Nonemotional features suppress early and enhance late emotional electrocortical responses to negative pictures☆
Research highlights
▶ Emotional pictures vary on various nonemotional features that may confound effects of emotion. ▶ Emotional ERPs (EPN and LPP) to many negative and neutral pictures were measured. ▶ Nonemotional features suppress the EPN and enhance the LPP, particularly so in small picture sets.
Introduction
Research on event-related potentials (ERPs) to emotional pictures has found two peaks that appear to code emotional value for both positive and negative valence along the visual processing stream (Olofsson et al., 2008). Specifically, when ERPs to emotional pictures are subtracted from those to neutral pictures, an early posterior negativity (EPN) and a late positive potential (LPP) are commonly observed (Schupp et al., 2006). Both peaks can also be obtained by emotional words (Kissler et al., 2009, Schacht and Sommer, 2009). The EPN refers to a bilateral temporal-occipital negativity about 200 ms after picture onset. Source localization studies suggest that potential sources reside in the occipital cortex with parietal contributions (Junghöfer et al., 2001). Although the EPN to pictures appears superimposed on the N1 in several studies (Junghöfer et al., 2001, Keil et al., 2001), it can also occur later than the N1 (Foti et al., 2009, Schupp et al., 2003b). Because the topography and timing is similar to the selection negativity in studies of directed attention (Codispoti et al., 2006), the EPN may reflect a call for attentional resources that is triggered by emotional stimuli because of their intrinsic relevance for the organism (motivated attention, Schupp et al., 2006).
The LPP is a central-parietal positivity about 300 ms after picture onset (Cuthbert et al., 2000, Palomba et al., 1997). Research emphasizes that the traditional P300 or P3b peak (Bashore and Van Der Molen, 1991, Sutton et al., 1965) is similar to the LPP in that both index the allocation of attentional resources to salient events (Bradley, 2009, Hajcak et al., 2010, Olofsson et al., 2008). Traditionally, the term P300 refers to the ERPs that are elicited by top-down, explicit attention to neutral stimuli, whereas the term LPP refers to ERPs that are elicited by bottom-up, motivated attention to emotional stimuli (Ferrari et al., 2008). The LPP is associated with activations in lateral occipital, inferotemporal, and parietal visual areas (Sabatinelli et al., 2007). Although the LPP can last several seconds, research suggests that later slow wave responses represent processes different from the initial response around 300 ms (Foti et al., 2009). The LPP has been used to index attentional processes in aging (Kisley et al., 2007), spider phobia (Leutgeb et al., 2009, Norberg et al., 2010), drug abuse (Franken et al., 2008), and emotion regulation (Hajcak et al., 2010).
Emotional pictures often differ from neutral pictures not only on emotional variables (e.g., valence and arousal) but also on many other features such as spatial frequency (Delplanque et al., 2007) and human content (Colden et al., 2008). These features may not carry any emotional information yet influence and, thus, confound emotional responses. To reduce or eliminate these potential confounds, many previous studies have either matched emotional and neutral pictures on nonemotional features or compared responses between different participant groups (e.g., phobics vs. controls). Although picture matching is a valid strategy, it seems tedious, particularly so because people respond emotionally to a wide variety of visual stimuli that vary greatly on many nonemotional features and, thus, seem difficult to match. In fact, strong emotions can be elicited by pictures of sporting events, traffic accidents, erotic scenes, hospital scenes, landscapes, and action scenes (Britton et al., 2006, Lang and Bradley, 2010). Clearly, studies that sample broadly from these real-life situations enhance the ecological and external validity of their results.
So, although emotion shows effects on EPN and LPP when pictures are matched on nonemotional features, it is unresolved how large the confounding effects of nonemotional features are on EPN and LPP if emotional pictures are not matched for nonemotional features. In other words, if it could be shown that the confounding effects of nonemotional features are small or average out relative to effects of emotion, such results would suggest that it may be unnecessary to match pictures on nonemotional features.
