Skip to main content
PMC home page
Social Cognitive and Affective Neuroscience logoLink to Social Cognitive and Affective Neuroscience
. 2009 Mar 6;4(2):191–198. doi: 10.1093/scan/nsp004

Wait, what? Assessing stereotype incongruities using the N400 ERP component

Katherine R White 1,✉, Stephen L Crites Jr 1, Jennifer H Taylor 1, Guadalupe Corral 1
PMCID: PMC2686231  PMID: 19270040

Abstract

Numerous discoveries regarding stereotypes have been uncovered by utilizing techniques and methods developed by cognitive psychologists. The present study continues this tradition by borrowing psychophysiological techniques used for the study of memory and language, and applying them to the study of stereotypes. In this study, participants were primed with either the gender category ‘Women’ or ‘Men’, followed by a word which was either consistent with gender stereotypes (e.g. Women: Nurturing) or inconsistent (e.g. Women: Aggressive). Their task was to indicate whether the words matched or did not match, according to gender stereotypes. Both response times and event-related brain potentials (ERPs) were recorded during performance of the task. As predicted, stereotype incongruent word pairs were associated with larger N400 ERP amplitudes and slower response times, relative to congruent word pairs. The potential utility of this approach as an independent measure of stereotypes is discussed.

Keywords: ERPs, N400, stereotypes, social cognition, gender

The primary objective of social cognition is to increase our understanding of how people perceive, evaluate and act toward other people and social groups. Social cognition has borrowed heavily (and fruitfully) from approaches used and developed in cognitive research. For instance, by utilizing procedures developed for the study of memory, learning and language comprehension, social cognition researchers have uncovered numerous insights into the possible cognitive mechanisms underlying stereotypes and prejudice (Gaertner and McLaughlin, 1983; Dovidio et al., 1986; Banaji and Hardin, 1996; Blair and Banaji, 1996; Wittenbrink et al., 1997, 2001; Oakhill et al., 2005). Building upon this research, the present study implemented ideas and methods from cognitive neuroscience to explore stereotypes.

To identify cognitive mechanisms underlying stereotypes, it is important to maintain a conceptual distinction between stereotypes and prejudice. While stereotypes refer to beliefs associated with a social group (usually held at the cultural level), prejudice generally consists of negative affect held toward a social group (Amodio and Devine, 2006). For example, encountering an African American may activate stereotypes that African Americans are good at sports and expressive. On the other hand, encountering an African American might activate prejudice in an individual who grew up in a family that taught intolerance of other social groups. This distinction has not always been maintained in previous research because stereotypes and prejudice are often perceived as unavoidably connected (Judd et al., 2004) and because both can be examined using conventional priming procedures (Fazio et al., 1995; Wittenbrink et al., 1997; Kawakami et al., 1998). This has led to difficulties when interpreting whether results are due to stereotype or prejudice activation (Judd et al., 2004). Although prejudice and stereotypes often act in concert, recent research suggests that stereotypes and prejudice may depend upon partially separate neural pathways/processes and can lead to different forms of discriminatory behavior (Damasio, 1989; Mackie and Smith, 1998; Davis and Whalen, 2001; Amodio and Devine, 2006). Thus, to fully understand discrimination, it is vital to maintain a conceptual distinction between the two constructs.

One approach for separating stereotypes and prejudice, and exploring the cognitive processes that underlie each construct has been to use well-controlled experimental paradigms (Judd et al., 2004; Amodio and Devine, 2006). Although these paradigms have been very useful, they have relied primarily upon behavioral data in the form of response latencies, which have inherent limitations. Specifically, response latencies have limited sensitivity; a response is executed hundreds of milliseconds after stimulus perception, during which time multiple processes have taken place (Ito and Cacioppo, 2000). Although response times and accuracy rates can be used to make inferences regarding a sequence of cognitive processes (Gaertner and McLaughlin, 1983; Dovidio et al., 1986; Blair and Banaji, 1996), these inferences are limited and a more direct measure of intervening processes, such as event-related potentials (ERPs), would be advantageous.

