emotions (M = 2.12, SD = 0.52 VS M = 1.89, SD

= 0.44).

4.2.2. Emotional Categorization

4.2.2.1. Frequency of Emotional Choice1

A 2 × 3 × 2 (Sex of the subjects × Valence of the tri-

1Note: A separate analysis of variance of categorization decision and

decision times was conducted for neutra l t riads (Table 1). The results

showed no si

nificant effect due to

artici

ant’s sex.

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1356

ads × Hand-side) ANOVA was used in which two latter

factors were treated as repeated measures.

Analysis of variance for frequency of emotional

choices revealed significant main effects of sex of the

subjects, F(1,78) = 4.64, p < 0.04; f = 0.48, valence of

the triads, F(2,156) = 17.11, p < 0.001; f = 0.66, and

hand-side, F (1,78) = 9.44, p < 0.003; f = 0.70. The sig-

nificant main effect of valence was due to the more fre-

quent choice of the su bjects for positive rather than neg-

ative emotional responses (M positive = 0.598 VS M

negative = 0.495 VS M neutral = 0.500). The main effect

of the sex of the subjects corresponds to a more frequent

emotional choice for wo men compared to men (M w om en

= 0.550 VS M men = 0.510). Effect of the hand-side

showed more left categorical choice than right (M left =

0. 56 1 VS M right = 0.500).

A two-way in teraction associating v alence of the triads

with hand-side, F(2,156) = 3.13, p < 0.05; f = 0.28, was

found, too. Inspection of the data showed most frequent

left-side emotional choice for positive triads and less for

negative ones (M positive = 0.657 VS M negative = 0 .505

VS M neutral = 0.522). For right-side emotional choice,

most frequent choice was also found for positive triads,

but less for neutral ones (M positive = 0.538 VS M nega-

tive = 0.484 VS M neutral = 0.478). Thus, emotional

choices were more frequent for positive triads, particu-

larly in the case of the left-sided presentations.

Separate analyses of the data for emotional triads

showed same significant effects and revealed modulation

of categorization of emotional information by valence

and side of presentation of the stimuli, particularly for

positive triads, F(1,78) = 5,56, p < 0.02; f = 0.53. The

most frequent choice wa s observed for positive triad wh en

emotional concept was presented at the left side, and the

least for negative triad with emotional co ncept presented

on the right side (see Table 1).

Analyses of the data taking into account subjects’ dif-

ferences for anxiety, depression, and alexithymic (identi-

fication and description) scores as covariable revealed

the same significant effects, except for externalization.

Indeed, after integration of externalization score as co-

variable the main significant effect of gender disap-

peared, without an y consequence for the o ther effects.

4.2.2.2. Categorization D ecision Time

Analyses of variance with a 2 × 3 × 2 (Sex of the

subjects x Valence of the triads x Hand-side) factorial

design were performed for categorization decision times

in which two latter factors were treated as repeated

measures. Except for main effect of sex of the subjects,

analyses after logarithmic transformation of categoriza-

tion decision times showed similar main effects and in-

teractions to those fo und for frequency of the catego riza-

tion decisio ns.

Tab le 1. Categorization Decision Frequency and Categoriza-

tion Decision Times as a function of Sex of the subjects,

Triad valence, and Hand-side.

Group/

Triade-Valence

Hand-side

Left Right

Male(n = 34)

Neutral

M 0.546 0.454

SD 0.156 0.156

M 2604a 2254

SD 0.205 0.588

Positive

M 0.603a 0.500

SD 0.223 0.192

M 5835 6549

SD 0.151 0.211

Negative

M 0.500 0.459

SD 0.222 0.194

M 4700 5333

SD 0.141 0.199

Female(n=46)

Neutral

M 0.498 0.502

SD 0.123 0.123

M 2221a 2327

SD 0.834 0.509

Positive

M 0.712a 0.574

SD 0.241 0.222

M 5988 7267

SD 0.173 0.178

Negative

M 0.511 0.509

SD 0.217 0.172

M 4619 6080

SD 0.177 0.151

Note: Categorization decision times are displayed in milliseconds.

Significant main effects of valence of the triads, F(2,

156) = 13.34, p < 0.001; f = 0.59, and hand-side, F(1,78)

= 5.85, p < 0.02; f = 0.55, were found. Careful in spection

of the data for these main effects showed faster decision

times for negative compared with positive triads (M Pos-

itive = 6440 ms VS M Negative = 5196 ms VS M Neu-

tral = 2366 ms), and for left hand categorization deci-

sions (M left = 4339 ms VS M right = 4996 ms).

A two way significant interaction effect was found

involving valen ce of the tr iads and ha nd-side; F(2,156) =

3.58, p < 0.03; f = 0.30. The data are relevant to the hy-

pothesis of differential processing of left- and right

presentations of the emotional information. Asymme-

trical processing of information seems to be related to

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1357

the

Table 2. Categorization Decision Frequency and Categoriza-

tion Decision Times as a function of Hand-side and Triad va-

lence.

Triad valence Hand side

Left Right

Neutral 0.522a 0.478c

2413a 2291a

Positive 0.657ab 0.538cb

5912ab 6909abc

Negative 0.505b 0.484b

4660ac 5707ac

Note: categorization decision tim es are displayed in milliseconds.

Means that share superscripts differ at p < 0.05 in post-hoc LSD tests.

valence of emotional information. In general, as we ex-

pected, for right-handed individuals the data (see Table

2) showed fastest decision times in the case of neutral

triads (M left-neutral = 2413 ms, M right-neutral = 2291

ms). In contrast, emotional decision times were slower

for positive triads; the fastest emotional choices occu rred

for left-side decisions based on negative valenced in-

formation (M left-positive = 5912 ms VS M

right-positive = 6909 ms; M left-negative = 4660 ms VS

M right- negative = 5707 ms). The configuration of re-

sults for emotional decisions supports our hypothesis. In

spite of a general deceleration of the decision times for

positive triads and acceleration for negative triads which

are relevant to a motivational hypothesis [32], these ten-

dencies were more noticeable whenever positive infor-

mation was more accessible to the right hemisphere and

negative information to the left hemisphere.

