- Email: [email protected]
Attentional orienting towards emotion: P2 and N400 ERP effects
Neuropsychologia 49 (2011) 3121–3129
Contents lists available at ScienceDirect
Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia
Attentional orienting towards emotion: P2 and N400 ERP effects Philipp Kanske a,b,∗ , Jan Plitschka c , Sonja A. Kotz a a
Max Planck Institute for Human Cognitive and Brain Sciences, Neurocognition of Rhythm in Communication Group, Leipzig, Germany Department of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, Mannheim, Germany c Bonn-Aachen International Center for Information Technology B-IT, Algorithmic Bioinformatics Group, Rheinische Friedrich-Wilhelms-Universität Bonn, Germany b
a r t i c l e
i n f o
Article history: Received 3 December 2010 Received in revised form 7 July 2011 Accepted 18 July 2011 Available online 23 July 2011 Keywords: EEG Event related potentials Emotion Linguistics Semantics Social cognition
a b s t r a c t Attention can be oriented to different spatial locations yielding faster processing of attended compared to unattended stimuli. Similarly attention can be oriented to a semantic category such as “animals” or “tools”. Words from the attended category will also be recognized faster than words from an unattended category. Here, we asked whether it is possible to orient attention to an emotional category, for example, “negative emotional stimuli”. Furthermore, we investigated which mechanisms facilitate processing of attended stimuli. In an attentional orienting paradigm in which cues are informative with regard to the spatial location, semantic category, or emotional category of subsequent target words, we found attention effects in all three cue conditions. Words at attended locations or of the attended semantic or emotional category were responded to faster than unattended categories. While spatial attention acted upon early visual processing and amplified occipital N1-P2 potentials, semantic cues modulated the N400 amplitude indexing semantic processing. Emotional cues also yielded an N400 modulation; however, in addition, a left anterior P2 effect was observed. The data clearly show that attention can be oriented to emotional categories. Emotional orienting yields facilitated processing of an attended emotional category through the modulation of early and late word processing stages. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction Human communication relies on the fast detection of others’ emotional states as these are highly predictive with regard to future behavior. Anger, for example, may result in an attack, while a neutral state may signal safety. Therefore, emotional stimuli are not only processed faster than neutral stimuli (Carretie, Hinojosa, Martin-Loeches, Mercado, & Tapia, 2004; Kanske & Kotz, 2007; Ortigue et al., 2004), but also attract attention (e.g. to their spatial location, see Stormark, Nordby, & Hugdahl, 1995; or in visual image processing, see Schupp et al., 2007). However, it is unclear whether attention can also be oriented in advance to signals of others’ emotional states. Thus, we asked whether non-emotional cues can lead to an active orienting of attention to the emotional quality of a stimulus. We investigated emotional orienting in a Posner-type cueing paradigm and compared it to spatial and semantic cueing effects (Cristescu & Nobre, 2008). Orienting of attention refers to the mechanisms that select information. It has primarily been investigated with perceptual features
∗ Corresponding author at: Department of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, Square J5, 68159 Mannheim, Germany. Tel.: +49 621 1703 6307; fax: +49 621 1703 6305. E-mail addresses: [email protected], [email protected] (P. Kanske). 0028-3932/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropsychologia.2011.07.022
in vision, mainly in visual spatial processing. Classic experimental paradigms include visual search (Treisman & Gelade, 1980) and cueing tasks (Posner, 1980). In the latter, a cue that informs the participant about the spatial location of an upcoming target is presented prior to that target. The modulatory influence of attention on stimulus processing can then be studied when comparing targets that are preceded by valid and invalid cues. Reaction times (RT) to validly cued targets are typically shorter than those to invalidly cued target sites (Posner, 1980). This indicates that visual spatial attention is oriented according to cue information, facilitating the processing of subsequent target stimuli. In eventrelated potentials (ERP) this facilitated processing is reflected in an amplification of early visual potentials over occipital electrodes. P1 and N1 amplitudes are enhanced for attended stimuli (Hillyard & Anllo-Vento, 1998); however, there are also reports of P2 modulations (Eason, 1981; Gomez, Vazquez, Vaquero, Lopez-Mendoza, & Cardoso, 1998; Hillyard & Mangun, 1986; Maenoa, Gjinia, Iraminaa, Etob, & Ueno, 2004; Shedden & Nordgaard, 2001; van Voorhis & Hillyard, 1977). Only a few studies have investigated whether attention can also select information that is not sensory-perceptual, but postperceptual. Recently, Cristescu and Nobre (2008) and Cristescu, Devlin, and Nobre (2006) proposed a version of the classical cueing paradigm that tests orienting of attention to the semantic category of a stimulus. Participants performed a lexical decision task. Prior to the words, an abstract cue symbol was presented that
3122
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
Fig. 1. Cueing task with spatial, semantic, and emotional cueing conditions. The cues validly (80%) or invalidly (20%) predicted the location, semantic category, or emotional connotation of subsequent target words. Participants identified the grammatical gender of the target words.
was informative with regard to the semantic category of the target word, which could either be an animal or a tool word. Valid cues led to shorter RTs in the lexical decision task. This behavioral effect was accompanied by a reduction of the N400 potential, an index of semantic analysis and integration (Kutas & Federmeier, 2000). A similar behavioral result in spatial and semantic orienting is thus achieved through distinct mechanisms that modify different levels of stimulus analysis (Cristescu & Nobre, 2008). Analysis of cue-related activity in fMRI revealed a similar fronto-parietal network for spatial and semantic orienting and additional activations for semantic orienting including left hemisphere areas involved in semantic analysis of words, such as the inferior frontal cortex (Cristescu et al., 2006). The semantic orienting task (Cristescu & Nobre, 2008; Cristescu et al., 2006) can be adapted to the current question of whether attention can be oriented to the emotional quality of a target stimulus. Accelerated processing of a stimulus that is preceded by an abstract cue validly predicting whether the stimulus is emotional or neutral would indicate that attention can select emotional categories similarly to selecting spatial locations. To test this possibility we extended the attentional orienting paradigm by including (1) spatially informative cues, (2) semantically informative cues, and (3) emotionally informative cues (see Fig. 1). The same abstract cue symbols were used in the three conditions; however, their meaning differed. Spatial cues predicted the location of the target to the right or left of a fixation point. Comparable to Cristescu and Nobre (2008), the semantic cues predicted whether the target words belonged to the category “animals” or “tools”. In the emotional orienting condition, the cues predicted the emotional connotation of the target words. Negative (e.g., tumor, corpse, jail) and neutral words (e.g., pea, belly, dam) were presented. The design thus largely resembles priming studies in which, for example, semantically related words are presented (Rugg, 1985). The prime word is expected to activate candidates in a semantic network, which then ease the processing and integration of subsequently presented target words. The critical difference in the present design is that prime words are replaced by abstract cue symbols that may only pre-activate a certain number of network candidates if attention is voluntarily and actively oriented towards them. In addition to the question whether it is possible to orient attention to a stimulus’ emotional characteristics, we asked which stages of emotional word processing would be modulated by attention, i.e. can be pre-activated. To this end, we also recorded ERPs during the task. With reference to the semantic orienting task (Cristescu & Nobre, 2008) it seems likely to expect an N400 effect. Stimuli of one emotional category (e.g., hatred, bomb, terror) also group semantically (Bradley & Lang, 1994; Moore, Romney, Hsia, & Rusch, 1999;
Romney, Moore, & Rusch, 1997; Wierzbicka, 1992). Consequently, comparable to semantic priming, an N400 has been observed in emotional priming (Steinbeis & Koelsch, 2009). Interestingly, emotional words also modulate ERPs in much earlier time-windows. Frequently, effects are observed in the P2 time-window (Begleiter & Platz, 1969; Begleiter, Projesz, & Garazzo, 1979; Bernat, Bunce, & Shevrin, 2001; Herbert, Kissler, Junghöfer, Peyk, & Rockstroh, 2006; Schapkin, Gusev, & Kuhl, 2000). For example, Kanske and Kotz (2007) observed a larger P2 amplitude for positive emotional than neutral words in a lexical decision task, however, a number of studies have also observed other early emotion effects (Herbert, Junghöfer, & Kissler, 2008; Hofmann, Kuchinke, Tamm, Vo, & Jacobs, 2009; Ortigue et al., 2004; Schacht & Sommer, 2009; Scott, O’Donnell, Leuthold, & Sereno, 2009; for reviews Kissler, Assadollahi, & Herbert, 2006; Kotz & Paulmann, 2011). These early emotion effects have often been interpreted as indicating rapid attention capture by emotional words (Herbert et al., 2008) or rudimentary semantic stimulus classification (Kissler, Herbert, Winkler, & Junghofer, 2009). Also, there is mixed evidence regarding the automaticity of these early effects. In a study directly contrasting different tasks, Begleiter et al. (1979) found early emotion effects only in an affective evaluation task, but not in a letter-identification condition. This would suggest that early ERP effects are only elicited when participants focus on judging the emotional connotation of presented words. Independent of the exact nature of these early emotion effects, if they represent an emotional word processing stage that can be selected by attention, than there should be ERP effects prior to the N400, potentially in the P2 time-window. In summary, our hypotheses are: (1) attentional orienting effects in RTs in all three cueing conditions (spatial, semantic, emotional). (2) An amplification of early visual potentials over occipital electrodes in the spatial orienting condition. (3) Modulation of the N400 in the semantic orienting condition (see Cristescu & Nobre, 2008). (4) Similarly, an N400 modulation is expected in the emotional orienting condition. Here, an effect may also show in an earlier time window (e.g., the P2; Kanske & Kotz, 2007). 