When used in concert with behavioral measures, ERPs have provided researchers with a more complete approach to the study of cognitive processes (Coles, 1989). ERPs are considered electrical manifestations of brain activity in response to, or in preparation for, specific events (Fabiani et al., 2007). That is, when people encounter a stimulus, a sequence of neural units activates to process this stimulus and prepare a response. The electrical activity associated with the entire sequence is referred to as the ERP, and segments of the ERP, which are referred to as components (e.g. N100, N400, P300), reflect relatively discrete neural units or collections of neural units that are performing a more specific cognitive process. Because ERPs provide an instantaneous assessment of underlying neural activity, they allow for easier identification and isolation of individual cognitive processes (Rhodes and Donaldson, 2008). By examining the way various factors affect these ERP components, cognitive research has gleaned important information on attention, categorization, memory, language processes, error monitoring and expectancy (Fabiani et al., 2007). Amodio and colleagues (2006), for instance, have recently demonstrated how one ERP component, which is referred to as either the error related negativity (ERN) or medial-frontal negativity (MFN), can be used to explore processes involved with regulation and control of stereotypes and prejudice. This component is thought to reflect response monitoring and conflict detection (Johnson et al., 2004) and therefore is useful for studying stereotype control and inhibition, but not necessarily stereotype activation. Another component of the ERP, the N400, is specifically sensitive to modulations of meaning at the semantic/associative level (Kutas and Federmeier, 2000), and may therefore provide a means of exploring stereotype activation.

The N400 is a negative deflection manifesting ∼400 ms following stimulus presentation, and is modulated by violations at the level of semantics or meaning. More specifically, the N400 is thought to reflect the ease of integrating a stimulus into a given context, based upon the cognitive effort needed to access information from long-term memory (for review, see Kutas and Federmeier, 2000). Items that are incongruent with the context, compared to items that are congruent, elicit N400s with larger amplitudes. Although first observed in response to semantically incongruent sentence endings (Kutas and Hillyard, 1980, 1984), the N400 effect is also elicited by the second word in semantically incongruent word pairs (Bentin et al., 1985). Word pair studies have been primarily used to examine associations in lexical memory, with larger N400 amplitudes indicating more difficulty in accessing information (Kutas and Federmeier, 2000). For example, the word ‘JELLY’ is more difficult to access from lexical memory when primed with ‘SHOE’ than when primed with ‘GRAPE’. Therefore, a larger N400 amplitude is associated with the word JELLY in the incongruent word pair SHOE: JELLY than with the congruent word pair GRAPE: JELLY. Because previous stereotype research has operated under the assumption that stereotypes are a specific class of semantic associations stored in memory (Gaertner and McLaughlin, 1983; Dovidio et al., 1986; Blair and Banaji, 1996), one might hypothesize that incongruities between stereotypically associated word pairs elicit the N400 effect. For example, within the context of ‘MEN’, the incongruent word ‘SKIRT’ would elicit a larger amplitude N400 than the congruent word ‘BRIEFCASE’. However, research to date has not demonstrated that the N400 is reliably modulated by stereotype incongruities.

Only a few researchers have attempted to use the N400 to investigate stereotypes. Osterhout et al. (1997) presented participants with sentences containing stereotypically male or female occupations and pronouns that did or did not match the gender stereotypically implied by the occupation (e.g. The beautician put herself through school vs The beautician put himself through school). More recently, Bartholow and colleagues (2001, 2003) sought to examine the cognitive processing of social expectancy violations, which they likened to the violation of expectancies based upon stereotypes. Participants were presented with a series of sentences that first established the personality traits of a target person and then were presented with a sentence that described the target's behavior in a manner that violated the previously established personality traits. For the above studies, contextual violations modulated a late positive potential (LPP) of the ERP but did not affect the N400. However, the failure to find the N400 effect might be due to the task and the participants’ processing strategies in these experiments, as research has shown the ERP to be dependent upon these factors (Mari-Beffa et al., 2005; Roehm et al., 2007). We therefore sought to implement a task and paradigm commonly used in both N400 and stereotype research.

The present research sought to determine whether incongruities within stereotypically associated word pairs elicit the N400 effect. We hypothesized that stereotypically incongruent word pairs would evoke larger N400 amplitudes, relative to those evoked by stereotypically congruent word pairs. A secondary objective was to replicate response latency effects observed in previous research on stereotypes (Banaji and Hardin, 1996; Blair and Banaji, 1996; Oakhill et al., 2005). We hypothesized that stereotypically incongruent word pairs would result in longer response latencies, relative to those associated with stereotypically congruent word pairs.