Analyses of the data in function of subjects’ differ-

ences for anxiety, depression, and alexithymic (identi-

fication, description, and externalization) scores as

covariable revealed the same significant effects.

4.3. Clinical Population

4.3.1. Emotional Characteristics

Analyses of individual d ifferences for anxiety, depres-

sion, and alexithymic scores for clinical sample showed

the same significant gender effect for emotion externali-

zation, F(1,38) = 8.86, p < 0.005; f = 0.97, with more

difficulties of externalization for men (M = 49.33, SD =

10. 7 2 VS M = 66.24, SD = 20.29).

4.3.2. Emotional Categorization

4.3.2.1. Frequency of Emotional Choice

Separate analysis of variance for neutral triads (Table

3) for participants with multiple sclerosis showed a sig-

nificant interaction for categorization decision involving

sex of the subjects and hand-side, F(1,38) = 4.91, p <

0.04; f = 0.72. Inspection of the data showed most fre

Table 3. Categorization Decision Frequency and Categoriza-

tion Decision Times as a function of Group and Sex of the

subjects, Triad val ence, a n d Hand-side.

Group/Triad va-

lence Hand-side

Left Right

Healthy Participants (40)

Neutral

Male (n=14)

M 0.508 0.492

SD 0.188 0.188

M 2178 2219a

SD 0.232 0.184

Positive

M 0.571 0.571

SD 0.206 0.205

M 4890 8084

SD 0.186 0.288

Negative

M 0.446 0.557ab

SD 0.223 0.160

M 3926 6340

SD 0.144 0.261

Female (n=26)

Neutral

M 0.496 0.504

SD 0.114 0.114

M 347 2319a

SD 0.164 0.154

Positive

M 0.673 0.577

SD 0.262 0.206

M 5915ab 7613

SD 0.202 0.194

Negative

M 0.481a 0.538

SD 0.199 0.176

M 4330 6715a

SD 0.195 0.173

Participants with Multiple Sclerosis (n=40)

Male (n=15)

Neutral

M 0.422 0.578

SD 0.086 0.086

M 3034 2188a

SD 0.162 0.169

Positive

M 0.667 0.537

SD 0.204 0.242

M 7440a 9439

SD 0.476 0.923

Negative

M 0.317a 0.373a

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1358

SD 0.291 0.237

M 4201 6280

SD 1.701 1.233

Female (n=25)

Neutral

M 0.511 0.489

SD 0.140 0.140

M 2265 2460a

SD 0.175 0.199

Positive

M 0.640 0.640

SD 0.315 0.252

M 5635ab 9268

SD 0.931 0.191

Negative

M 0.410 0.344ab

SD 0.269 0.147

M 4192 4820a

SD 1.376 0.720

Note: categorization decision times are displayed in millseconds.

Means in the same row that do not share superscripts differ at p<.05

in post-hoc LSD tests.

quent right-side and less left-side choices for men (M

left-side = 0.422 VS M right-side = 0.578; p < 0.02).

Reverse tendencies were found for women (M left-side =

0.511 VS M right-side = 0.489). It seemed false alarm

were significantly more frequent for men in the case of

the right-sided information, whereas women did the same

for left-sided ones.

Analysis of variance (Sex of the subjects 2 × Valence

of the triads 3 × Hand-side 2) for frequency of emotional

choices revealed significant main effects of valence of

the triads, F(2,76) = 22.72, p < 0.0001; f = 1.10, which

was due to the more frequent choices for positive or

neutral rather than negative emotional responses (M pos-

itive = 0.621 VS M negative = 0.361 VS M neutral =

0.500).

A three-way interactio n associating sex of the subjects

with valence of the triads and hand-side, F(2,76) = 3.18,

p < 0.05; f = 0.41, was found, too. Inspection of the data

revealed for men most frequent left-side emotional

choices for positive triads and less frequent emotional

choices for negative ones (Positive: M left-side = 0.667

VS M right-side = 0.537; Negative: M left-side = 0.317

VS M right-side = 0.373. Neutral: M left-side = 0.422.

VS M right-side = 0.578). For women, we did not find

any differences related to hand-side for positive triads.

However, women made relatively more categorical deci-

sions for left-sided negative and neutral information

(Positive: M left-side = 0.640 VS M right-side = 0.640;

Negative: M left-side = 0.410 VS M right-side = 0.344.

Neutral: M left-side = 0.511 VS M right-side = 0.489).

Thus, emotional choices were more frequent for positive

triads, particularly for men in the case of the left-sided

presentations, in volving right hemisphere activation. For

negative triads, men’s left-sided and women’s right-

sided were least choices.

Analyses of the data in function of subjects’ differ-

ences for anxiety, depression, and alexithymic (identifi-

cation, description, and externalization) scores as cova-

riable revealed the same significant effects.

4.3.2.2. Categorization D ecision Time

Analyses of variance performed for categorization de-

cision times after logarithmic transformation of the data

showed significant main effects of valence of the triads,

F(2,76) = 9.37, p < 0.001; f = 0.70, and hand-side,

F(1,38) = 6.45, p < 0.02; f = 0.82, and also a significant

interaction effect involving same variables, F(2,76) =

5.17, p < 0.01; f = 0.52. Inspection of the data for these

main effects showed faster decision times for negative

compared with positive triads (M Negative = 4874 ms

VS M Positive = 7946 ms VS M Neutral = 2487 ms),

and for left hand compared with right hand categoriza-

tion decisions (M left = 4461 ms VS M right = 5743

ms).

