Emotion word processing

Chair:  Lars Kuchinke

Ruhr-University Bochum, Department of Psychology, Bochum, Germany


Abstract:

The processing of words with an emotional connotation has recently attracted much
attention in emotion research and the affective neurosciences. An advantage of
words as a unit of experimental research seems to lie in the fact that many of the
variables that affect word recognition are known and can be controlled for. Different
studies revealed influences of emotional valence (how positive or negative a word is)
and arousal on the speed and the time course of word recognition. In a first study,
Estes & Adelman examined the relationship between these emotional variables and
show that both contribute in a non-independent manner to word recognition
performance. Thus, the question is not whether the processing of emotion words
differs from that of neutral words, but how and at which stage word recognition is
affected. To investigate this question Méndez-Bértolo & Hinojosa additionally
manipulated word frequency to functionally define the moment of lexical access. The
results point to an involvement of attentional mechanisms at a late processing stage
that are at a neural level often explained by back-projections from the amygdala to
the visual cortex. But Kissler et al. show that the amygdala might play only a small
role herein. While the Kissler et al. study also examined the modulation of neural
activations by induced mood, the Herbert study reveals that words describing ones
own emotion are processed differently in the brain. Finally, Kuchinke & Fritsch
investigated how the emotional connotation of symbolic stimuli such as words may
be acquired by learning associations to emotional contents.

Talk 1:

Contributions of Arousal and Valence to Word Recognition

Zachary Estes¹ & James S. Adelman^2
1 Department of Marketing, Bocconi University, Italy
² Department of Psychology, University of Warwick


Common words such as “kitten” and “coffin” have emotional connotations that are
often described in terms of arousal (from calming to exciting) and valence (from
negative to positive). Moreover, these factors of arousal and valence are
automatically detected, and they influence the recognition of and responding to
words. Specifically, negative words such as “coffin” tend to elicit slower responses
than positive words such as “kitten”, and arousing words such as “dead” tend to be
recognized faster than calming words such as “sad”.
However, the relationships among arousal, valence, and word recognition are the
subject of much current debate. First, some argue that arousal and valence are
independent factors, but others argue that they are inherently related. Furthermore,
some argue that valence exerts a categorical effect on word recognition, such that
moderately negative (e.g., “dirt”) and extremely negative words (e.g., “death”) are
recognized equally slowly, whereas others argue instead that valence has a graded
influence on word recognition.
To examine (a) the relationship between arousal and valence in ordinary language,
and (b) their independent and/or interactive contributions to word recognition, we
aggregated emotionality ratings of over 10,000 English words and merged it with
response times and accuracy rates in the lexical decision and reading aloud tasks
(acquired from E-Lexicon). Regression analyses revealed that arousal and valence
(1) are non-independent, (2) contribute uniquely to word recognition, and (3) exert
graded effects on word recognition. Implications for neuroscience research are
discussed.

Talk 2:

Effects of Word Frequency during the Processing of Emotional Words

Constantino Méndez-Bértolo¹ & Jose Antonio Hinojosa^2
1 Centro de Tecnología Biomédica, Universidad Politécnica de Madrid, 28223 Pozuelo de Alarcón, Spain.
² Instituto Pluridisciplinar, Universidad Complutense de Madrid, 28040 Madrid, Spain.


High frequency words are usually processed faster compared to those words that are
less frequently used in a language. This word frequency effect has been
demonstrated to modulate the processing of words with an emotional connotation. In
this regard, several studies have reported an interaction between word frequency
and affective content with behavioural and neural activity measures in word
recognition. Despite methodological differences in the experimental parameters
examined in these studies, it can be shown that some of the variables that affect the
processing of emotional connotation were not adequately controlled in some of these
studies. This might account for the lack of convergent results. In the present study we
aimed at clarifying the effects of word frequency on the processing of emotional
words. Event-related potentials were recorded while participants made lexical
decisions on high- and low-frequency negative and neutral nouns. Those
components that reflect interactions between word frequency and emotion were
detected with temporal and spatial principal component analyses. Low-frequency
negative nouns were recognized faster than low-frequency neutral nouns. Low-
frequency neutral nouns also elicited reduced amplitudes in a late positive
component compared to low-frequency negative nouns. No differences were evident
between high- and low-frequency neutral nouns. In sum, these findings are
discussed to reflect an involvement of attentional mechanisms during the evaluation
of the lexicality of a presented letter string that facilitate the processing of low-
frequency negative nouns.