METHOD

Participants

The sample included data from 23 participants (13 female), all in good health. The age of the participants ranged from 18 to 27 with an average age of 20. For ethnicity, 16 participants self-identified as Hispanic/Latino, and the remaining as Not Hispanic/Latino. For race, 11 participants self-identified as White, two Native American, one Asian, one African American and seven as other. No specific exclusionary criteria were enforced.

Stimuli

Stimuli were word pairs that consisted of a gender category (Men or Women) followed by a target word. The target word was either a trait (e.g. nurturing, aggressive) or non-trait (e.g. makeup, mechanic) stereotypically associated with one of the gender categories. Target words included 14 female traits (e.g. caring, gossipy), 14 male traits (e.g. aggressive, powerful), 14 female non-traits (e.g. secretary, doll) and 14 male non-traits (e.g. engineer, cigars) chosen from previous gender stereotype research (Bem, 1974; Blair and Banaji, 1996; Oakhill et al., 2005). Each target word was paired with both gender categories, making a total of 28 congruent word pairs (e.g. Women : Caring; Men : Engineer) and 28 incongruent word pairs (e.g. Women : Cigars; Men : Gossipy).

Procedure

Participants first read and signed an informed consent form and were prepared for EEG recording. Each participant was then taken to an isolated room and seated in a chair ∼0.5 m in front of a computer monitor. The experimenter informed participants that they would see a series of word pairs and that their task was to indicate whether the second word matched or did not match the gender category that preceded it, according to social gender stereotypes.1 Thus, participants made a binary decision. Each of the four experimental blocks contained 120 word pairs—the first eight word pairs were practice trials containing target words not used in the experimental trials and the remaining 112 were experimental trials. Word pairs were flashed centrally upon the computer monitor and participants indicated their response via a response pad. Prime words were preceded by a 200 ms focus ‘+’ and a 100 ms blank screen. The prime word was then presented for 150 ms followed by a 100 ms ISI before the target word appeared for 150 ms. The target word was followed by a 100 ms blank screen and then a 1050 ms response window displaying a ‘Match/Mismatch’ instruction. The inter-trial interval was 1500 ms. If a participant failed to respond within the 1050 ms, the program automatically continued to the next word pair.

Data acquisition and reduction

Bioelectrical activity was recorded using Ag/AgCl electrodes from 29 scalp locations and referenced to the right mastoid. Vertical electrooculographic (VEOG) and horizontal electrooculographic (HEOG) activity were also recorded. Neuroscan amplifiers were used to amplify, filter (bandpass of 0.05–30 Hz) and digitize (1000 Hz) the bioelectrical signals that were recorded continuously during the experiment.

A number of steps were taken to reduce and quantify the bioelectrical data. The EEG data were re-referenced to a digital, linked-mastoids reference (Hagemann et al., 2001). Epochs associated with each prime–target stimulus pair were extracted, and each epoch and electrode site was baseline corrected to the mean of its pre-stimulus period. A regression procedure for removing VEOG artifacts from the EEG recordings was applied (Semlitsch et al., 1986) after epochs containing extreme activity at VEOG were excluded. Data were manually reviewed, and electrodes were deleted from further analyses if there was a problem (e.g. if an electrode came loose). Epochs containing extreme activity at any scalp site (± 60 μV) were excluded from further analyses. The EEG recordings over each recording site for each participant were averaged separately within each of the experimental conditions.

RESULTS

Prior to conducting analyses, response accuracy and response latency were transformed (arcsine and log transformations, respectively) so that the assumptions of the ANOVA model were met. For ease of interpretation, however, accuracy is reported in percentages and response times in milliseconds. Participants were considered accurate if their match/mismatch judgments aligned with shared cultural gender stereotypes, as determined by prior research and pilot testing in the local population. For accuracy scores and response latencies, we conducted two 2 (Prime–Target Congruency) × 2 (Prime: Men/Women Category) × 2 (Target: Trait/non-trait Word) ANOVAs (only correct trials were included for response latencies). We examined the N400 amplitude using a 2 (Prime–Target Congruency) × 2 (Prime: Men/Women Category) × 2 (Target: Trait/non-trait Word) × 5 (Electrode site: Fz, FCz, Cz, CPz, Pz)2 ANOVA (only correct trials were included in the analysis). Because physiological data frequently violate the sphericity assumption of the univariate ANOVA, we assessed significance using a multivariate (Wilks’ Lambda) test and report the F-statistics and significance associated with the Wilks’ Lambda.