A significant interaction effect involving valence of

the triads and hand-side was observed concerning the

hypothesis of asymmetrical processing of information

dependent on the valence of emotional information. In

general, as we expected, for right-handed individuals the

data (see Ta bl e 4) showed fastest decision times in the

case of neutral triads (M left-sided = 2649 ms, M right-

sided = 2323 ms). For emotional decision times, right-

sided decision were the slowest, specifically in the case

of positive triads (M left-positive = 6538 ms VS M

right-positive = 9353 ms; M left-negative = 4197 ms VS

M right-negative = 5550 ms). The data’s directions for

emotional decisions was similar to that of the general

population in the present study, specifically when posi-

tive information was more accessible to the right he-

misphere and negative information to the left hemis-

phere.

Analyses of the data in function of subjects’ differ-

ences for anxiety, depression, and alexithymic (identifi-

cation, description, and externalization) scores as cova-

riable revealed the same significant effects.

4.4. Comparative Analyses

In order to compare performance of participants

with multiple sclerosis with the general population,

we conducted a separate analysis comparing 40 con-

trol participants matched on sex, age, and so-

cio-economical profile with those participants with

multiple sclerosis.

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1359

4.4.1. Emotional Characteristics

Analyses of individual d ifferences for anxiety, depres-

sion, and alexithymic scores comparing healthy partici-

Table 4. Categorization Decision Frequency and Categoriza-

tion Decision Times as a function of Group of the subjects,

Hand-side, and Triad valence.

Group/Triad va-

lence Hand-side

Left Right

Healthy subjects

Neutral 0.502 0.498

2262 2269

Positive 0.622 0.574

5402bc 7848abc

Negative 0.464ab 0.548ab

4128bc 6528ac

Participants with Sclerosis

Neutral 0.467 0.533

2649 2324

Positive 0.653 0.588

6538a 9354ab

Negative 0.363a 0.359b

4197a 5550ab

Note: Categorization decision times are displayed in milliseconds.

Means that share superscripts differ at p < 0.05 in post-hoc LSD tests.

pants with those with multiple sclerosis showed signifi-

cantly higher levels of anxiety, F(1,76) = 8.43, p < 0.005;

f = 0.67; M control = 37.51 VS M sclerosis = 44.35, and

alexithymic scores, F(1,76) = 16.09, p < .0001; f = 0.56;

M control = 43.38 VS M sclerosis = 52.31, for the clini-

cal group. Concerning alexithymic values, although in

both groups there were no differences between men and

women in terms of difficulties to identify emotions, F(1,

76) = 20.31, p < 0.0001; f = 1.04; M control = 14.01 VS

M sclerosis = 19.25, for externalization of emotions in

addition to main effects related to participants’ sex and

health, a significant two-way interaction involving both

variables is found, F(1,76) = 10.69, p < 0.002; f = 0.75.

Compared to women, men of the non-clinical

sub-sample experience more difficulties to express their

emotion (M men = 17.14, SD = 3.40 VS M women =

14.92, SD = 3.53). A reversed trend is observed for the

clinical group: Women more than men find difficult to

express their emotions (M men = 49.33, SD = 10.71 VS

M women = 66.24, SD = 20.29). No significant main

effects or interaction were found for the scores on de-

pression.

4.4.2. Emotional Categorization

4.4.2.1. Frequency of Emotional Choice 1

For emotional triads, a 2 × 2 × 3 × 2 (Sex of the sub-

jects × Group of the subjects × Valence of the triads ×

Hand-side) ANOVA was used in which two latter factors

were treated as repeated measures.

Analysis of variance for frequency of emotional

choices revealed significant main effects of valence of

the triads, F(2,152) = 27.41, p < 0.001; f = 0.85, and its

interaction with group of the subj ects, F(2,15 2) = 7.14, p

< 0.002; f = 0.43. The significant main effect of valence

was due to the most frequent choices of the subjects for

positive rather than negative emotional responses (M

positive = 0.610 VS M negative = 0.433 VS M neutral =

0.500).

A two-way interaction associating valence of the tri-

ads with group of the subjects, corresponding to the most

frequent emotional choices for positive triads and least

for negative ones was observed in the case of the pa rtic-

ipants with multiple sclerosis (M positive = 0.621 VS M

negative = 0.361 VS M neutral = 0.500) compared to

healthy ones (M p ositive = 0.598 VS M negative = 0.506

VS M neutral = 0.500). In fact, differences between

non-clinical and clinical participants were significant

only for negative choices, F(1,76) = 13.27, p < 0.0005; f

= 0.82, M control = 0.506 VS M sclerosis = 0.361.

4.4.2.2. Categorization D ecision Time

A 2 × 2 × 3 × 2 (Sex of the subjects x Group of the

subjects x Valence of the triads x Hand-side) factorial

analysis was performed for categorization decision times

in which the two latter factors were treated as repeated

measures. Analyses after logarithmic transformation of

categorization decision times showed in addition to the

main effects of valence of the triads, F(2,152) = 8.83, p

< 0.0002; f = 0.48, and hand-side, F(1,76) = 7.59, p <

0.008; f = 0.63, two two-way interactions, F(2,152) =

9.47, p < 0.0001; f = 0.50; F(1,76) = 5.06, p < 0.03; f =

0.52, and a three-way, F(2,152) = 5.23, p < 0.007; f =

0.35, involving gr oup of the subjects and the two former

variables. Inspection of the data showed longest time

decisions for positive information (M positive = 7286 ms

VS M negative = 5101 ms VS M neutral = 2376 ms),

right-sided information (M right-side = 5645 ms VS M

left-side = 4196 ms), specifically for participants with

multiple sclerosis in terms of valenced information (Mul-

tiple sclerosis: M positive = 7946 ms VS M negative =

4874 ms VS M neutral = 2187 ms; Control: M positive =

6625 ms VS M negative = 5325 ms VS M neutral =

2265 ms), or its presentation side (Multiple sclerosis: M

right-side = 5743 ms VS M left-side = 4461 ms; Control:

M right-side = 5548 ms VS M left-side = 3931 ms).