2. Methods 2.1. Participants Twenty-five participants (13 females) volunteered for the experiment. All were native speakers of German and right-handed according to the Edinburgh Handedness Inventory (Oldfield, 1971). Three participants (2 females) were excluded due to increased error rates (more than 40% in at least one condition). Data from two further participants (2 females) could not be analyzed due to severe artifacts. The remaining 20 participants had a mean age of 24.8 years (SD 3.0) and a handedness laterality quotient of 96.0 (SD 6.4). All reported normal or corrected-to-normal vision. 2.2. Stimuli To assess orienting of attention to semantic and to emotional categories, two word lists were constructed (see Supplement 1 for a complete stimulus list). The list for the semantic task contained 32 concrete animal and 32 tool words. The emotional word list included 32 negative emotional and 32 neutral words, which differed significantly in valence and arousal ratings, but not in rated concreteness (see Table 1). The ratings were taken from a previous rating study with 32 participants (16 females). The word groups did not differ in number of letters and number of syllables. Also, the words were controlled for frequency of usage according to the Wortschatzlexikon of the University of Leipzig (http://wortschatz.uni-leipzig.de/; Biemann, Bordag, Heyer, Quasthoff, & Wolf, 2004). All words were German nouns, half of which were of male gender (“der”), and the other half of female gender (“die”). 2.3. Task and procedure A trial sequence is displayed in Fig. 1. First, a cue was presented in the center of the screen, followed by a target word in either the right or the left hemifield. Participants decided whether the target word required the article “der” or “die”. However, the cue contained different types of information in the different runs of the experiment. (1) In the spatial orienting condition, the cue predicted the location
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
3123
Table 1 Psycholinguistic parameters of the selected words. Means and standard deviations (in parentheses) are given. Frequency
# of letters
# of syllables
Valence
Arousal
Concreteness
Semantic Wordlist
Animals Tools
13.69 (1.64) 14.47 (1.59)
5.41 (1.27) 5.72 (1.20)
1.94 (0.56) 1.94 (0.35)
– –
– –
– –
Emotional Wordlist
Neutral Negative
13.47 (2.36) 13.47 (2.53)
6.00 (1.08) 5.50 (1.14)
1.84 (0.37) 1.81 (0.40)
5.18 (0.20) 2.32 (0.28)
2.56 (0.66) 6.40 (0.69)
4.63 (2.25) 5.39 (2.13)
of the target words (both, the semantic and the emotional word list were presented in this condition). (2) In the semantic orienting condition (using the semantic word list), the cue was informative with regard to a semantic category (animal or tool), and (3) in the emotional orienting condition (using the emotional word list), the cue predicted the emotional category of the target words (negative or neutral). The cue was valid in 80% of all trials, and predicted the wrong location or category in 20% of the trials. The timing of events in a trial was as follows: a fixation dot was presented for a variable duration (1250–1750 ms) followed by the cue (100 ms). After an interval of 1100–1500 ms, the target appeared for 150 ms on the screen. Participants had to respond within 2000 ms. Feedback was presented for 1000 ms before a new trial started. The feedback consisted of the word “falsch” (wrong) for erroneous and missed responses, and of the RT in ms for correct responses. The assignment of the cue symbols “×” and “+” to a certain location or category was counterbalanced across participants. There were two experimental sessions, which were at least seven days apart. Each session included (1) a spatial orienting task with one wordlist (e.g., spatial orienting with the semantic list consisting of animal and tool words) and (2) the semantic or emotional orienting task with the other wordlist (e.g., emotional orienting with the emotional wordlist consisting of negative and neutral words). The order of sessions and tasks was counterbalanced across participants. Each word was presented twice in the invalid cueing condition (once in the left and once in the right hemifield), and eight times in the valid condition to have 20% invalid and 80% valid trials. Thus, there were 320 trials in each task, which were split into 4 blocks with short breaks in between. Stimuli were presented on a computer screen in white on a black background. The words were presented in upper case 3◦ of visual angle to the left or right of fixation. The computer monitor was positioned 100 cm in front of the participants’ eyes. Participants responded with a right- or left-hand button press. Assignment of the responses (the articles “der” and “die”) to the response hands was counterbalanced. A practice block was presented before the experiment with longer stimulus presentation and response times to help the participants adjust to the task. Prior to this practice, to ensure that participants correctly understood the target word categories, they also (1) produced items for each category or (2) sorted presented words into the categories (either with or without a prior cue). 2.4. EEG recording
For the statistical analyses of both the behavioral and the EEG data, the software package SAS (SAS Copyright© SAS Institute Inc., Cary, NC, USA) was used. Repeated measures ANOVAs including validity (valid and invalid cues) and cue type (spatial, semantic, emotional) were computed for the behavioral data. The analysis of the EEG data additionally included the factors region (anterior and posterior) as well as hemisphere (left and right). Additionally, we compared processing of emotional to neutral words by calculating repeated measures ANOVAs including the word’s emotional quality (negative and neutral) and cue validity (valid and invalid) in the emotional orienting task. The mean numbers of averaged trials per condition are reported in Supplement 2.
3. Results 3.1. Behavioral results Validly cued targets were reacted to faster than invalidly cued targets (see Fig. 3; F(1,19) = 39.73, p < .0001). Validity interacted significantly with cue type (F(3,57) = 4.64, p < .01). Follow-up analyses yielded shorter RTs for valid compared to invalid trials for each cue type; however, the effect was stronger for spatial cues (semantic cues: F(1,19) = 11.11, p < .01; emotional cues: F(1,19) = 12.15, p < .01; spatial cues, semantic wordlist: F(1,19) = 13.10, p < .001; spatial cues, emotional wordlist: F(1,19) = 25.71, p < .0001). Accuracy was higher following a valid compared to an invalid cue (see Fig. 3; F(1,19) = 24.02, p < .0001). Validity interacted significantly with cue type (F(3,57) = 5.30, p < .01), indicating that validly cued targets were responded to more accurately in the spatial cueing condition, but not in the semantic or emotional cues, where the difference between valid and invalid cues did not reach significance (semantic cues: F(1,19) = 3.27, p > .05; emotional cues: F(1,19) = 0.0, p > .10; spatial cues, semantic wordlist: F(1,19) = 12.39, p < .01; spatial cues, emotional wordlist: F(1,19) = 11.49, p < .01).
Ag–AgCl electrodes were used to record the EEG from 64 scalp positions according to the international 10-20 system, see Fig. 2 (Jasper, 1958). Vertical eye movements were registered with two electrodes positioned above and below the right eye. The horizontal electrooculograms were recorded with lateral electrodes from both eyes. The ground electrode was located on the sternum. Data were referenced to the left mastoid, and the right mastoid was actively recorded. Impedances were below 5 k for all recordings. Brain vision professional recorder software by Brain Products GmbH and the corresponding amplifier Brainamps were used. Sampling rate was 500 Hz. Data were stored on hard disk and were filtered online with a bandpass between DC and 250 Hz. Further filtering was only done for graphical display (14 Hz lowpass filter). Data were re-referenced offline to the average of both mastoids. 2.5. Data analysis EEG data were analyzed with the software package EEP (ERP Evaluation Package EEP 3.2, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany). Epochs of 1000 ms after stimulus onset were computed according to a 200 ms pre-stimulus baseline. All trials were visually inspected for artifacts. An automatic rejection of trials exceeding 30 V for the two eye channels and 40 V for CZ and PZ within a sliding window of 200 ms was additionally applied. Trials were averaged for each condition and participant and for each condition across participants (grand average). After visual inspection of the grand averages and a time-line analysis with 50 ms time windows, three time windows were chosen for further analysis. These time windows represent the following deflections: N1-P2 amplitude difference was computed between 140 and 255 ms over occipital electrodes O1, OZ, and O2 to obtain a measure of early visual processing. Here, we analyzed the amplitude difference in accordance with previous attention studies (e.g., Eason, 1981; Woldorff, Hansen, & Hillyard, 1987). The frontal P2 was computed as mean amplitude between 190 and 290 ms and for the N400 between 460 and 590 ms. Channels were grouped into four regions of interest (ROI) with 9 electrodes each (see Fig. 2).
Fig. 2. Electrode positions and regions of interest (ROI) in gray. For the analysis of early visual potentials, the black encircled electrodes were included.
3124
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
tudes than neutral words (see Fig. 6; F(1,19) = 16.91, p < .001). This difference was not significant in the invalid condition.
4. Discussion
Fig. 3. Reaction times (A) and the percentage of trials correctly responded to (B) are displayed for the spatial, semantic, and emotional cueing conditions for the two wordlists (semantic and emotional). Means and S.E.M. are given.