Talk 3:

Effects of mood and emotional content on visual word processing – an fMRI study

Johanna Kissler¹, Bianka Gerling² , Reka Daniel² , Claus Tempelmann^2
1 University of Bielefeld
² Otto von Guericke University Magdeburg


Compared to neutral words, processing of emotional words has been shown to be
associated with distinct activity in the extended visual cortex as well as, less
consistently, frontal cortices, the amygdala, insula and cingulate cortex. Some of the
inconsistencies are likely due to task differences, others have been attributed to
implicit effects of mood congruency. Here, we examine to what extent experimentally
induced transient mood alters brain activation during processing of emotional and
neutral words using a semantic monitoring task. Different moods states were induced
following which participants were instructed to monitor sequences of positive,
negative and neutral adjectives for occasional occurences of color adjectives. Effects
of mood and word content were analysed, excluding responses evoked by color
words. Analyses revealed robust effects of emotion on extended visual cortex activity
for both positve and negative compared to neutral words. Overall, mood affected
cerebral activity in the cingulate gyrus and an interaction between mood and
emotional content occurred in the left fusiform gyrus. Amygdala activity could only be
identified using a region of interest approach and only following negative mood
induction. Across different mood states, findings confirm enhanced extended visual
cortex activity in response to emotional compared to neutral words, underscoring the
robustness of these effects, but partially also interacting with mood states. Although
amygdala activity was identified, it was considerably less consistent than visual
cortex activity, casting doubt on the idea that back-projections from the amygdala are
an obligatory driving source of visual cortex activity in emotional word processing.

Talk 4:

Emotion processing and its regulation: What words can tell us about it

Cornelia Herbert
Department of Psychology, University of Würzburg, Germany


Emotion perception in self and others is important for successful social interactions. It
is involved in the generation of subjective emotional experiences (i.e., feelings) and
the regulation of emotions. Building upon previous research using verbal material for
emotion induction the present studies investigated by means of ERP and functional
imaging methods how emotional words describing the own emotion are processed in
the brain and what the underlying mechanisms are. In association with this, we
investigated if verbally negating the own emotion serves as an effective emotion
regulation strategy when exposed to emotional facial expressions.
The results allow a number of conclusions: During reading, emotional words
describing the own emotion are more deeply processed compared to unreferenced or
other-related emotional words. Second, processing of self-related emotional words
increases activity in medial prefrontal brain structures involved in conscious emotion
processing, whereas reading of emotional words, particularly unpleasant ones, leads
to an increase in amygdala and insula activity irrespective of the word’s reference.
Third, reframing one’s emotion by using negated emotional cue words decreases
cortical processing of fearful faces and spontaneously triggers emotion regulation
strategies that appear more closely associated with cognitive reappraisal than with
emotion suppression. Theoretically, these findings support an embodied view of
language. More specifically, they demonstrate that investigating emotional word
processing in social contexts could tell us much about the neural mechanisms
underlying the most private and subjective aspects of emotion processing, i.e., of
emotional experience and its regulation.

Talk 5:

Processing emotional words and nonwords: an evaluative conditioning ERP study

Lars Kuchinke¹², Nathalie Fritsch^1
1 Ruhr-University Bochum, Department of Psychology, Bochum, Germany
² The Cluster of Excellence 'Languages of Emotion', Free University Berlin,Germany

Numerous studies have shown that word recognition differs depending on the
emotional connotation of a particular word. Emotional connotation, i.e. differences in
emotional valence and arousal, have been shown to affect response times and
accuracy measures in the lexical decision task and to modulate very early (80-
120ms; e.g., Hofmann et al., 2009) and later components of the event-related
potentials (ERP). It is widely accepted that words receive their emotional connotation
through the learning of emotional-semantic associations, but direct evidence for this
proposal is lacking. To address this question we conducted an evaluative
conditioning study using meaningless pseudowords. Participants learned
associations between randomly selected 50 pseudowords and 150 affective pictures
and between another 50 pseudowords and 150 neutral pictures on five consecutive
days. Each pseudoword was associated with more than one picture and with different
pictures (of the same connotation) each day to guaranty that only the emotional
connotation and not a particular association to one picture was learned. This was
tested in both, a subsequent ERP lexical decision study and in an explicit valence
judgements task. The results reveal effects of learned negative connotations in
pseudowords in early and late ERP components, replicating the effects known from
word processing. These findings support the assumption of learned associations as
the basis of a words' emotional connotation. Still, the nature of the very early effects
that are most probably related to modulated attention to emotional (pseudo)words is
in need of further clarification.