A three-way interaction showed that compared with

non-clinical population those from clinical sample took

more time to make their decisions in th e case of the pos-

itive and neutral triads, with the longest decision times

for right-sided positive information (Right-side: M con-

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1360

trol = 7848 ms VS M sclerosis = 9353 ms; Left-side: M

control = 5402 ms VS M sclerosis = 6537 ms), and the

fastest time decisions for right-sided neutral ones

(Right-side: M control = 2268 ms VS M sclerosis =

2323 ms; Left-side: M control = 2262 ms VS M sclerosis

= 2649 ms). The most interesting effect was observed in

the case of negative triads. Indeed, although there were

no differences between healthy participants and multiple

sclerosis patients when processing left-sided negative

information (M control = 4127 ms VS M sclerosis =

4197 ms), in the case of the right-sided negative infor-

mation participants with multiple sclero sis took less time

to make their emotional decision (M control = 6527 ms

VS M sclerosis = 5550 ms, see Table 4).

Thus, although participants with multiple sclerosis

needed much more time to make their decision about

categorization of emotional information, when facing

negative information they were relatively rapid. This

effect could be interpreted as a consequence of their

disease and the worries it generates, in which case a

negative correlation between anxiety or depression

scores and frequencies and time decisions for negative

information would be expected. Nevertheless, the

results of our study reveal rather positive and non sig-

nificant correlations between these variables.

5. DISCUSSION

Our data are consistent with previous studies, showing

a) right hemisphere specialization for the categorization of

emotional concepts (e.g. [44]), especially negative ones

(e.g. [45]) in general population and in a sample of pa-

tients suffering from multiple sclerosis. In addition, our

data provide b) evidence for asymmetrical processing of

emotional information and bring some methodological

insights related to measurement of negative-valenced

information processing, suggestion correspondence be-

tween valenced information (positive/negative) and their

presentation-side (left/right). As a matter of fact, the lack

of differential processing of negative information in the

absence of positive ones in some studies could be ex-

plained in terms of the lack of those types of correspon-

dence (e.g. [30,35,36]). Nevertheless, they show, for

both populations, c) no gender-related differences re-

garding individuals’ emotional characteristics, specifi-

cally in terms of alexithymia features (e.g. [56]).

On the basis of these findings our hypothesis that

emotion-related individual differences could modulate

processing of affective information is partially supported.

In accordance with previous studies, analyses show that

dimensions relevant to emotional disorders are generally

associated with asymmetrical categorization, with slow

left-sided decisions, particularly in the case of negative

information. Anxious or/and alexithymic people seemed

to solicit preferentially right hemisphere resources, c au s in g

slower treatment of right-sided negative stimuli, except

in the clinical sample. Indeed, participants with multiple

sclerosis treated right-sided negative information faster

than non-clinical participants. Two different but related

explanations could be proposed to explain this effect.

First, people with multiple sclerosis may rely first on left

hemisphere activation for processing emotional informa-

tion regardless of the valence of the information. Second,

we could imagine that this preferentially activation is

due to their general negative emot i onal stat e.

Our analyses of variance showed that people develop

some preferential processes for categorization decisions,

which result in longer decision times for preferred infor-

mation (positive information, [32]). However, even tho ugh

participants preferred , in general, to us e positive info rma-

tion as the basis of their categorization, they chose more

left-sided than right-sided po sitive emotional alternatives.

At the same time, these left preferential decisions were

faster than right-sided decisions. When participants had

to choose between emotional or non-emotional alterna-

tives they showed the same tendencies, more frequent

and faster left-sided emotional decisions. The processing

of emotional information, specifically positive informa-

tion seemed to be more efficient whenever it was access-

ible to the right hemisphere.

This preferential processing seems to vary depending

on gender of the individuals. Women compared to men

chose more frequently emotional information for making

categorical decisions, but they took more time to make

these decisions. Regarding gender, we noticed exactly

the same tendencies in clinical populatio n inv estigated in

present study.

The main finding of the present study was that there

exists patterns of hemispheric asymmetry that vary sys-

tematically as an interactive function of the valence of

emotional information involved in categorical decisions,

and the spatial presentation of this information (left- or

right side), resulting in more or less accessibility of in-

formation to the left or right hemisphere. Additionally,

the results showed a contribution of certain individual

emotional characteristics to the organization of these

patterns. The cen tral finding regarding hemispheric late-

rality was that the asymmetry emerged from longer

processing for the right side decisions (stimuli/response)

and shorter processing for left side, thus from more ac-

cessibility of the information for the right hemisphere

than the left, except in the case of neutral triads. The

slowest and the fastest decision s were respectively f ound

for positive and neutral right-sided triads, supporting a

motivational hypothesis, according to which people are

motivated to search and preserve their positive feeling s .

Taken together, the results concerning individuals’

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1361

emotional differences and cognitive tasks revealed a

greater implication of the right hemisphere in processing

affectivo-verbal information. This observation favors a

functional conception of hemispheric asymmetry [13],

and provides evidence concerning the dynamic relation-

ship between individuals’ emotional differences (alexi-

thymic features), the nature of information and the spa-

tial context of their presentation. The participants high in

emotion identification and externalization needed more

time to process emotional information presented to the

right visual field (left hemisphere activation).

In the present study, the emotional stimuli were pre-

sented in the both hemi-spaces, and the right hemisphere

dominance was present across both positiv e and neg ativ e

stimuli. All participants with a higher emotional sensi-

tivity (anxiety, emotion id entificatio n and externalization)

seemed to search emotional information in their left vis-

ual space (right hemisphere involvement), and process it

more efficiently. However, those with multiple sclerosis

processed faster but less efficiently (less than 40% of

emotional categorization decisions) right-sided informa-

tion (involving the left hemisphere). These findings

support those of previous studies an d suggest the impor-

tance of integration of individual differences in the

analysis of hemispheric asymmetrical processing of

emotional information. In fact, although the patterns of

asymmetries favor right hemisphere processing of nega-

tive stimuli compared to left hemisphere treatment of

positive stimuli (faster negative left-side decisio n), in the

case of non-clinical participants when comparing deci-

sion times only for emotional triads, the interactive ef-

fect associating valence of triads with their side presen-

tation disappeared. These complex dynamic relation-

ships between dispositional and situational features in

cognitive tasks require more investigation. It may be

valuable for future studies on hemispheric laterality of

emotional information to consider the functions of each

hemisphere depending on emotional states. The present

research suggests the importance of right-hemispheric

activity in processing emotional information as a func-

tion of dispositional characteristics. By understanding

basic processes involved in treating emotional informa-

tion we should be in a better pos ition to predict relation-

ships between emotion and cogn ition.