Effect of congruency

Behavioral data replicated previous stereotype research. Participants demonstrated relatively high accuracy with regards to identifying when target words were a stereotypical match or mismatch with the preceding gender category (M = 78.39%). As was the case in Oakhill et al. (2005), participants were more accurate for congruent word pairs (M = 86.02%) than for incongruent word pairs (M = 70.75%), F(1,22) = 27.78, P < 0.001, ω2 = 0.538. For response latencies, participants were faster to respond to congruent word pairs (M = 711 ms) than incongruent word pairs (M = 793 ms), F(1,22) = 67.88, P < 0.001, ω2 = 0.744. This also replicates prior research (Banaji and Hardin, 1996, Experiment 2; Blair and Banaji, 1996; Oakhill et al., 2005), in which priming effects were interpreted to reflect spreading activation in associative memory (Collins and Loftus, 1975; Neely, 1977; Gawronski and Bodenhausen, 2005).

The expected ERP effect for congruency was also found. We hypothesized that negative amplitude in the 300–500 ms time window would be larger following incongruent word pairs than following congruent word pairs. As can be seen in Figure 1, N400 peak amplitude was more negative for incongruent word pairs (M = −3.69 μV) relative to congruent word pairs (M = −2.77 μV), F(1,17) = 11.58, P = 0.003, ω2 = 0.315.3 Note that the N400 is ∼650 ms after the onset of the prime in the figure, which equates to 400 ms after target onset (250 SOA between prime and target words). Due to its latency (within the window of 300–500 ms) and modulation by word pair congruency, we believe that the effect observed belongs to the N400 family.4

Fig. 1.

Fig. 1

Open in a new tab

Effect of congruency upon N400 amplitude. The five graphs depict ERP waveforms over the frontal (Fz) to posterior (Pz) midline scalp sites. The solid line reflects stereotypically congruent word pairs (e.g. Women : Secretary) and the dashed line reflects stereotypically incongruent pairs (e.g. Women : Mechanic). The PRIME was presented at 0 ms and the TARGET at 250 ms. Thus, the N400 which is a negative-going potential (negative is plotted up in the graphs) in response to TARGET words is observed at 650 ms.

Effect of trait vs non-trait target words

Accuracy, response latency and ERP amplitude all revealed effects for the type (trait/non-trait) of target word. Accuracy scores were higher for non-trait target words (M = 83.72) than for trait target words (M = 73.06), F(1,22) = 90.51, P < 0.001, ω2 = 0.796. Faster reaction times occurred for non-trait target words (M = 481 ms) than trait target words (M = 522 ms), F(1,22) = 50.64, P < 0.001, ω2 = 0.683, conceptually replicating previous findings (Blair and Banaji, 1996). Finally, the N400 amplitudes associated with word pairs using non-trait target words (M = −4.23 μV) were significantly more negative than those associated with word pairs using trait target words (M = −2.23 μV), F(1,17) = 50.472, P < 0.001, ω2 = 0.683.

Exploratory analyses

Because previous stereotype manipulations demonstrated modulation of a late positive potential (Bartholow et al., 2001; Osterhout et al., 1997), the LPP amplitude (quantified as the most positive peak 700–1200 ms following prime presentation; 450–950 ms following target presentation) was examined using the same analysis performed for the N400. In contrast with previous research (Bartholow et al., 2001; Osterhout et al., 1997), the late positive potential was not modulated by congruency (F < 1). All other effects were non-significant.

Similar analyses were also conducted on response latencies, the N400 and the LPP with participant gender as an additional factor. While the addition of participant gender had no effect upon the N400 results, several interesting interactions arose for response latencies and the LPP. There was a significant interaction for response latency between prime gender category and participant gender, F(1,21) = 17.587, P < 0.001, and also a significant three-way interaction between participant gender, gender category and congruency for response times [F(1,21) = 14.363, P = 0.001] and the LPP [F(1,15) = 7.428, P = 0.016] (Table 1). For congruent word pairs, all participants were quicker to make judgments when the prime gender category matched their own. For incongruent word pairs, this pattern held for male participants, but not for female participants. For the LPP, when word pairs were incongruent participants exhibited larger amplitudes to word pairs with a prime gender category that matched their own gender. For congruent word pairs, this pattern held for female participants but not for male participants.5

Table 1.