3.2. Electrophysiological results 3.2.1. Occipital N1-P2 We observed a modulation of the N1-P2 amplitude difference between 140 and 255 ms over occipital electrodes. For spatially valid cues, the amplitude difference was enlarged compared to spatially invalid cues (see Fig. 4; spatial cues, semantic wordlist F(1,19) = 7.21, p < .01; spatial cues, emotional wordlist: F(1,19) = 7.63, p < .01). There was no significant difference for semantic and emotional cues. 3.2.2. Frontal P2 Between 190 and 290 ms after word onset there was an interaction of validity and cue type with hemisphere and region (see Fig. 5; F(1,19) = 7.02, p < .01). This interaction indicated an enlarged P2 amplitude for emotionally valid compared to invalid cues over left anterior electrodes (F(1,19) = 4.49, p < .05). No such effect was present for any of the other cue types. There were no differences between negative and neutral words. 3.2.3. N400 In the specified N400 time window between 460 and 590 ms, we found a significant validity by cue type interaction that was broadly distributed (see Fig. 5; F(1,19) = 11.41, p < .01). For the semantic cues, a validity effect was present over all electrodes, indicating larger amplitudes for invalid compared to valid cues (F(1,19) = 6.27, p < .05). For emotional cues, we observed an interaction of validity and region (F(1,19) = 6.44, p < .05). Emotionally valid cues yielded reduced N400 amplitudes, but only over anterior sites (F(1,19) = 5.99, p < .05). Spatially informative cues also modulated the N400 amplitude; in the opposite direction, however. Here, valid cues elicited larger N400 amplitudes compared to invalid cues (semantic wordlist: F(1,19) = 12.12, p < .01; emotional wordlist: F(1,19) = 6.21, p < .05). In the emotional orienting condition we also observed a significant effect of the emotional word quality (F(1,19) = 4.88, p < .05) and an interaction of validity and emotion (F(1,19) = 4.63, p < .05). In the valid condition, emotional words elicited larger N400 ampli-
The present study yields new insights into the mechanisms of attentional orienting. It is the first to show that attention can be actively oriented to the emotional quality of a word. We presented different types of cues that (1) predict the spatial location, (2) the semantic category, and (3) the emotional category of subsequently presented target words. For each cue type validly cued words were responded to faster than invalidly cued words. The ERP data show that this orienting effect results from the modulation of different stimulus processing stages. While spatial orienting enhances amplitudes of early visual potentials, semantic orienting yields an increased N400. Orienting to the emotional category of a word also increases the N400 amplitude, but also a left anterior P2. A large number of studies have shown that attention can be oriented to perceptual features such as spatial location or color (Hillyard & Anllo-Vento, 1998; Posner, 1980; Treisman & Gelade, 1980). Cristescu and Nobre (2008) showed that the semantic category of a word can also be selected by attention yielding speeded processing of these words. The main question we addressed here was whether attention can be oriented to the emotional status of word stimuli. The present data clearly show that attention can be oriented to emotion. Abstract cues that predict the emotional category of a target word speed up word processing if the prediction is valid and slow down word processing if invalid. The size of this effect is smaller than the spatial orienting effect, but comparable to semantic orienting (for a similar spatial/semantic orienting difference see Cristescu & Nobre, 2008). This demonstrates that participants build up expectations of the emotional target word category based on non-emotional cues. This is a highly adaptive process for human communication as it ensures the rapid prediction of others’ emotional states via all available cues. The process that enables this effect acts upon semantic analysis and integration of target words as indexed in the altered N400 amplitude. This effect is in agreement with data from affective priming (Steinbeis & Koelsch, 2009) and resembles the N400 modulation in the semantic orienting condition. Thus, it reflects the semantic coherence among words of one emotional category (Bradley & Lang, 1994; Moore et al., 1999; Romney et al., 1997; Wierzbicka, 1992). Cues that predict an emotional category seem to activate the related semantic network yielding facilitated integration of the upcoming words that fit this category, which is reflected in the reduced N400 amplitude (Kutas & Federmeier, 2011). Emotional orienting additionally acted upon earlier processing stages. We observed a modulation of a left anterior P2, which was larger for words whose emotional category was validly predicted by the preceding cue. The interpretation of this P2 effect is not as straightforward as the N400 effect; however, the following suggestions can be considered. We know from a number of studies that the amplitude of the P2 is, under certain conditions, sensitive to emotionality in words (Begleiter & Platz, 1969; Begleiter et al., 1979; Bernat et al., 2001; Herbert et al., 2006; Kanske & Kotz, 2007; Schapkin et al., 2000). The most prominent suggestion is that this reflects attentional capturing by emotional words. Following this suggestion, the present data would indicate that cues signaling the emotional quality of a subsequent target can boost the attention capture effect, which is reflected in an increased P2 amplitude. This “amplification” view is in line with the well-established interpretation of the spatial attention effects on early visual potentials (Hillyard & AnlloVento, 1998). A second possibility is that early emotion effects in the ERP reflect rudimentary semantic stimulus categorization (Kissler et al., 2009). Pre-N400 semantic effects have indeed been reported
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
3125
Fig. 4. ERPs to validly and invalidly cued target words at occipital electrodes for the spatial, semantic, and emotional cueing conditions for the two wordlists (semantic and emotional).
(for an extensive review see Dien, 2009). However, this raises the question why the P2 effects were not observed in the semantic orienting condition in the present experiment. In consequence, emotional words must represent an exceptional semantic category as has been previously suggested (mainly, however, for emotion words denoting feeling such as “anger, love”, but not for emotional words such as “catastrophe, grave” as used here; Altarriba & Bauer, 2004; Altarriba, Bauer, & Benvenuto, 1999; Paivio, Yuille, & Madigan, 1968). Even though the present study cannot unequivocally define the mechanism underlying the P2 emotion effect, the effect is clearly linked to the orienting of attention to emotion at an early stage of word processing. One question that arises from this early P2 effect is whether such an effect should have any behavioral consequences. This is difficult to answer as the semantic and emotional orienting effects did not
differ in the RT data. However, such a direct comparison is complicated. It would require that semantic and emotional categories (animal vs. tool and negative vs. neutral words) are of equivalent size and coherence. Otherwise one would, metaphorically speaking, compare the effects of spatial cues that orient attention to certain locations of different size. Future research should address this open issue which goes beyond the question targeted in the present experiment. We did not observe ERP differences prior to the N400 when directly contrasting negative with neutral words in the present study which seems to contradict the observed P2 emotional orienting effects. However, the targeted psychological mechanisms are very different. The orienting effect demonstrates that attention oriented towards the emotional quality of a word selects characteristics of the word processed in the P2 (and N400) time-window,
3126
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
Fig. 5. ERPs at selected electrodes and difference maps for validly and invalidly cued target words for the spatial, semantic, and emotional cueing conditions for the two wordlists (semantic and emotional).
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
3127
Fig. 6. ERPs at selected electrodes and difference maps to emotional and neutral words preceded by cues either validly or invalidly predicting the emotional category of the target words.
thereby yielding facilitated processing of those words matching the selected features (e.g., being negative emotional when the cue predicted negative emotion). In contrast, the direct comparison of negative and neutral words will give information about differential processing of the two categories, which seems to be adaptively adjusted to the current task demands. There is some evidence that only tasks requiring participants to evaluate the emotional quality of words elicit P2 emotion effects (Begleiter et al., 1979, but see also Bernat et al., 2001). In the present study the task was to identify the grammatical gender of words. The lack of an early emotion effect may, thus, be due to these non-affective task demands. Interestingly, we also found a modulation of the N400 emotion effect. The N400 amplitude was reduced for negative compared to neutral words (see also Kanske & Kotz, 2007), however, only in valid cue trials. This suggests that the integration advantage of emotional over neutral words is abolished when integration is complicated by preactivation of the wrong context (e.g., of “negative emotion” when
a neutral word is presented). In line with this interpretation, the disappearance of the N400 effect in invalid trials is due to an amplitude increase for negative words, not a decrease for neutral words. A limitation regarding the contrast of negative and neutral words is that words were repeatedly shown in the present experiment. As repetition priming effects have been reported for emotional, but not neutral words, this may have influenced the comparison of the two (for details see Luo et al., 2004). The present data also confirm the results of Cristescu and Nobre (2008). We replicate the behavioral facilitation effect for semantically cued words and also report an N400 amplitude modulation. This is interesting as there are a number of differences between the two experimental designs. First, Cristescu and Nobre (2008) and Cristescu et al. (2006) had participants perform a lexical decision task which required the presentation of words and pseudowords. While this is not problematic in the spatial cueing condition as spatial cues provide information about the location of the target and
3128
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
are, thus, applicable to words and pseudowords, semantic cues only apply to words, as the pseudowords have no semantic meaning and do not belong to a semantic category. We therefore avoided the use of pseudowords by applying a grammatical gender decision task. Second, Cristescu and Nobre (2008) presented each word only once to each participant. While this has the advantage that no arbitrary “non-semantic” relations of cue symbols and words can be learned, it also means that comparisons between the spatial and semantic cueing conditions had to be between-subject, and comparisons between valid and invalid conditions were between-item. For the current experiment, words were repeatedly presented in each condition, allowing within-subject and within-item comparison. Third, Cristescu and Nobre (2008) only analyzed words presented in the right hemifield based on left-hemispheric dominance for word processing. Nevertheless, modulatory influence of attention on word processing should be the same, even if the information needs to be transferred from the right to the left hemisphere. We therefore included all words in the present data analysis. Despite these substantial differences between the studies, the results largely resemble each other, underscoring the reliability of the semantic orienting effect. We also observed the modulation of early visual processing in the spatial orienting conditions. The amplified N1-P2 amplitude for attended words is in agreement with a large number of visual attention studies (Eason, 1981; Gomez et al., 1998; Hillyard & Mangun, 1986; Johannes, Munte, Heinze, & Mangun, 1995; Maenoa et al., 2004; Shedden & Nordgaard, 2001; van Voorhis & Hillyard, 1977). Spatial attention also influenced the N400 such that validly cued words elicited larger amplitudes. This pattern replicates previous data (Cristescu & Nobre, 2008; McCarthy & Nobre, 1993) and supports the notion that spatial attention can modulate semantic analysis and integration of linguistic stimuli (Miniussi, Marzi, & Nobre, 2005). Specifically, the reduced N400 amplitude for unattended words has been interpreted as showing that semantic analysis is not automatic but depends on visual selective attention (e.g., McCarthy & Nobre, 1993). The topographic maps of the N400 effects in the present study may suggest that different subprocesses generating the N400 were involved in the spatial orienting compared to the other conditions as the distributions seem to be more posterior. However, the lack of statistically significant distribution differences makes it difficult to draw definite conclusions and renders answer to the question of specific subprocesses to future research. To conclude, the present study clearly demonstrates that attention can be actively oriented to the emotional category of upcoming word stimuli. Correct expectations about the emotionality of stimuli yield faster responses to these stimuli. Attention also operates at different processing stages. Attentional orienting to emotion was similar to orienting to a semantic category in the N400 response, but additionally entailed an earlier modulation of the P2 component. This early effect indicates attentional selection of stimulus characteristics that are processed prior to word semantics and ensures the rapid detection of potentially threatening situations, an evolutionary highly adaptive mechanism.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neuropsychologia.2011.07.022.