For example, working memory involved in categoriza-

tion processes seems to have two major levels of

processing in terms of access to encoded information:

maintenance, and manipulation. Irrespective of which of

these two systems are responsible for impaired perfor-

mance in working memory, another mediating variable

influencing working memory performance accuracy is

information processing speed [91]. Several studies have

revealed significant difficulties in information

processing in individuals with multiple sclerosis [92]. In

fact, decreased efficiency in processing speed is a pri-

mary determinant of impaired working memory perfor-

mance. However, a recent research using MRI suggested

that patients, with working memory impairment, at the

early stage of MS may partially compensate by a greater

cognitive control [93]. Also, our results suggest that in-

vestigation of processes which underlay hemispheric

specialization and interaction between emotion and cog-

nition is warranted among subjects with multiple sclero-

sis and may ultimately help us to better understand and

generate novel techniques to provid e a better life-quality

for them.

REFERENCES

[1] Schwarz, N. and Clore, G.L. (1983) Mood, misattribution,

and judgements of well-being: informative and directive

functions of affective states. Journal of Personality and

Soci al Psyc hology, 45, 513-523.

[2] Ellis, H.C. and Ashbrook, P.W. (1989) The “state” of

mood and memory research: A selective review. Journal

of Social Behavior and Personality, 4, 1-21.

[3] Bless, H. and Fielder, K. (1995) Affective states and the

influence of activated general knowledge structures. Perso-

nality and Social Psychology Bulletin, 21, 766-778.

[4] Isen, A.M., Means, B., Patrick, R. and Nowichi, G. (1982)

Some factors influencing decision-making strategy and

risk-taking. In Clark M.S. and Fiske, S.T. Eds., Affect

and cognition: The 17th annual Carnegie Symposium on

Cognition, Hillsdale, NJ: Erlbaum, 243-261.

[5] Pahlavan, F., Bonnet, P. and Duda, D. (2000) Human

motor responses to simultaneous aversive stimulation

and failure on a valued task. Psychological Reports, Per-

ceptual and Motor Skills, 86, 232-242.

[6] Wyer, R. S. (2004) Social comprehension and judgment:

The role of situation models, narratives, and implicit

theories, Erlbaum, New Jersey.

[7] Gray, J.R. (2004) Integration of Emotion and Cognitive

Control. Current Directions in Psychological Science, 33,

46-48.

[8] Bower, G.H. (1981) Mood and memory. American Psy-

chologist, 36, 129-148.

[9] Blaney, P.H. (1986) Affect and memory: A review.

Psychological Bulletin, 99, 229-246.

[10] Niedenthal, P.M. and Setterlund, M.B. (1994) Emotion

congruence in perception. Personality and Social Psy-

chology Bulletin, 20, 401-411.

[11] Niedenthal, P.M., Halberstadt, J. and Setterlund, M.B.

(1997) Being happy and seeing “happy”: Emotional state

facilitates visual word recognition. Cognition and Emo-

tion, 11, 594-624.

[12] Carver, C.S., Sutton, S.K. and Scheier, M.F. (2000) Ac-

tion, emotion, and personality: Emerging conceptual in-

tegration. Personality and Social Psychology Bulletin, 26,

741-751.

[13] Davidson, R.J. (1998) Affective style and affective di so rd er s:

Perspectives from affective neuroscience. Cognition and

Emotion, 12, 307-330.

[14] Salovey, P., Mayer, J.D. and Caruso, D. (2002) The posi-

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1362

tive psychology of emotional intelligence. In Snyder, C.R.

and Lopez, S.J. Eds., The handbook of positive psychol-

ogy, Oxford University Press, New York, 159-171.

[15] Borkenau, P. and Mauer, N. (2007) Well-being and the

accessibility of pleasant and unpleasant concepts. Euro-

pean Journal of Personality, 21, 169-189.

[16] Niedenthal, P.M. and Halberstadt, J.H. (2000). Groun-

ding categories in emotional response. In: Forgas, J.P.

Ed., Feeling and Thinking: The Role of Affect in Social

Cognition. Studies in Emotion and Social Interaction,

Second Series, New York, Cambridge University Press,

357-386.

[17] Manguno-Mire, G. M., Constans, J.I. and Geer, J.H. (2005)

Anxiety-related differences in affective categorizations of

lexical stimuli. Behaviour Research and Therapy, 43,

197-213.

[18] Fox, E. (1996) Selective processing of threatening words

in anxiety: The role of awareness. Cognition & Emotion,

10, 449-480.

[19] Forgas, J.P. (2000). The interplay of affect and cognition:

The mediating role of general knowledge structures. In

Forgas, J.P. Ed., Feeling and Thinking: The Role of Affect

in Social Cognition, Cambridge University Press, Cam-

bridge, 1-28.

[20] Rusting, C. L. (1999) Interactive effects of personality

and mood on emotion-congruent memory and judgment.

Journal of Personality and Social Psychology, 77, 1073-

1086.

[21] Miller, L. (1986) Is alexithymia a disconnection syn-

drome? A neuropsychological perspective. International

Journal of Psychiatry in Medicine, 16, 199-209.

[22] Shipko, S. (1982) Further reflections on Psychosomatic

theory: Alexithymia and Interhemispheric specialization.

Psychotherapy and Psychosomatics, 37, 83-86.

[23] Tabibnia, G. and Zaidel, E. (2005) Alexithymia, interhe-

mispheric transfer, and right hemispheric specialization:

A critical review. Psychotherapy and Psychosomatics, 74,

81-92.