Because no prior study has addressed this question (Olofsson et al., 2008), we selected 400 pictures in the range of neutral to negative valence from the International Affective Picture System (IAPS) that is a cross-culturally validated set of emotional and neutral pictures from various contexts and currently contains more than 1000 pictures (Lang et al., 2008). Although the IAPS is used in many ERP studies (Olofsson et al., 2008), the IAPS pictures vary substantially in many nonemotional features (Colden et al., 2008, Delplanque et al., 2007). For each picture, we determined the following nonemotional features:
- •
Picture composition: In a study by Bradley et al. (2007), ERPs to pictures with clear figure-ground organization (simple composition) were compared with ERPs to pictures with scenes (complex composition). The LPP to emotional (vs. neutral) pictures was relatively unaffected by composition. In contrast, the ERPs between 150 and 300 ms were strongly affected by composition: scenes elicited a stronger posterior positivity than did figure-ground compositions. Although no reliable EPN was obtained, effects of composition overlapped in time and location with those typically reported for the EPN (Schupp et al., 2006). These findings suggest that composition might confound EPN results obtained with IAPS. In fact, a recent study found a reverse EPN in that ERPs were more negative for neutral than negative IAPS pictures (Van Strien et al., 2009). To explain this reverse EPN, the authors argued that picture composition confounded their EPN results because negative pictures were more complex than neutral pictures and, thus, elicited a posterior positivity that suppressed the EPN.
- •
Human content: Colden et al. (2008) classified IAPS pictures as to whether they depicted something inanimate (i.e., objects), animal nonfaces, animal faces, human nonfaces, and human faces. Results showed that low-arousing, neutral pictures contained mainly objects, whereas high-arousing, emotional pictures contained mainly human faces and nonfaces. Although potentially confounding effects of content have yet to be determined, an effect on EPN might be predicted: faces elicit an N170 (Bentin et al., 1996), and this peak overlaps temporally and spatially with the EPN. Therefore, an apparent EPN to emotional versus neutral pictures might be enhanced by a greater proportion of faces in emotional than neutral pictures.
- •
Spatial frequency: As discussed by Delplanque et al. (2007), typical picture selections from the IAPS show clear differences in spatial frequency. For example, selections of pleasant and unpleasant pictures that were matched on arousal had more energy in all frequency bands than did neutral, low-arousing pictures. Therefore, confounding effects of spatial frequency on emotional ERPs cannot be ruled out (Boeschoten et al., 2005). Notably, the present question is whether energy in low and high spatial frequency bands per se affects EPN and LPP rather than whether effects of emotional content differ for different spatial frequencies. To illustrate, effects of emotional content for low frequencies would be studied with low-pass filtered emotional and neutral pictures that have identical energy in the low frequency bands (Alorda et al., 2007, Carretié et al., 2007, Pourtois et al., 2005).
- •
Luminance and contrast: In ERP studies, prime candidates for low-level confounds are luminance and contrast (Luck, 2005), but potentially confounding effects on emotional ERPs are unresolved.
In the present study, participants viewed IAPS pictures and rated the emotionality of each picture while ERPs were recorded. Data were aggregated over individual pictures to perform an item analysis over pictures. Multiple regression analyses with simultaneous and hierarchical models served to assess the independent or unique contributions (direct effects) of emotion and nonemotional features on EPN and LPP mean amplitudes. Subsequent mediation analyses served to determine whether correlations of emotion with EPN and LPP were mediated by nonemotional features (indirect effects of emotion). Because emotion studies commonly use fewer than 400 pictures, we also performed data simulations to determine whether stronger confounding effects of nonemotional features can be anticipated in studies with small picture sets, as might be expected because of greater random variation.
Section snippets
Participants
Forty-three students (30 women) with a mean age of 24.0 (age range: 19–37 years) were recruited from Stockholm University and Karolinska Institute as part of another study on the effects of a low dose of a benzodiazepine on emotional memory (Olofsson et al., in press). The study used informed consent and was approved by the regional review board. In a double-blind design, 22 participants (15 women) received a pill of 20 mg oxazepam whereas 21 participants (15 women) received placebo. Results
Results
Fig. 1 shows scatterplots of valence norms on arousal norms for the pictures in the 2008 IAPS pool (left panel) and for the IAPS pictures in the final sample (n = 333) together with IAPS pictures used in 13 other studies (right panel; Amrhein et al., 2004, Bradley et al., 2007, Cuthbert et al., 2000, Foti et al., 2009, Hajcak et al., 2006, Hajcak and Nieuwenhuis, 2006, Hot et al., 2006, Keil et al., 2002, Moser et al., 2006, Pollatos et al., 2005, Sabatinelli et al., 2007, Schupp et al., 2004,
Discussion
We performed an item-analysis of negative to neutral pictures from the IAPS database (n = 333). This analysis allowed us to assess the relative contribution of emotion and nonemotional features on EPN and LPP. The main findings were that (a) emotion (as a continuous variable) had independent effects on EPN and LPP; (b) nonemotional features, in particular composition, had substantial effects on EPN; (c) for studies that use few pictures and do not control for nonemotional features, effects of
Acknowledgements
Stefan Wiens thanks Sylvain Delplanque for his Matlab scripts and advice, and Carin Tillman, Emily Kappenman, and Andreas Keil for helpful discussions. Stefan Wiens also acknowledges the inspiration from Steven Luck and his ERP boot camp in 2009.