References Altarriba, J., & Bauer, L. M. (2004). The distinctiveness of emotion concepts: A comparison between emotion, abstract, and concrete words. American Journal of Psychology, 117, 389–410.
Altarriba, J., Bauer, L. M., & Benvenuto, C. (1999). Concreteness, context availability, and imageability ratings and word associations for abstract, concrete, and emotion words. Behavior Research Methods Instruments & Computers, 31, 578–602. Begleiter, H., & Platz, A. (1969). Cortical evoked potentials to semantic stimuli. Psychophysiology, 6, 91–100. Begleiter, H., Projesz, B., & Garazzo, R. (1979). Evoked brain potentials and behavior. In H. Begleiter (Ed.), Visual evoked potentials and affective ratings of semantic stimuli (pp. 127–143). New York: Plenum Press. Bernat, E., Bunce, S., & Shevrin, H. (2001). Event-related brain potentials differentiate positive and negative mood adjectives during both supraliminal and subliminal visual processing. International Journal of Psychophysiology, 42, 11–34. Biemann, C., Bordag, S., Heyer, G., Quasthoff, U., & Wolf, C. (2004). Languageindependent methods for compiling monolingual lexical data. Lecture Notes in Computer Science Berlin and Heidelberg: Springer., pp. 217–228 Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The Self-Assessment Manikin and the Semantic Differential. Journal of Behavior Therapy and Experimental Psychiatry, 25, 49–59. Carretie, L., Hinojosa, J. A., Martin-Loeches, M., Mercado, F., & Tapia, M. (2004). Automatic attention to emotional stimuli: Neural correlates. Human Brain Mapping, 22, 290–299. Cristescu, T. C., & Nobre, A. C. (2008). Differential modulation of word recognition by semantic and spatial orienting of attention. Journal of Cognitive Neuroscience, 20, 787–801. Cristescu, T. C., Devlin, J. T., & Nobre, A. C. (2006). Orienting attention to semantic categories. Neuroimage, 33, 1178–1187. Dien, J. (2009). The neurocognitive basis of reading single words as seen through early latency ERPs: A model of converging pathways. Biological Psychology, 80, 10–22. Eason, R. G. (1981). Visual evoked potential correlates of early neural filtering during selective attention. Bulletin of the Psyehonomie Society, 18, 203–206. Gomez, C. M., Vazquez, M., Vaquero, E., Lopez-Mendoza, D., & Cardoso, M. J. (1998). Frequency analysis of the EEG during spatial selective attention. International Journal of Neuroscience, 95, 17–32. Herbert, C., Kissler, J., Junghöfer, M., Peyk, P., & Rockstroh, B. (2006). Processing of emotional adjectives: Evidence from startle EMG and ERPs. Psychophysiology, 43, 197–206. Herbert, C., Junghöfer, M., & Kissler, J. (2008). Event related potentials to emotional adjectives during reading. Psychophysiology, 45, 487–498. Hillyard, S. A., & Anllo-Vento, L. (1998). Event-related brain potentials in the study of visual selective attention. Proceedings of the National Academy of Sciences of the United States of America, 95, 781–787. Hillyard, S. A., & Mangun, G. R. (1986). The neural basis of visual selective attention: A commentary on Harter and Aine. Biological Psychology, 23, 265–279. Hofmann, M. J., Kuchinke, L., Tamm, S., Vo, M. L., & Jacobs, A. M. (2009). Affective processing within 1/10th of a second: High arousal is necessary for early facilitative processing of negative but not positive words. Cognitive, Affective & Behavioral Neuroscience, 9, 389–397. Jasper, H. (1958). The ten-twenty electrode system of the international federation. Electroencephalography and Clinical Neurophysiology, 10, 371–375. Johannes, S., Munte, T. F., Heinze, H. J., & Mangun, G. R. (1995). Luminance and spatial attention effects on early visual processing. Brain Research Cognitive Brain Research, 2, 189–205. Kanske, P., & Kotz, S. A. (2007). Concreteness in emotional words: ERP evidence from a hemifield study. Brain Research, 1148, 138–148. Kissler, J., Assadollahi, R., & Herbert, C. (2006). Emotional and semantic networks in visual word processing: Insights from ERP studies. Progress in Brain Research, 156, 147–183. Kissler, J., Herbert, C., Winkler, I., & Junghofer, M. (2009). Emotion and attention in visual word processing: An ERP study. Biological Psychology, 80, 75–83. Kotz, S. A., & Paulmann, S. (2011). Emotion, language and the brain. Language and Linguistics Compass, 5, 108–125. Kutas, M., & Federmeier, K. D. (2000). Electrophysiology reveals semantic memory use in language comprehension. Trends in Cognitive Sciences, 4, 463–470. Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: Finding meaning in the N400 component of the event-related brain potential (ERP). Annual Review of Psychology, 62, 621–647. Luo, Q., Peng, D., Jin, Z., Xu, D., Xiao, L., & Ding, G. (2004). Emotional valence of words modulates the subliminal repetition priming effect in the left fusiform gyrus: An event-related fMRI study. Neuroimage, 21, 414–421. Maenoa, T., Gjinia, K., Iraminaa, K., Etob, F., & Ueno, S. (2004). Event-related potential P2 derived from visual attention to the hemi-space. Source localization with LORETA. International Congress Series, 1270, 262–265. McCarthy, G., & Nobre, A. C. (1993). Modulation of semantic processing by spatial selective attention. Electroencephalography and Clinical Neurophysiology, 88, 210–219. Miniussi, C., Marzi, C. A., & Nobre, A. C. (2005). Modulation of brain activity by selective task sets observed using event-related potentials. Neuropsychologia, 43, 1514–1528. Moore, C. C., Romney, A. K., Hsia, T.-L., & Rusch, C. D. (1999). The universality of the semantic structure of emotion terms: Methods for the study of inter- and intra-cultural variability. American Anthropologist, 101, 529–546. Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia, 9, 97–113. Ortigue, S., Michel, C. M., Murray, M. M., Mohr, C., Carbonnel, S., & Landis, T. (2004). Electrical neuroimaging reveals early generator modulation to emotional words. Neuroimage, 21, 1242–1251.
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129 Paivio, A., Yuille, J. C., & Madigan, S. A. (1968). Concreteness, imagery, and meaningfulness values for 925 nouns. Journal of Experimental Psychology, 76(Suppl. 1–25). Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Romney, A. K., Moore, C. C., & Rusch, C. D. (1997). Cultural universals: Measuring the semantic structure of emotion terms in English and Japanese. Proceedings of the National Academy of Sciences of the United States of America, 94, 5489–5494. Rugg, M. D. (1985). The effects of semantic priming and work repetition on eventrelated potentials. Psychophysiology, 22, 642–647. Schacht, A., & Sommer, W. (2009). Time course and task dependence of emotion effects in word processing. Cognitive, Affective & Behavioral Neuroscience, 9, 28–43. Schapkin, S. A., Gusev, A. N., & Kuhl, J. (2000). Categorization of unilaterally presented emotional words: An ERP analysis. Acta Neurobiologiae Experimentalis (Wars), 60, 17–28. Schupp, H. T., Stockburger, J., Codispoti, M., Junghöfer, M., Weike, A. I., & Hamm, A. O. (2007). Selective visual attention to emotion. Journal of Neuroscience, 27, 1082–1089.
3129
Scott, G. G., O’Donnell, P. J., Leuthold, H., & Sereno, S. C. (2009). Early emotion word processing: Evidence from event-related potentials. Biological Psychology, 80, 95–104. Shedden, J. M., & Nordgaard, C. L. (2001). ERP time course of perceptual and post-perceptual mechanisms of spatial selection. Brain Research Cognitive Brain Research, 11, 59–75. Steinbeis, N., & Koelsch, S. (2009). Affective priming effects of musical sounds on the processing of word meaning. Journal of Cognitive Neuroscience. Stormark, K. M., Nordby, H., & Hugdahl, K. (1995). Attentional shifts to emotionally charged cues: Behavioral and ERP data. Cognition & Emotion, 9, 507–523. Treisman, A. M., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12, 97–136. van Voorhis, S., & Hillyard, S. A. (1977). Visual evoked potentials and selective attention to points in space. Perception & Psychophysics, 22, 54–62. Wierzbicka, A. (1992). Defining emotion concepts. Cognitive Science, 16, 539–581. Woldorff, M., Hansen, J. C., & Hillyard, S. A. (1987). Evidence for effects of selective attention in the mid-latency range of the human auditory event-related potential. Electroencephalography and Clinical Neurophysiology Supplement, 40, 146–154.