[24] Groom, C.J., Sherman, J.W., Lu, L., Conrey, F.R. and

Keijzer, B.S.C. (2005) Judging compound social catego-

ries: Compound familiarity and compatibility as deter-

minants of processing mode. Social Cognition, 23, 291-

323.

[25] Locke, V., Macrae, C.N. and Eaton, J.L. (2005) Is person

categorization modulated by exemplar typicality? Is per-

son categorization modulated by exemplar typicality?

Social Cognition, 23, 417-428.

[26] Murphy, G.L. and Medin, D.L. (1985) The role of theo-

ries in conceptual coherence. Psychological Review, 92,

289-316.

[27] Smith, E.E. and Sleman, S.A. (1994) Similarity-versus

rule-based categorization. Memory and Cognition, 22,

377-386.

[28] Wittenbrink, B., Hilton, J.L. and Gist, P.L. (1998) In

search of similarity: Stereotypes as naive theories in so-

cial categorization. Social Cognition, 16, 31-55.

[29] Barsalou, L.W. (1983) Ad hoc categories. Memory and

Cognition, 11, 211-217.

[30] Niedenthal, P.M., Halberstadt, J.H. and Innes-Ker, A.H.

(1999) Emotional response categorization. Psychological

Review, 106, 337-361.

[31] Forgas, J.P. and George, J.M. (2001) Affective influences

on judgments and behavior in organizations: An informa-

tion processing perspective. Organizational Behavior

and Human Decision Processes, 86, 3-34.

[32] Isen, A.M. (2000) Some perspectives on positive affect

and self-regulation. Psychological Inquiry, 11, 184-187.

[33] Dalle, N. and Niedenthal, P.M. (2003) Conceptual

reorganization during emotional states/La reorganisation

de l'espace conceptuel au cours des etats emotionnels.

L’Année Psychologique, 103, 585-616.

[34] Niedenthal, P.M., Auxiette, C., Nugier, A., Dalle, N.,

Bonin, P. and Fayol, M. (2004) A prototype analysis of

the French category “emotion”. Cognition and Emotion,

18, 289-312.

[35] Isen, A.M. and Daubman, K.A. (1984) The influence of

affect on categorization. Journal of Personality and So-

cial Psychology, 47, 1206-1217.

[36] Isen, A.M., Niedenthal, P.M. and Cantor, N.(1992) An

influence of positive affect on social categorization.

Motivation and Emotion, 16, 65-78.

[37] Hansen, C.H. and Hansen, R.D. (1994) Automatic emo-

tion: Attention and facial efference. In: Niedenthal, P.M.

and Kitayama, S. Eds., The Heart’S Eye: Emotional In-

fluences in Perception and Attention, San Diego, Aca-

demic Press, Inc, 217-243.

[38] Ledoux, J. (1994) The emotional brain: The mysterious

underpinnings of emotional life. Touchstone book, New

York.

[39] MacLeod, C. and Hagan, R. (1992) Individual differenc-

es in the selective processing of threatening information,

and emotional responses to a stressful life event. Beha-

viour Research and Therapy, 30, 151- 16 1.

[40] Mogg, K., Bradley, B. P. and Williams, R. (1993) Subli-

minal processing of emotional information in anxiety and

depression. Journal of Abnormal Psychology, 102,

304-311.

[41] Mathews, A., Mackintosh, B. and Fulcher, E. P. (1997)

Cognitive biases in anxiety and attention to threat. Trends

in Cognitive Sciences, 1, 340-345.

[42] Vallortigara, G. and Rogers, L.J. (2005) Survival with an

asymmetrical brain: Advantages and disadvantages of

cerebral lateralization. Behavioral and Brain Sciences,

28, 574-633.

[43] Davidson, R. and Hugdahl, K. (1995) Brain asymmetry.

The MIT Press, Cambridge.

[44] Borod, J.C., Andelman, F. , Obler, L.K., Tweedy, J.R. and

Welkowitz, J. (1992) Right hemiphere specialization for

the identification of emotional words and sentences:

Evidence from Stroke patients. Neuropsychologia, 30,

827-844.

[45] Van Strien, J.W. and Morpurgo, M. (1992) Opposite he-

mispheric activations a s a resul t of emotiona lity t hr ea ten i ng

and non-threatening words. Neuropsychologia, 30,

845-848.

[46] Enloe, A.A., Ilardi, S.S., Atchley, R.A., Cromwell, R.L.,

and Sewell, K.W. (2001) Word valence, attention, and

hemispheric activity in depressed, remitted, and nonde-

pressed controls. Brain and Cognition, 46, 129-133.

[47] Meyer, M.B. and Smith, B.D. (1992) Cerebral processing

of nonverbal affective stimuli: Differential effects of

cognitive and affective sets on hemispheric asymmetry.

Biological Psychology, 24, 67-84.

[48] Sanders, J.D., McClure, K.A. and Zarate, M.A. (2004).

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1363

Cerebral hemisphere asymmetries in social perception:

Perceiving and responding to the individual and the

group. Social Cognition, 22, 279-291.

[49] Koivisto, M. and Laine, M. (1999) Strategies of semantic

categorization in the cerebral hemispheres. Brain and

Language, 66, 341-357.

[50] Zaidel, D.W. (1987) Hemispheric asymmetry in long-

term semantic relationships. Cognitive Neuropsychology,

4, 321-332.

[51] Vitkovitch, M. and Underwood, G. (1991) Hemispheric

differences in the processing of pictures of typical and

atypical semantic category members. Cortex, 27, 475-

480.

[52] Henriques, J.B. and Davidson, R.J (1991) Left frontal

hypoactivation in depression. Journal of Abnormal Psy-

chology, 100, 535-545.

[53] Howard, R., Fenwick, P., Brown, D. and Norton, R.

(1992) Relationship between CNV asymmetries and in-

dividual differences in cognitive performance, personali-

ty and gender. International Journal of Psychophysiology,

13, 191-197.