Simple comparison means for response times and LPP amplitude (Participant Gender × Prime Gender Category × Congruency)

Female participants
Prime Target Mean response times Mean RT difference Mean LPP amplitude Mean LPP difference
Men Congruent 490 ms −83 ms** 9.00 μV 0.36 μV
Women Congruent 407 ms 9.36 μV
Men Incongruent 526 ms −7 ms 8.65 μV 0.71 μV
Women Incongruent 519 ms 9.36 μV
Male participants
Men Congruent 431 ms −50 ms* 5.19 μV −0.90 μV
Women Congruency 481 ms 6.09 μV
Men Incongruent 542 ms −24 ms* 6.16 μV 0.57 μV
Women Incongruent 566 ms 5.59 μV
Open in a new tab

Mean RT difference reflects decreases in response times associated with a match between prime category and participant gender. Mean LPP difference reflects increases in LPP amplitude associated with a match between prime category and participant gender.

*P <.04; **P <.01.

DISCUSSION

Findings from the present study replicate previous behavioral research and suggest that the N400 component is sensitive to violations of gender stereotypes in an associative priming paradigm. Participants made quicker judgments for stereotypically congruent prime–target word pairs (Women : Nurturing), relative to incongruent word pairs (Men : Nurturing). This congruency effect was also present in the ERP data—congruent prime–target word pairs elicited smaller N400 amplitudes relative to incongruent word pairs. Together, these findings suggest that target words were easier to access from associative memory due to facilitation from the preceding context of ‘Men’ or ‘Women’ when they matched gender stereotypes (Kutas and Federmeier, 2000; Gawronski and Bodenhausen, 2005). For example, target words such as ‘Nurturing, Gossipy, Flowers, and Secretary’ were easier to access and integrate when the preceding prime context was ‘Women’ than ‘Men’. This led to a decrease in both the N400 amplitude and response latency.

We have interpreted the observed effect as the N400, based on behavioral and ERP similarities with N400 research on semantic priming, and believe that the effect reflects differential memory activation due to priming (Bentin et al., 1985; Brown and Hagoort, 1993; Koivisto and Revonsuo, 2001; Franklin et al., 2007; Rhodes and Donaldson, 2008). However, given that this is the first demonstration of N400 modulation due to stereotypes, it is important to consider other possible explanations. One issue is the extent to which the observed effect reflects neural activity associated solely with the prime. That is, because target onset was 250 ms after prime onset, a portion of the ERP after target onset likely reflects neural processes associated with the prime and not with the target. Because there were only two primes (the words ‘Men’ and ‘Women’) that each appeared in half of the trials, we systematically examined the entire ERP associated with these primes (i.e. collapsing across target congruency). The only effect was a larger P2 (peak latency = 198 ms) associated with the prime ‘Women’ compared to ‘Men’. There was no evidence of any ERP differences associated with the primes after the onset of the target. These analyses, in conjunction with the large extant literature that has observed comparable N400 effects regardless of long or short SOAs (Boddy, 1986; Holcomb and Anderson, 1993; Anderson and Holcomb, 1995; Babad, 1997; Chwilla et al., 1998; Hill et al., 2002), suggest that the observed effects reflect the incongruity between prime and target rather than neural processes associated with the prime (Rugg, 1985; Kiefer et al., 1998; Deacon et al., 2000; Franklin et al., 2007).