Contents lists available at ScienceDirect
Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia
Attentional orienting towards emotion: P2 and N400 ERP effects Philipp Kanske a,b,∗ , Jan Plitschka c , Sonja A. Kotz a a
Max Planck Institute for Human Cognitive and Brain Sciences, Neurocognition of Rhythm in Communication Group, Leipzig, Germany Department of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, Mannheim, Germany c Bonn-Aachen International Center for Information Technology B-IT, Algorithmic Bioinformatics Group, Rheinische Friedrich-Wilhelms-Universität Bonn, Germany b
a r t i c l e
i n f o
Article history: Received 3 December 2010 Received in revised form 7 July 2011 Accepted 18 July 2011 Available online 23 July 2011 Keywords: EEG Event related potentials Emotion Linguistics Semantics Social cognition
a b s t r a c t Attention can be oriented to different spatial locations yielding faster processing of attended compared to unattended stimuli. Similarly attention can be oriented to a semantic category such as “animals” or “tools”. Words from the attended category will also be recognized faster than words from an unattended category. Here, we asked whether it is possible to orient attention to an emotional category, for example, “negative emotional stimuli”. Furthermore, we investigated which mechanisms facilitate processing of attended stimuli. In an attentional orienting paradigm in which cues are informative with regard to the spatial location, semantic category, or emotional category of subsequent target words, we found attention effects in all three cue conditions. Words at attended locations or of the attended semantic or emotional category were responded to faster than unattended categories. While spatial attention acted upon early visual processing and amplified occipital N1-P2 potentials, semantic cues modulated the N400 amplitude indexing semantic processing. Emotional cues also yielded an N400 modulation; however, in addition, a left anterior P2 effect was observed. The data clearly show that attention can be oriented to emotional categories. Emotional orienting yields facilitated processing of an attended emotional category through the modulation of early and late word processing stages. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction Human communication relies on the fast detection of others’ emotional states as these are highly predictive with regard to future behavior. Anger, for example, may result in an attack, while a neutral state may signal safety. Therefore, emotional stimuli are not only processed faster than neutral stimuli (Carretie, Hinojosa, Martin-Loeches, Mercado, & Tapia, 2004; Kanske & Kotz, 2007; Ortigue et al., 2004), but also attract attention (e.g. to their spatial location, see Stormark, Nordby, & Hugdahl, 1995; or in visual image processing, see Schupp et al., 2007). However, it is unclear whether attention can also be oriented in advance to signals of others’ emotional states. Thus, we asked whether non-emotional cues can lead to an active orienting of attention to the emotional quality of a stimulus. We investigated emotional orienting in a Posner-type cueing paradigm and compared it to spatial and semantic cueing effects (Cristescu & Nobre, 2008). Orienting of attention refers to the mechanisms that select information. It has primarily been investigated with perceptual features
∗ Corresponding author at: Department of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, Square J5, 68159 Mannheim, Germany. Tel.: +49 621 1703 6307; fax: +49 621 1703 6305. E-mail addresses: [email protected], [email protected] (P. Kanske). 0028-3932/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropsychologia.2011.07.022
in vision, mainly in visual spatial processing. Classic experimental paradigms include visual search (Treisman & Gelade, 1980) and cueing tasks (Posner, 1980). In the latter, a cue that informs the participant about the spatial location of an upcoming target is presented prior to that target. The modulatory influence of attention on stimulus processing can then be studied when comparing targets that are preceded by valid and invalid cues. Reaction times (RT) to validly cued targets are typically shorter than those to invalidly cued target sites (Posner, 1980). This indicates that visual spatial attention is oriented according to cue information, facilitating the processing of subsequent target stimuli. In eventrelated potentials (ERP) this facilitated processing is reflected in an amplification of early visual potentials over occipital electrodes. P1 and N1 amplitudes are enhanced for attended stimuli (Hillyard & Anllo-Vento, 1998); however, there are also reports of P2 modulations (Eason, 1981; Gomez, Vazquez, Vaquero, Lopez-Mendoza, & Cardoso, 1998; Hillyard & Mangun, 1986; Maenoa, Gjinia, Iraminaa, Etob, & Ueno, 2004; Shedden & Nordgaard, 2001; van Voorhis & Hillyard, 1977). Only a few studies have investigated whether attention can also select information that is not sensory-perceptual, but postperceptual. Recently, Cristescu and Nobre (2008) and Cristescu, Devlin, and Nobre (2006) proposed a version of the classical cueing paradigm that tests orienting of attention to the semantic category of a stimulus. Participants performed a lexical decision task. Prior to the words, an abstract cue symbol was presented that
3122
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
Fig. 1. Cueing task with spatial, semantic, and emotional cueing conditions. The cues validly (80%) or invalidly (20%) predicted the location, semantic category, or emotional connotation of subsequent target words. Participants identified the grammatical gender of the target words.
was informative with regard to the semantic category of the target word, which could either be an animal or a tool word. Valid cues led to shorter RTs in the lexical decision task. This behavioral effect was accompanied by a reduction of the N400 potential, an index of semantic analysis and integration (Kutas & Federmeier, 2000). A similar behavioral result in spatial and semantic orienting is thus achieved through distinct mechanisms that modify different levels of stimulus analysis (Cristescu & Nobre, 2008). Analysis of cue-related activity in fMRI revealed a similar fronto-parietal network for spatial and semantic orienting and additional activations for semantic orienting including left hemisphere areas involved in semantic analysis of words, such as the inferior frontal cortex (Cristescu et al., 2006). The semantic orienting task (Cristescu & Nobre, 2008; Cristescu et al., 2006) can be adapted to the current question of whether attention can be oriented to the emotional quality of a target stimulus. Accelerated processing of a stimulus that is preceded by an abstract cue validly predicting whether the stimulus is emotional or neutral would indicate that attention can select emotional categories similarly to selecting spatial locations. To test this possibility we extended the attentional orienting paradigm by including (1) spatially informative cues, (2) semantically informative cues, and (3) emotionally informative cues (see Fig. 1). The same abstract cue symbols were used in the three conditions; however, their meaning differed. Spatial cues predicted the location of the target to the right or left of a fixation point. Comparable to Cristescu and Nobre (2008), the semantic cues predicted whether the target words belonged to the category “animals” or “tools”. In the emotional orienting condition, the cues predicted the emotional connotation of the target words. Negative (e.g., tumor, corpse, jail) and neutral words (e.g., pea, belly, dam) were presented. The design thus largely resembles priming studies in which, for example, semantically related words are presented (Rugg, 1985). The prime word is expected to activate candidates in a semantic network, which then ease the processing and integration of subsequently presented target words. The critical difference in the present design is that prime words are replaced by abstract cue symbols that may only pre-activate a certain number of network candidates if attention is voluntarily and actively oriented towards them. In addition to the question whether it is possible to orient attention to a stimulus’ emotional characteristics, we asked which stages of emotional word processing would be modulated by attention, i.e. can be pre-activated. To this end, we also recorded ERPs during the task. With reference to the semantic orienting task (Cristescu & Nobre, 2008) it seems likely to expect an N400 effect. Stimuli of one emotional category (e.g., hatred, bomb, terror) also group semantically (Bradley & Lang, 1994; Moore, Romney, Hsia, & Rusch, 1999;
Romney, Moore, & Rusch, 1997; Wierzbicka, 1992). Consequently, comparable to semantic priming, an N400 has been observed in emotional priming (Steinbeis & Koelsch, 2009). Interestingly, emotional words also modulate ERPs in much earlier time-windows. Frequently, effects are observed in the P2 time-window (Begleiter & Platz, 1969; Begleiter, Projesz, & Garazzo, 1979; Bernat, Bunce, & Shevrin, 2001; Herbert, Kissler, Junghöfer, Peyk, & Rockstroh, 2006; Schapkin, Gusev, & Kuhl, 2000). For example, Kanske and Kotz (2007) observed a larger P2 amplitude for positive emotional than neutral words in a lexical decision task, however, a number of studies have also observed other early emotion effects (Herbert, Junghöfer, & Kissler, 2008; Hofmann, Kuchinke, Tamm, Vo, & Jacobs, 2009; Ortigue et al., 2004; Schacht & Sommer, 2009; Scott, O’Donnell, Leuthold, & Sereno, 2009; for reviews Kissler, Assadollahi, & Herbert, 2006; Kotz & Paulmann, 2011). These early emotion effects have often been interpreted as indicating rapid attention capture by emotional words (Herbert et al., 2008) or rudimentary semantic stimulus classification (Kissler, Herbert, Winkler, & Junghofer, 2009). Also, there is mixed evidence regarding the automaticity of these early effects. In a study directly contrasting different tasks, Begleiter et al. (1979) found early emotion effects only in an affective evaluation task, but not in a letter-identification condition. This would suggest that early ERP effects are only elicited when participants focus on judging the emotional connotation of presented words. Independent of the exact nature of these early emotion effects, if they represent an emotional word processing stage that can be selected by attention, than there should be ERP effects prior to the N400, potentially in the P2 time-window. In summary, our hypotheses are: (1) attentional orienting effects in RTs in all three cueing conditions (spatial, semantic, emotional). (2) An amplification of early visual potentials over occipital electrodes in the spatial orienting condition. (3) Modulation of the N400 in the semantic orienting condition (see Cristescu & Nobre, 2008). (4) Similarly, an N400 modulation is expected in the emotional orienting condition. Here, an effect may also show in an earlier time window (e.g., the P2; Kanske & Kotz, 2007). 