[54] Davidson, R.J., Marshall, J.R., Tomarken, A.J. and He-

nriques, J.B. (2000) While a phobic waits: Regional brain

electrical and autonomic activity in social phobics during

anticipation of public speaking. Biological Psychiatry, 47,

85-95.

[55] Lumley, M.A. and Sielky, K. (2000) Alexithymia, gender,

and hemispheric functioning. Comprarative Psychiatry,

41, 352-359.

[56] Sp al le tt a, G., Pas i ni , A . , Co s ta, A ., De An ge l is , D. , Ramundo,

N., Paolucci, S., et al. (2001) Alexithymic features in

stroke: Effects of laterality and gender. Psychosomatic

Medicine, 63, 944-950

[57] Bar-On, R., Tranel, D., Denburg, N.L., and Bechara, A.

(2003) Exploring the neurological substrate of emotional

and social intelligence. Brain, 126, 1790-1800.

[58] DeHouwer, J., Hermans, D., Rothermund, K. and Wen-

tura, D. (2002) Affective priming of semantic categorisa-

tion responses. Cognition and Emotion, 16, 643-666.

[59] Gohm, C. L. and Clore, G. L. (2000) Individual differ-

ences in emotional experience: Mapping available scales

to processes. Personality and Social Psychology Bulletin,

26, 679-697.

[60] Jacob, S., Lubart, T. I. and Getz, I. (1999) Interrelations

entre les aspects émotionnels de la personnalité. In:

Huteau, M. and Lautrey, J. Eds, Approches différentielles

en psychologie, Universitaire de Rennes, Rennes Pre sses,

365-368.

[61] Zenasni, F., Lubart , T., Pahlavan, F., Jac ob, S., Jac quet, A.

Y. and Lemoine, C. (2003) Traits émotionnels, intelli-

gence émotionnelle: Intérêt de ces concepts et étude de

leur interre lati ons. Revue Psychiatrie, Sciences Humaines

et Neurosciences, 1, 19-27.

[62] Montreuil, M. And Pedinielli, J.L. (1995) Caractéristiques

du traitement visuel parallèle chez des sujets sains

alexithymiques. Application de la Toronto Alexithymia

Scale et du Test de Saisie Visuelle Parallèle. L’Encéphale,

21, 589-595.

[63] Parker J.D.A., Keightley M.L., Smith, C.T. and Taylor

G.J. (1999) Interhemispheric transfer deficit in alexithy-

mia: an experimental study. Psychosomatic Medicine, 61,

464-468.

[64] TenHouten, W.D., Hoppe, K.D., Bogen, J.E. and Walter,

D.O.(1989) Alexithymia: an experimental study of cere-

bral commissurotomy patients and normal control sub-

jects. American Journal of Psychiatry, 143, 312-316.

[65] Zeitlin, S.B., Lane, R.D., O’Leary, D.S. and Schrift, M.J.

(1989) Interhemispheric transfer deficit and alexithymia.

American Journal of Psychiatry, 146, 1434-1439.

[66] Blonder, L.X., Bowers, D and Heilman, K.M. (1991) The

role of the right hemisphere in emotional communication.

Brain, 114, 1115-1127.

[67] Mandal, M.K, Borod, J.C., Hasthana, H.S., Mohanty, A.,

Mohanty, S, & Koff E. (1999) Effects of lesion variables

and emotion type on the perception of facial emotion.

Journal of Nervous Mental Diseases, 187, 603-609.

[68] Rief, W., Heuser, J. and Fichter, M.M. (1996) What does

the Toronto Alexithymia Scale TAS-R measure? Journal

of Clinical Psychology, 52, 423-429.

[69] Hendryx, M.S., Haviland, M.G., Shaw, D.G. (1991) Di-

mensions of alexithymia and their relationships to anxie-

ty and depression. Journal of Personality Assessment, 56,

227-237.

[70] Pelletier, J., Montreuil, M., Habib, M., Ali Cherif, A. and

Ly on-Caen, O. (1996) Alexithymia and multiple sclerosis:

alteration of interhemispheric transfer. European Journal

of Neurology, 3, 63.

[71] Goldman Consensus Group, New York City (2005) The

Goldman Consensus statement on depression in multiple

sclerosis. Multiple Sclerosis, 11, 328-337.

[72] Jouvent, R., Montreuil, M., Benoit, N., Lubetzki, C., Des

Lauriers, A., Tournier-Lasserve, E., Widlocher, D.,

Lhermitte, F. and Lyon-Caen, O. (1989) Cognitive im-

pairment, emotional disturbances and duration of mul-

tiple sclerosis. In: Jensen, K., Knudsen, L., Stenager, E.

and Grant, I. Eds., Mental disorders, cognitive deficits

and their treatment in multiple sclerosis, John Libbey

London, 139-145.

[73] Korostil, M. and Feinstein, A. (2007) Anxiety disorders

and their clinical correlates in multiple sclerosis patients.

Multiple Sclerosis, 13, 67-72.

[74] Feinstein, A., Feinstein, K., Gray, T. and O’Connor, P.

(1997) Prevalence and neurobehavioral correlates of pa-

thological laughing and crying in multiple sclerosis. Arc-

hive Neurology, 54, 1116-1121.

[75] Feinstein, A. (2004) The neuropsychiatry of multiple

sclerosis. Canadian Journal of Psychiatry, 49, 157-163.

[76] Schiffer, R. B. (2001) Psychiatric management in multi-

ple sclerosis. American Psychiatric Association, Annual

Meeting.

[77] Stenager, E., Knudsen, L. and Jensen, K. (1994) Multiple

sclerosis: Correlation of anxiety, physical impairment

and cognitive dysfunction. The Italian Journal of Neuro-

logical Sciences, 15, 99-103.