References (52)
- et al.
Low spatial frequency filtering modulates early brain processing of affective complex pictures
Neuropsychologia
(2007) - et al.
Modulation of event-related brain potentials during affective picture processing: a complement to startle reflex and skin conductance response?
International Journal of Psychophysiology
(2004) - et al.
Discovery of the P300: a tribute
Biological Psychology
(1991) - et al.
Time-varying differences in evoked potentials elicited by high versus low spatial frequencies: a topographical and source analysis
Clinical Neurophysiology
(2005) - et al.
Measuring emotion—the self-assessment mannequin and the semantic differential
Journal of Behavior Therapy and Experimental Psychiatry
(1994) - et al.
Facial expressions and complex IAPS pictures: common and differential networks
Neuroimage
(2006) - et al.
An electrophysiological study on the interaction between emotional content and spatial frequency of visual stimuli
Neuropsychologia
(2007) - et al.
Implicit and explicit categorization of natural scenes
- et al.
Brain potentials in affective picture processing: covariation with autonomic arousal and affective report
Biological Psychology
(2000) - et al.
Spatial frequencies or emotional effects? A systematic measure of spatial frequencies for IAPS pictures by a discrete wavelet analysis
Journal of Neuroscience Methods
(2007)
Event-related P3a and P3b in response to unpredictable emotional stimuli
Biological Psychology
An ERP investigation of emotional processing in European and Japanese individuals
Brain Research
Effects of emotional arousal in the cerebral hemispheres: a study of oscillatory brain activity and event-related potentials
Clinical Neurophysiology
Emotion and attention in visual word processing—an ERP study
Biological Psychology
Emotion and the motivational brain
Biological Psychology
An event-related potential study on exposure therapy for patients suffering from spider phobia
Biological Psychology
Visual event-related potentials index focused attention within bilateral stimulus arrays. 2. Functional dissociation of P1 and N1 components
Electroencephalography and Clinical Neurophysiology
Affective picture processing: an integrative review of ERP findings
Biological Psychology
Visual evoked potentials, heart rate responses and memory to emotional pictorial stimuli
International Journal of Psychophysiology
On the relationship between interoceptive awareness, emotional experience, and brain processes
Cognitive Brain Research
Emotions in word and face processing: early and late cortical responses
Brain and Cognition
Emotion and attention: event-related brain potential studies
Combined perception of emotion in pictures and musical sounds
Brain Research
Electrophysiological studies of face perception in humans
Journal of Cognitive Neuroscience
Natural selective attention: orienting and emotion
Psychophysiology
Brain potentials in perception: picture complexity and emotional arousal
Psychophysiology
Cited by (63)
The influence of social context and social connection on visual perceptual processes
2021, Acta PsychologicaTime-dependent effects of perceptual load on processing fearful and neutral faces
2020, NeuropsychologiaIndividual differences in processing emotional images after reading disgusting and neutral sentences
2020, NeuropsychologiaCitation Excerpt :Neutral images depicted simple objects (e.g. cups and plates). Due to the known influences on emotional ERP components from picture complexity (Wiens et al., 2011), all images were of a simple composition depicting only a single central entity against a background. The images were selected to provoke automatic affective responses (of either disgust or fear) without requiring the integration of information from multiple parts of the image to understand the scene.
The neural chronometry of threat-related attentional bias: Event-related potential (ERP) evidence for early and late stages of selective attentional processing
2019, International Journal of Psychophysiology
- ☆
This research was supported by grants from the Swedish Research Council and Stockholm Brain Institute to Stefan Wiens.