2. Methods 2.1. Participants Twenty-five participants (13 females) volunteered for the experiment. All were native speakers of German and right-handed according to the Edinburgh Handedness Inventory (Oldfield, 1971). Three participants (2 females) were excluded due to increased error rates (more than 40% in at least one condition). Data from two further participants (2 females) could not be analyzed due to severe artifacts. The remaining 20 participants had a mean age of 24.8 years (SD 3.0) and a handedness laterality quotient of 96.0 (SD 6.4). All reported normal or corrected-to-normal vision. 2.2. Stimuli To assess orienting of attention to semantic and to emotional categories, two word lists were constructed (see Supplement 1 for a complete stimulus list). The list for the semantic task contained 32 concrete animal and 32 tool words. The emotional word list included 32 negative emotional and 32 neutral words, which differed significantly in valence and arousal ratings, but not in rated concreteness (see Table 1). The ratings were taken from a previous rating study with 32 participants (16 females). The word groups did not differ in number of letters and number of syllables. Also, the words were controlled for frequency of usage according to the Wortschatzlexikon of the University of Leipzig (http://wortschatz.uni-leipzig.de/; Biemann, Bordag, Heyer, Quasthoff, & Wolf, 2004). All words were German nouns, half of which were of male gender (“der”), and the other half of female gender (“die”). 2.3. Task and procedure A trial sequence is displayed in Fig. 1. First, a cue was presented in the center of the screen, followed by a target word in either the right or the left hemifield. Participants decided whether the target word required the article “der” or “die”. However, the cue contained different types of information in the different runs of the experiment. (1) In the spatial orienting condition, the cue predicted the location
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
3123
Table 1 Psycholinguistic parameters of the selected words. Means and standard deviations (in parentheses) are given. Frequency
# of letters
# of syllables
Valence
Arousal
Concreteness
Semantic Wordlist
Animals Tools
13.69 (1.64) 14.47 (1.59)
5.41 (1.27) 5.72 (1.20)
1.94 (0.56) 1.94 (0.35)
– –
– –
– –
Emotional Wordlist
Neutral Negative
13.47 (2.36) 13.47 (2.53)
6.00 (1.08) 5.50 (1.14)
1.84 (0.37) 1.81 (0.40)
5.18 (0.20) 2.32 (0.28)
2.56 (0.66) 6.40 (0.69)
4.63 (2.25) 5.39 (2.13)
of the target words (both, the semantic and the emotional word list were presented in this condition). (2) In the semantic orienting condition (using the semantic word list), the cue was informative with regard to a semantic category (animal or tool), and (3) in the emotional orienting condition (using the emotional word list), the cue predicted the emotional category of the target words (negative or neutral). The cue was valid in 80% of all trials, and predicted the wrong location or category in 20% of the trials. The timing of events in a trial was as follows: a fixation dot was presented for a variable duration (1250–1750 ms) followed by the cue (100 ms). After an interval of 1100–1500 ms, the target appeared for 150 ms on the screen. Participants had to respond within 2000 ms. Feedback was presented for 1000 ms before a new trial started. The feedback consisted of the word “falsch” (wrong) for erroneous and missed responses, and of the RT in ms for correct responses. The assignment of the cue symbols “×” and “+” to a certain location or category was counterbalanced across participants. There were two experimental sessions, which were at least seven days apart. Each session included (1) a spatial orienting task with one wordlist (e.g., spatial orienting with the semantic list consisting of animal and tool words) and (2) the semantic or emotional orienting task with the other wordlist (e.g., emotional orienting with the emotional wordlist consisting of negative and neutral words). The order of sessions and tasks was counterbalanced across participants. Each word was presented twice in the invalid cueing condition (once in the left and once in the right hemifield), and eight times in the valid condition to have 20% invalid and 80% valid trials. Thus, there were 320 trials in each task, which were split into 4 blocks with short breaks in between. Stimuli were presented on a computer screen in white on a black background. The words were presented in upper case 3◦ of visual angle to the left or right of fixation. The computer monitor was positioned 100 cm in front of the participants’ eyes. Participants responded with a right- or left-hand button press. Assignment of the responses (the articles “der” and “die”) to the response hands was counterbalanced. A practice block was presented before the experiment with longer stimulus presentation and response times to help the participants adjust to the task. Prior to this practice, to ensure that participants correctly understood the target word categories, they also (1) produced items for each category or (2) sorted presented words into the categories (either with or without a prior cue). 2.4. EEG recording
For the statistical analyses of both the behavioral and the EEG data, the software package SAS (SAS Copyright© SAS Institute Inc., Cary, NC, USA) was used. Repeated measures ANOVAs including validity (valid and invalid cues) and cue type (spatial, semantic, emotional) were computed for the behavioral data. The analysis of the EEG data additionally included the factors region (anterior and posterior) as well as hemisphere (left and right). Additionally, we compared processing of emotional to neutral words by calculating repeated measures ANOVAs including the word’s emotional quality (negative and neutral) and cue validity (valid and invalid) in the emotional orienting task. The mean numbers of averaged trials per condition are reported in Supplement 2.
3. Results 3.1. Behavioral results Validly cued targets were reacted to faster than invalidly cued targets (see Fig. 3; F(1,19) = 39.73, p < .0001). Validity interacted significantly with cue type (F(3,57) = 4.64, p < .01). Follow-up analyses yielded shorter RTs for valid compared to invalid trials for each cue type; however, the effect was stronger for spatial cues (semantic cues: F(1,19) = 11.11, p < .01; emotional cues: F(1,19) = 12.15, p < .01; spatial cues, semantic wordlist: F(1,19) = 13.10, p < .001; spatial cues, emotional wordlist: F(1,19) = 25.71, p < .0001). Accuracy was higher following a valid compared to an invalid cue (see Fig. 3; F(1,19) = 24.02, p < .0001). Validity interacted significantly with cue type (F(3,57) = 5.30, p < .01), indicating that validly cued targets were responded to more accurately in the spatial cueing condition, but not in the semantic or emotional cues, where the difference between valid and invalid cues did not reach significance (semantic cues: F(1,19) = 3.27, p > .05; emotional cues: F(1,19) = 0.0, p > .10; spatial cues, semantic wordlist: F(1,19) = 12.39, p < .01; spatial cues, emotional wordlist: F(1,19) = 11.49, p < .01).
Ag–AgCl electrodes were used to record the EEG from 64 scalp positions according to the international 10-20 system, see Fig. 2 (Jasper, 1958). Vertical eye movements were registered with two electrodes positioned above and below the right eye. The horizontal electrooculograms were recorded with lateral electrodes from both eyes. The ground electrode was located on the sternum. Data were referenced to the left mastoid, and the right mastoid was actively recorded. Impedances were below 5 k for all recordings. Brain vision professional recorder software by Brain Products GmbH and the corresponding amplifier Brainamps were used. Sampling rate was 500 Hz. Data were stored on hard disk and were filtered online with a bandpass between DC and 250 Hz. Further filtering was only done for graphical display (14 Hz lowpass filter). Data were re-referenced offline to the average of both mastoids. 2.5. Data analysis EEG data were analyzed with the software package EEP (ERP Evaluation Package EEP 3.2, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany). Epochs of 1000 ms after stimulus onset were computed according to a 200 ms pre-stimulus baseline. All trials were visually inspected for artifacts. An automatic rejection of trials exceeding 30 V for the two eye channels and 40 V for CZ and PZ within a sliding window of 200 ms was additionally applied. Trials were averaged for each condition and participant and for each condition across participants (grand average). After visual inspection of the grand averages and a time-line analysis with 50 ms time windows, three time windows were chosen for further analysis. These time windows represent the following deflections: N1-P2 amplitude difference was computed between 140 and 255 ms over occipital electrodes O1, OZ, and O2 to obtain a measure of early visual processing. Here, we analyzed the amplitude difference in accordance with previous attention studies (e.g., Eason, 1981; Woldorff, Hansen, & Hillyard, 1987). The frontal P2 was computed as mean amplitude between 190 and 290 ms and for the N400 between 460 and 590 ms. Channels were grouped into four regions of interest (ROI) with 9 electrodes each (see Fig. 2).
Fig. 2. Electrode positions and regions of interest (ROI) in gray. For the analysis of early visual potentials, the black encircled electrodes were included.
3124
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
tudes than neutral words (see Fig. 6; F(1,19) = 16.91, p < .001). This difference was not significant in the invalid condition.
4. Discussion
Fig. 3. Reaction times (A) and the percentage of trials correctly responded to (B) are displayed for the spatial, semantic, and emotional cueing conditions for the two wordlists (semantic and emotional). Means and S.E.M. are given.