[78] Deloire, M.S.A., Salort, E., Bonnet, M., Arimone, Y.,

Boudineau, M., Amieva, H., Barroso, B., Ouallet, J.C.,

Pachai, C. , Galliau d, E., P etry, K.G., Dousset, V., Fabrigoule,

C. and Brochet, B. (2005) Cognitive impairment as

marker of diffuse brain abnormalities in early relapsing

remitting multiple sclerosis. Journal of Neurology, Neu-

rosurgery and Psychiatry, 76, 519-526.

[79] Sanfilipo, M.P., Benedict, R.H.B., Weinstock-Guttman,

B., & Bakshi, R. (2006) Gray and white matter brain

atrophy and neuropsychological impairment in multiple

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1364

sclerosis. Neurology, 66, 685-692.

[80] Kurtzke, J.F. (1983) Rating neurologic impairment in

multiple sclerosis and the disability status scale (EDSS).

Neurology, 33, 1444-1452.

[81] Prigatano, G.P., Amin, K. and Rosenstein, L.D. (1995)

Administration and scoring manual for the BNI screen

for higher cerebral functions. Phoenix, AZ: Barrow Neu-

rological Institute.

[82] Spielberger (1993) State-Trait Anxiety Inventory, Editions

du Centre de Psychologie Appliquée.

[83] Beck, R.A. and Steer, G. K. (1998) Inventaire de dépressio n

II. Les éditions du Centre de Psychologie Appliquée,

Paris.

[84] Bagby, R.M., Parker, J.D.A. and Taylor, G.J. (1994) The

twenty-item Toronto Alexithymia Scale: Item selection

and cross-validation of the factor structure. Journal of

Psychosomatic Research, 38, 2 3-32.

[85] Tayl or, G.J. and Taylor, H.L. (1997) Alexithymia. In: M.

McCallum & W. E. Piper Psychological mindedness: A

contemporary understanding. Mahwah, NJ, Lawrence Erl-

baum Associates Publishers, US, 77-104.

[86] Taylor, G. J., Bagby, R. M. and Luminet, O. (2000) As-

sessment of alexithymia: Self-report and observer-rated

measures. In: Bar-On, R. and Parker, J.D.A. Eds., The

handbook of emotional intelligence: Theory, development,

assessment, and application at home, school, and in the

workplace, San Francisco, CA, Jossey-Bass, US, 301-

319.

[87] Jessimer, M. and Markham, R. (1997) Alexthymia: A

right hemisphere dysfunction specific to recognition of

certain facial expressions. Brain and Cognition, 34,

246-258.

[88] Judd, C. M. (2000) Everyday data analysis in social psy-

chology: Comparisons of linear models. In: Reis, H.T.

and Judd, C.M. Eds., Handbook of research methods in

social and personality psychology, New York, NY, Cam-

bridge University Press, US, 2000, 370-393.

[89] Smith, S. D., Dixon, M. J., Bulman-Fleming, M. B.,

Birch, C., Laudi, N. and Wagar, B. (2005) Experience

with a category alters hemispheric asymmetries for the

detection of anomalies. Neurops ycholo gia, 43, 1911-1915.

[90] Campbell, A. and Muncer, S. (1994) Sex differences in

aggression: Social representation and social roles. British

Journal of Social Psychology, 33, 233-240.

[91] Salthouse, T. A. (1996) The processing-speed theory of

adult age differences in cognition. Psychological Review,

103, 403-428.

[92] Kujala, P., Portin, R., Revonsuo, A., Ruutiai n en, J. (1994)

Automatic and controlled information processing in mul-

tiple sclerosis, Brain, 117, 1115-1126.

[93] Audoin, B., Van Au Dong, M., Malikova, I., Confort-

Gouny, S., Ibarrola, D., Cozzone, P.J., Pelletier, J. and

Ranjeva, J.P. (2006 ) Functional magnetic resonance im-

aging and cognition at the very early stage of MS. Jour-

nal of the Neurological Sciences, 245, 87-91.

F. Pahlavan et al. / Health 2 (2010) 1349-1365

1365

ANNEX

Order of presentation and valence of the Triads:

French version

Target Concept Alternative Concepts

Neutral Triads

1)cuisinier serveur artiste

3)peau écorce cheveux

5)muscle os moteur

6)patinoire stade miroir

12)dents perles gorge

13)fourchette dard louche

17)football ping-pong œuf

18)chaussure pneu chemise

21)moquette parquet fourrure

27)chapeau chaussettes toit (roof)

30)serpent spaghettis tortue

Positive triads

2)soleil rire allumette

7)succès incertitude vacances

10)chanson cris barbe à papa

14)mer liberté plaine

15)plage falaise tendresse

20)repos fleur extinction

23)escalade monte-charge pique-nique

26)biberon câlin assiette

29)diamant fer étoile

31)promenade beauté course

Negative Triads

4)béquilles fauteuil roulant jambe

8)avalanche neige accident

9)ruine achats ouleur

11)avocat contrat divorce

16)hôpital hôtel ennui

19)poubelle odeur sac

22)doberman requin caniche

24)incendie enterrement grille-pain

25)dentiste souffrance brosse

28)ambulance fourgonnette guerre

English version

Target Concept Alternative Concepts

Neutral Triads

1)chief waiter artist

3)skin bark hair

5)muscle bone motor

6)skating rink stadium mirror

12)teeth pearls throat

13)fork stinger ladle

17)football ping-pong egg

18)shoe tire shirt

21)carpet floor fur

27)hat shoes roof

30)snake spaghetti turtle

Positive triads

2)sun laugh matches

7)success uncertainty vacation

10)song screams cotton-candy

14)sea freedom plain

15)beach cliff tenderness

20)rest flower extinguish

23)climbing hoist picnic

26)bottle cuddle plate

29)diamond iron star

31)walk beauty race

Negative Triads

4)crutch wheelchair leg

8)avalanche snow accident

9)ruin purchases pain

11)lawyer contract divorce

16)hospital hotel boredom

19)trash smell bag

22)doberman shark poodle

24)fire burial toaster

25)dentist suffering brush

28)ambulance truck war

_________________________________________

Note: Trial 1 Neutral was used as an example triad