Modulation of event-related brain potentials during affective picture processing: a complement to startle reflex and skin conductance response?

Abstract

The present study compared startle response, skin conductance response (SCR) and subjective variables (valence and arousal ratings, viewing time) assessed in an affective picture paradigm with simultaneously registered event-related brain potentials (ERPs) parameters such as P300 and positive slow waves (PSW). Pleasant, neutral and unpleasant pictures from the International Affective Picture System [Lang, P.J., Bradley, M.M., Cuthbert, B.N., 1999. International Affective Picture System (IAPS): Instruction manual and affective ratings. Technical Report A-4, Center for Research in Psychophysiology. University of Florida, Gainesville, Florida] were presented for 8 s, and startle probes were delivered during picture presentation. Startle response was modulated by picture valence, and SCR by picture arousal. ERP positivity was greater for pleasant and unpleasant than for neutral pictures for the P300 amplitude and the positive slow wave (PSW). ERPs showed characteristic differences and a distinct time course for pictures of different valence categories and may deliver useful information not contained in startle response or SCR measures. The simultaneous registration of startle responses and ERPs in the affective picture paradigm seems valuable.

Introduction

Lang et al. Lang, 1985, Lang et al., 1998 postulated that emotions can be measured on two basic dimensions: affective valence (from an appetitive to an aversive pole) and arousal. In order to systematically vary emotional valence and arousal during experiments, they have collected a large set of photographic stimuli (International Affective Picture System, IAPS; Lang et al., 1999). In the so-called “affective picture paradigm”, these pictures are presented typically for several seconds (6–8 s), and subjective as well as physiological responses are registered. Covariation between dependent measures and emotional valence and arousal of picture stimuli can be analysed (see Lang et al., 1998, for a review).

Psychophysiological variables of interest during affective picture processing have been skin conductance response (SCR) and heart rate (HR) changes. SCR was found to be a reliable indicator of arousal (e.g. Bradley et al., 1993), while HR systematically varies with picture valence (Vrana et al., 1989). Reactions triggered by probe stimuli delivered during picture viewing may also serve as indicators of emotional processing. For example, the startle response triggered by acoustic startle stimuli and quantified by the amplitude of its eye-blink component was found to be increased during viewing of aversive pictures and to be reduced for appetitive pictures (e. g. Vrana et al., 1988, Hamm et al., 1997). However, later studies demonstrated that affective startle modulation is also influenced by arousal, with stronger modulation effects for high- than for low-arousing pleasant and unpleasant stimuli Schupp et al., 1997, Hamm et al., 1997.

Event-related brain potentials (ERPs) have been studied in a variety of cognitive paradigms, but less extensively as indicators of emotional processes. However, they may provide important information about cortical activity evoked by emotional stimuli (e. g. Pauli et al., 1997). Two positive ERP components are of special interest: the P300 (a positive peak occurring approximately 300 ms after stimulus onset) seems to represent attentional processes and processing of new or meaningful stimuli (see Donchin, 1987, Rockstroh et al., 1989, for a review), and the Positive Slow Wave (PSW), which occurs 400 ms after stimulus onset and beyond, and has been found to covary with controlled cognitive processing (Rösler, 1986) and with evaluation and memory storage (Rohrbaugh et al., 1984).

There are two approaches to studying the influences of emotional stimuli like IAPS pictures on ERPs. On the one hand, two studies Schupp et al., 1997, Cuthbert et al., 1998 examined the effects of affective foreground stimuli (IAPS pictures) on the P300 elicited by probe stimuli (startling noises or nonstartling tones). The probe P300 was primarily modulated by picture arousal, with smaller amplitudes elicited when viewing affective (pleasant and unpleasant) as compared to neutral pictures. This reduction of the probe P300 presumably reflects a greater allocation of attentional resources to affective visual stimuli (Cuthbert et al., 1998).

On the other hand, several studies examined ERPs triggered directly by emotional stimuli, adjectives (Naumann et al., 1992), self-selected affective pictures Johnston et al., 1986, Johnston and Wang, 1991 or IAPS pictures (e. g. Palomba et al., 1997, Keil et al., 2002). These studies found the P300 as well as the PSW to be enhanced for pleasant and unpleasant compared to neutral picture stimuli Mini et al., 1996, Diedrich et al., 1997, Palomba et al., 1997, Cuthbert et al., 2000, Schupp et al., 2000, Keil et al., 2002. Cuthbert et al. (2000) also found that the PSW was more positive for the higher arousing compared to the lower arousing pictures within each category. Overall, the enhanced late positive components of the ERPs (from about 400–600 ms on), triggered by emotional stimuli seem to be associated with the arousal dimension and presumably reflect sustained attentive processing of emotionally relevant stimuli (Cuthbert et al., 2000).

Finally, some studies Diedrich et al., 1997, Palomba et al., 1997, Cuthbert et al., 2000 observed an enhanced positivity in an earlier latency range (200–300 ms) only for pleasant pictures. This effect was considered to be elicited by picture valence (Diedrich et al., 1997), but it could also be due to other characteristics differing between pleasant and unpleasant pictures. Recent studies Junghöfer et al., 2001, Schupp et al., 2003, however, with brief presentations of affective pictures and multi-channel ERP recordings confirmed the effect of emotional arousal on early ERPs. Besides producing an enhanced late PSW, emotional stimuli also affected early ERP components (as early as 200 ms after picture onset), which can be related to sources of activation in the extended visual system. The early ERP effects presumably reflect facilitated sensory encoding of affective stimuli by early implicit selective attention, while the later ERP effects seem to be related to higher stages of processing (Schupp et al., 2003).

In order to understand the significance of ERPs triggered by emotional stimuli, it seems important to simultaneously record other variables — both psychophysiological and self report measures — and to examine their relationship. ERP, startle and peripheral responses may complement each other. They reflect different response systems (cortical versus subcortical) which differ in response speed and variability. However, only a few studies have done this systematically. Palomba et al. (1997) measured heart rate and memory performance, and found significant correlations between EEG positivity from 600 to 900 ms and heart rate deceleration, and between P300 amplitude and the number of remembered picture stimuli. Cuthbert et al. (2000) recorded heart rate, skin conductance, corrugator EMG, and valence and arousal ratings. EEG positivity from 400 to 1000 ms loaded on the same factor as SCR and arousal ratings.

The present study used an affective picture paradigm to evaluate changes in EEG parameters such as P300 and PSW related to the processing of emotionally pleasant, neutral and unpleasant pictures. PSW was analysed in the latency range from 200 to 2000 ms after picture onset, since this time window seems appropriate to capture both early and late ERP components (see Cuthbert et al., 2000) and was not confounded by startle probes, which were delivered earliest 2.5 s after picture onset. Startle reflex modulation, SCR, viewing time and subjective valence and arousal ratings were also assessed. Since only few studies simultaneously recorded EEG responses and startle reflex modulation, we were especially interested in examining the association between these variables.

The aims of the present study were:

• to replicate the well-known effects (e.g. Lang et al., 1993) of affective pictures on valence and arousal ratings, viewing time, startle reflex magnitude and SCR amplitude;

• to confirm previous findings (e.g. Cuthbert et al., 2000) that P300 and PSW are enhanced for pleasant and unpleasant pictures compared to neutral ones;

• to evaluate differences in the time course of the PSW for pleasant, neutral and unpleasant pictures;

• to demonstrate the concurrent validity of ERPs and startle responses in the affective picture paradigm.