3.2. Electrophysiological results 3.2.1. Occipital N1-P2 We observed a modulation of the N1-P2 amplitude difference between 140 and 255 ms over occipital electrodes. For spatially valid cues, the amplitude difference was enlarged compared to spatially invalid cues (see Fig. 4; spatial cues, semantic wordlist F(1,19) = 7.21, p < .01; spatial cues, emotional wordlist: F(1,19) = 7.63, p < .01). There was no significant difference for semantic and emotional cues. 3.2.2. Frontal P2 Between 190 and 290 ms after word onset there was an interaction of validity and cue type with hemisphere and region (see Fig. 5; F(1,19) = 7.02, p < .01). This interaction indicated an enlarged P2 amplitude for emotionally valid compared to invalid cues over left anterior electrodes (F(1,19) = 4.49, p < .05). No such effect was present for any of the other cue types. There were no differences between negative and neutral words. 3.2.3. N400 In the specified N400 time window between 460 and 590 ms, we found a significant validity by cue type interaction that was broadly distributed (see Fig. 5; F(1,19) = 11.41, p < .01). For the semantic cues, a validity effect was present over all electrodes, indicating larger amplitudes for invalid compared to valid cues (F(1,19) = 6.27, p < .05). For emotional cues, we observed an interaction of validity and region (F(1,19) = 6.44, p < .05). Emotionally valid cues yielded reduced N400 amplitudes, but only over anterior sites (F(1,19) = 5.99, p < .05). Spatially informative cues also modulated the N400 amplitude; in the opposite direction, however. Here, valid cues elicited larger N400 amplitudes compared to invalid cues (semantic wordlist: F(1,19) = 12.12, p < .01; emotional wordlist: F(1,19) = 6.21, p < .05). In the emotional orienting condition we also observed a significant effect of the emotional word quality (F(1,19) = 4.88, p < .05) and an interaction of validity and emotion (F(1,19) = 4.63, p < .05). In the valid condition, emotional words elicited larger N400 ampli-
The present study yields new insights into the mechanisms of attentional orienting. It is the first to show that attention can be actively oriented to the emotional quality of a word. We presented different types of cues that (1) predict the spatial location, (2) the semantic category, and (3) the emotional category of subsequently presented target words. For each cue type validly cued words were responded to faster than invalidly cued words. The ERP data show that this orienting effect results from the modulation of different stimulus processing stages. While spatial orienting enhances amplitudes of early visual potentials, semantic orienting yields an increased N400. Orienting to the emotional category of a word also increases the N400 amplitude, but also a left anterior P2. A large number of studies have shown that attention can be oriented to perceptual features such as spatial location or color (Hillyard & Anllo-Vento, 1998; Posner, 1980; Treisman & Gelade, 1980). Cristescu and Nobre (2008) showed that the semantic category of a word can also be selected by attention yielding speeded processing of these words. The main question we addressed here was whether attention can be oriented to the emotional status of word stimuli. The present data clearly show that attention can be oriented to emotion. Abstract cues that predict the emotional category of a target word speed up word processing if the prediction is valid and slow down word processing if invalid. The size of this effect is smaller than the spatial orienting effect, but comparable to semantic orienting (for a similar spatial/semantic orienting difference see Cristescu & Nobre, 2008). This demonstrates that participants build up expectations of the emotional target word category based on non-emotional cues. This is a highly adaptive process for human communication as it ensures the rapid prediction of others’ emotional states via all available cues. The process that enables this effect acts upon semantic analysis and integration of target words as indexed in the altered N400 amplitude. This effect is in agreement with data from affective priming (Steinbeis & Koelsch, 2009) and resembles the N400 modulation in the semantic orienting condition. Thus, it reflects the semantic coherence among words of one emotional category (Bradley & Lang, 1994; Moore et al., 1999; Romney et al., 1997; Wierzbicka, 1992). Cues that predict an emotional category seem to activate the related semantic network yielding facilitated integration of the upcoming words that fit this category, which is reflected in the reduced N400 amplitude (Kutas & Federmeier, 2011). Emotional orienting additionally acted upon earlier processing stages. We observed a modulation of a left anterior P2, which was larger for words whose emotional category was validly predicted by the preceding cue. The interpretation of this P2 effect is not as straightforward as the N400 effect; however, the following suggestions can be considered. We know from a number of studies that the amplitude of the P2 is, under certain conditions, sensitive to emotionality in words (Begleiter & Platz, 1969; Begleiter et al., 1979; Bernat et al., 2001; Herbert et al., 2006; Kanske & Kotz, 2007; Schapkin et al., 2000). The most prominent suggestion is that this reflects attentional capturing by emotional words. Following this suggestion, the present data would indicate that cues signaling the emotional quality of a subsequent target can boost the attention capture effect, which is reflected in an increased P2 amplitude. This “amplification” view is in line with the well-established interpretation of the spatial attention effects on early visual potentials (Hillyard & AnlloVento, 1998). A second possibility is that early emotion effects in the ERP reflect rudimentary semantic stimulus categorization (Kissler et al., 2009). Pre-N400 semantic effects have indeed been reported
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
3125
Fig. 4. ERPs to validly and invalidly cued target words at occipital electrodes for the spatial, semantic, and emotional cueing conditions for the two wordlists (semantic and emotional).
(for an extensive review see Dien, 2009). However, this raises the question why the P2 effects were not observed in the semantic orienting condition in the present experiment. In consequence, emotional words must represent an exceptional semantic category as has been previously suggested (mainly, however, for emotion words denoting feeling such as “anger, love”, but not for emotional words such as “catastrophe, grave” as used here; Altarriba & Bauer, 2004; Altarriba, Bauer, & Benvenuto, 1999; Paivio, Yuille, & Madigan, 1968). Even though the present study cannot unequivocally define the mechanism underlying the P2 emotion effect, the effect is clearly linked to the orienting of attention to emotion at an early stage of word processing. One question that arises from this early P2 effect is whether such an effect should have any behavioral consequences. This is difficult to answer as the semantic and emotional orienting effects did not
differ in the RT data. However, such a direct comparison is complicated. It would require that semantic and emotional categories (animal vs. tool and negative vs. neutral words) are of equivalent size and coherence. Otherwise one would, metaphorically speaking, compare the effects of spatial cues that orient attention to certain locations of different size. Future research should address this open issue which goes beyond the question targeted in the present experiment. We did not observe ERP differences prior to the N400 when directly contrasting negative with neutral words in the present study which seems to contradict the observed P2 emotional orienting effects. However, the targeted psychological mechanisms are very different. The orienting effect demonstrates that attention oriented towards the emotional quality of a word selects characteristics of the word processed in the P2 (and N400) time-window,
3126
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
Fig. 5. ERPs at selected electrodes and difference maps for validly and invalidly cued target words for the spatial, semantic, and emotional cueing conditions for the two wordlists (semantic and emotional).
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
3127
Fig. 6. ERPs at selected electrodes and difference maps to emotional and neutral words preceded by cues either validly or invalidly predicting the emotional category of the target words.
thereby yielding facilitated processing of those words matching the selected features (e.g., being negative emotional when the cue predicted negative emotion). In contrast, the direct comparison of negative and neutral words will give information about differential processing of the two categories, which seems to be adaptively adjusted to the current task demands. There is some evidence that only tasks requiring participants to evaluate the emotional quality of words elicit P2 emotion effects (Begleiter et al., 1979, but see also Bernat et al., 2001). In the present study the task was to identify the grammatical gender of words. The lack of an early emotion effect may, thus, be due to these non-affective task demands. Interestingly, we also found a modulation of the N400 emotion effect. The N400 amplitude was reduced for negative compared to neutral words (see also Kanske & Kotz, 2007), however, only in valid cue trials. This suggests that the integration advantage of emotional over neutral words is abolished when integration is complicated by preactivation of the wrong context (e.g., of “negative emotion” when
a neutral word is presented). In line with this interpretation, the disappearance of the N400 effect in invalid trials is due to an amplitude increase for negative words, not a decrease for neutral words. A limitation regarding the contrast of negative and neutral words is that words were repeatedly shown in the present experiment. As repetition priming effects have been reported for emotional, but not neutral words, this may have influenced the comparison of the two (for details see Luo et al., 2004). The present data also confirm the results of Cristescu and Nobre (2008). We replicate the behavioral facilitation effect for semantically cued words and also report an N400 amplitude modulation. This is interesting as there are a number of differences between the two experimental designs. First, Cristescu and Nobre (2008) and Cristescu et al. (2006) had participants perform a lexical decision task which required the presentation of words and pseudowords. While this is not problematic in the spatial cueing condition as spatial cues provide information about the location of the target and
3128
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129
are, thus, applicable to words and pseudowords, semantic cues only apply to words, as the pseudowords have no semantic meaning and do not belong to a semantic category. We therefore avoided the use of pseudowords by applying a grammatical gender decision task. Second, Cristescu and Nobre (2008) presented each word only once to each participant. While this has the advantage that no arbitrary “non-semantic” relations of cue symbols and words can be learned, it also means that comparisons between the spatial and semantic cueing conditions had to be between-subject, and comparisons between valid and invalid conditions were between-item. For the current experiment, words were repeatedly presented in each condition, allowing within-subject and within-item comparison. Third, Cristescu and Nobre (2008) only analyzed words presented in the right hemifield based on left-hemispheric dominance for word processing. Nevertheless, modulatory influence of attention on word processing should be the same, even if the information needs to be transferred from the right to the left hemisphere. We therefore included all words in the present data analysis. Despite these substantial differences between the studies, the results largely resemble each other, underscoring the reliability of the semantic orienting effect. We also observed the modulation of early visual processing in the spatial orienting conditions. The amplified N1-P2 amplitude for attended words is in agreement with a large number of visual attention studies (Eason, 1981; Gomez et al., 1998; Hillyard & Mangun, 1986; Johannes, Munte, Heinze, & Mangun, 1995; Maenoa et al., 2004; Shedden & Nordgaard, 2001; van Voorhis & Hillyard, 1977). Spatial attention also influenced the N400 such that validly cued words elicited larger amplitudes. This pattern replicates previous data (Cristescu & Nobre, 2008; McCarthy & Nobre, 1993) and supports the notion that spatial attention can modulate semantic analysis and integration of linguistic stimuli (Miniussi, Marzi, & Nobre, 2005). Specifically, the reduced N400 amplitude for unattended words has been interpreted as showing that semantic analysis is not automatic but depends on visual selective attention (e.g., McCarthy & Nobre, 1993). The topographic maps of the N400 effects in the present study may suggest that different subprocesses generating the N400 were involved in the spatial orienting compared to the other conditions as the distributions seem to be more posterior. However, the lack of statistically significant distribution differences makes it difficult to draw definite conclusions and renders answer to the question of specific subprocesses to future research. To conclude, the present study clearly demonstrates that attention can be actively oriented to the emotional category of upcoming word stimuli. Correct expectations about the emotionality of stimuli yield faster responses to these stimuli. Attention also operates at different processing stages. Attentional orienting to emotion was similar to orienting to a semantic category in the N400 response, but additionally entailed an earlier modulation of the P2 component. This early effect indicates attentional selection of stimulus characteristics that are processed prior to word semantics and ensures the rapid detection of potentially threatening situations, an evolutionary highly adaptive mechanism.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neuropsychologia.2011.07.022.
References Altarriba, J., & Bauer, L. M. (2004). The distinctiveness of emotion concepts: A comparison between emotion, abstract, and concrete words. American Journal of Psychology, 117, 389–410.
Altarriba, J., Bauer, L. M., & Benvenuto, C. (1999). Concreteness, context availability, and imageability ratings and word associations for abstract, concrete, and emotion words. Behavior Research Methods Instruments & Computers, 31, 578–602. Begleiter, H., & Platz, A. (1969). Cortical evoked potentials to semantic stimuli. Psychophysiology, 6, 91–100. Begleiter, H., Projesz, B., & Garazzo, R. (1979). Evoked brain potentials and behavior. In H. Begleiter (Ed.), Visual evoked potentials and affective ratings of semantic stimuli (pp. 127–143). New York: Plenum Press. Bernat, E., Bunce, S., & Shevrin, H. (2001). Event-related brain potentials differentiate positive and negative mood adjectives during both supraliminal and subliminal visual processing. International Journal of Psychophysiology, 42, 11–34. Biemann, C., Bordag, S., Heyer, G., Quasthoff, U., & Wolf, C. (2004). Languageindependent methods for compiling monolingual lexical data. Lecture Notes in Computer Science Berlin and Heidelberg: Springer., pp. 217–228 Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The Self-Assessment Manikin and the Semantic Differential. Journal of Behavior Therapy and Experimental Psychiatry, 25, 49–59. Carretie, L., Hinojosa, J. A., Martin-Loeches, M., Mercado, F., & Tapia, M. (2004). Automatic attention to emotional stimuli: Neural correlates. Human Brain Mapping, 22, 290–299. Cristescu, T. C., & Nobre, A. C. (2008). Differential modulation of word recognition by semantic and spatial orienting of attention. Journal of Cognitive Neuroscience, 20, 787–801. Cristescu, T. C., Devlin, J. T., & Nobre, A. C. (2006). Orienting attention to semantic categories. Neuroimage, 33, 1178–1187. Dien, J. (2009). The neurocognitive basis of reading single words as seen through early latency ERPs: A model of converging pathways. Biological Psychology, 80, 10–22. Eason, R. G. (1981). Visual evoked potential correlates of early neural filtering during selective attention. Bulletin of the Psyehonomie Society, 18, 203–206. Gomez, C. M., Vazquez, M., Vaquero, E., Lopez-Mendoza, D., & Cardoso, M. J. (1998). Frequency analysis of the EEG during spatial selective attention. International Journal of Neuroscience, 95, 17–32. Herbert, C., Kissler, J., Junghöfer, M., Peyk, P., & Rockstroh, B. (2006). Processing of emotional adjectives: Evidence from startle EMG and ERPs. Psychophysiology, 43, 197–206. Herbert, C., Junghöfer, M., & Kissler, J. (2008). Event related potentials to emotional adjectives during reading. Psychophysiology, 45, 487–498. Hillyard, S. A., & Anllo-Vento, L. (1998). Event-related brain potentials in the study of visual selective attention. Proceedings of the National Academy of Sciences of the United States of America, 95, 781–787. Hillyard, S. A., & Mangun, G. R. (1986). The neural basis of visual selective attention: A commentary on Harter and Aine. Biological Psychology, 23, 265–279. Hofmann, M. J., Kuchinke, L., Tamm, S., Vo, M. L., & Jacobs, A. M. (2009). Affective processing within 1/10th of a second: High arousal is necessary for early facilitative processing of negative but not positive words. Cognitive, Affective & Behavioral Neuroscience, 9, 389–397. Jasper, H. (1958). The ten-twenty electrode system of the international federation. Electroencephalography and Clinical Neurophysiology, 10, 371–375. Johannes, S., Munte, T. F., Heinze, H. J., & Mangun, G. R. (1995). Luminance and spatial attention effects on early visual processing. Brain Research Cognitive Brain Research, 2, 189–205. Kanske, P., & Kotz, S. A. (2007). Concreteness in emotional words: ERP evidence from a hemifield study. Brain Research, 1148, 138–148. Kissler, J., Assadollahi, R., & Herbert, C. (2006). Emotional and semantic networks in visual word processing: Insights from ERP studies. Progress in Brain Research, 156, 147–183. Kissler, J., Herbert, C., Winkler, I., & Junghofer, M. (2009). Emotion and attention in visual word processing: An ERP study. Biological Psychology, 80, 75–83. Kotz, S. A., & Paulmann, S. (2011). Emotion, language and the brain. Language and Linguistics Compass, 5, 108–125. Kutas, M., & Federmeier, K. D. (2000). Electrophysiology reveals semantic memory use in language comprehension. Trends in Cognitive Sciences, 4, 463–470. Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: Finding meaning in the N400 component of the event-related brain potential (ERP). Annual Review of Psychology, 62, 621–647. Luo, Q., Peng, D., Jin, Z., Xu, D., Xiao, L., & Ding, G. (2004). Emotional valence of words modulates the subliminal repetition priming effect in the left fusiform gyrus: An event-related fMRI study. Neuroimage, 21, 414–421. Maenoa, T., Gjinia, K., Iraminaa, K., Etob, F., & Ueno, S. (2004). Event-related potential P2 derived from visual attention to the hemi-space. Source localization with LORETA. International Congress Series, 1270, 262–265. McCarthy, G., & Nobre, A. C. (1993). Modulation of semantic processing by spatial selective attention. Electroencephalography and Clinical Neurophysiology, 88, 210–219. Miniussi, C., Marzi, C. A., & Nobre, A. C. (2005). Modulation of brain activity by selective task sets observed using event-related potentials. Neuropsychologia, 43, 1514–1528. Moore, C. C., Romney, A. K., Hsia, T.-L., & Rusch, C. D. (1999). The universality of the semantic structure of emotion terms: Methods for the study of inter- and intra-cultural variability. American Anthropologist, 101, 529–546. Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia, 9, 97–113. Ortigue, S., Michel, C. M., Murray, M. M., Mohr, C., Carbonnel, S., & Landis, T. (2004). Electrical neuroimaging reveals early generator modulation to emotional words. Neuroimage, 21, 1242–1251.
P. Kanske et al. / Neuropsychologia 49 (2011) 3121–3129 Paivio, A., Yuille, J. C., & Madigan, S. A. (1968). Concreteness, imagery, and meaningfulness values for 925 nouns. Journal of Experimental Psychology, 76(Suppl. 1–25). Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Romney, A. K., Moore, C. C., & Rusch, C. D. (1997). Cultural universals: Measuring the semantic structure of emotion terms in English and Japanese. Proceedings of the National Academy of Sciences of the United States of America, 94, 5489–5494. Rugg, M. D. (1985). The effects of semantic priming and work repetition on eventrelated potentials. Psychophysiology, 22, 642–647. Schacht, A., & Sommer, W. (2009). Time course and task dependence of emotion effects in word processing. Cognitive, Affective & Behavioral Neuroscience, 9, 28–43. Schapkin, S. A., Gusev, A. N., & Kuhl, J. (2000). Categorization of unilaterally presented emotional words: An ERP analysis. Acta Neurobiologiae Experimentalis (Wars), 60, 17–28. Schupp, H. T., Stockburger, J., Codispoti, M., Junghöfer, M., Weike, A. I., & Hamm, A. O. (2007). Selective visual attention to emotion. Journal of Neuroscience, 27, 1082–1089.
3129
Scott, G. G., O’Donnell, P. J., Leuthold, H., & Sereno, S. C. (2009). Early emotion word processing: Evidence from event-related potentials. Biological Psychology, 80, 95–104. Shedden, J. M., & Nordgaard, C. L. (2001). ERP time course of perceptual and post-perceptual mechanisms of spatial selection. Brain Research Cognitive Brain Research, 11, 59–75. Steinbeis, N., & Koelsch, S. (2009). Affective priming effects of musical sounds on the processing of word meaning. Journal of Cognitive Neuroscience. Stormark, K. M., Nordby, H., & Hugdahl, K. (1995). Attentional shifts to emotionally charged cues: Behavioral and ERP data. Cognition & Emotion, 9, 507–523. Treisman, A. M., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12, 97–136. van Voorhis, S., & Hillyard, S. A. (1977). Visual evoked potentials and selective attention to points in space. Perception & Psychophysics, 22, 54–62. Wierzbicka, A. (1992). Defining emotion concepts. Cognitive Science, 16, 539–581. Woldorff, M., Hansen, J. C., & Hillyard, S. A. (1987). Evidence for effects of selective attention in the mid-latency range of the human auditory event-related potential. Electroencephalography and Clinical Neurophysiology Supplement, 40, 146–154.