Another potential explanation for the observed ERP effect is that it reflects match/mismatch processes and not spreading memory activation. Previous literature indicates that other negative-going ERP components which peak between 200 and 300 ms (N2 and N250) are modulated by stimulus mismatches of a perceptual, physical or orthographic nature (Holcomb and Grainger, 2007; Folstein and Van Petten, 2008). That is, stimuli that are physically discrepant from preceding stimuli evoke a large amplitude potential ∼200 ms following stimulus onset. It is possible that the N400 observed here is a result of this type of mismatch effect, and occurs later than has been previously observed because it reflects a more complicated semantic/associative mismatch rather than a mismatch in the physical nature of stimuli. Additional research will be needed to determine whether the effects observed here reflect spreading memory activation (less negative potential when the prime is associated with the target) or a semantic mismatch (more negative potential when the prime and target are incongruent). At this time, we are more inclined to think of the effect as spreading activation because the stimuli are more similar to those used in N400 research, and the conceptualization of stereotypes as semantic/associative memory structures fits with existing N400 research and theory. One way to disentangle these two possibilities is to explore whether the effect remains when the task does not require matching the prime and target. If the effect observed here is due to spreading activation, it should remain even when there is not an explicit matching requirement. If it is due to matching, however, the effect should disappear or be significantly diminished (Folstein and Van Petten, 2008).

It is also possible that the observed ERP effect reflects response conflict processes and not spreading memory activation. Research using ERPs to explore response conflict has examined either stimulus-locked ERPs (EEG is synchronized according to stimulus onset), as was done in this study, or response-locked ERPs (EEG is synchronized according to response onset). In stimulus-locked ERPs, response conflict can be seen in more negative-going potentials that peak between 200 and 500 ms after stimulus onset, depending on the complexity of the stimulus (Folstein and Van Petten, 2008). For example, when the Stroop task is used, incongruent stimuli are associated with a negative deflection peaking around 450 ms (Folstein and Van Petten, 2008). Because the stimulus-locked ERP negativity evoked by response conflict is often accompanied by an anterior LPP (Folstein and Van Petten, 2008), analyses were conducted on the LPP to investigate whether the observed effect might be due to response conflict. Findings revealed no significant LPP effects. In addition, we derived and examined response-locked ERPs to explore whether there was evidence that response conflict occurred during execution of the participants’ task (Coles et al., 1985; Johnson et al., 2004). There was no difference between the ERPs evoked by congruent and incongruent word pairs and thus no evidence for response conflict. Finally, the nature of the task argues against response conflict as an explanation. That is, in most response conflict research, a stimulus or stimulus set is presented that activates two different responses (Klauer and Teige-Mocigemba, 2007)—a central one that must be executed based on the task instructions (e.g. respond to the color of ink in the Stroop task) and an irrelevant one that must be inhibited (the name of the color word in the Stoop task). In the present study, there was no ‘irrelevant’ stimulus because the task required participants to compare the two stimuli. Thus it is unlikely that two competing responses were activated. The lack of any ERP effects associated with response conflict in conjunction with the nature of the task suggests that the observed effects were probably not driven by response conflict.

The findings of this study suggest that the N400 is modulated by stereotype violations, however, two previous studies found no such effect (Bartholow et al., 2001; Osterhout et al., 1997). The N400 is thought to reflect the difficulty of accessing information from semantic/associative memory (Kutas and Federmeier, 2000). This form of memory is accumulated slowly over a long period of time, and a larger N400 is elicited when the context does not facilitate access to associative memory (Kiefer, 2007). The present study required participants to use this form of memory and demonstrated an N400 effect when the target was incongruent with stereotypes stored in associative memory. In Bartholow and colleagues’ (2001) study, violations were manipulated differently. In their study, participants were presented with a series of sentences which first established the personality traits of a target person and then violated behavioral expectations of the same target person. Social expectancies were therefore created within the context of the experimental design, and violations constituted a mismatch between the local context and the target stimulus rather than a lack of facilitation in associative memory. Bartholow and colleagues found no evidence of an N400, which is consistent with the idea that the N400 effect reflects facilitated access to associative memory. They also found that the LPP was greater when there was a mismatch with the local context, which is consistent with research on the LPP (Nieuwenhuis et al., 2005).

The effects observed by Osterhout and colleagues (1997) may also be due to a mismatch between local context and stimuli. In their study, participants were presented with sentences which contained stereotypically male or female occupations and pronouns that did or did not match the gender stereotypically implied by the occupation (e.g. The beautician put herself through school vs The beautician put himself through school). Incongruent sentences elicited an LPP, which the authors interpret as an indication that stereotype incongruent pronouns were perceived as grammatically incorrect. In other words, the stereotypical occupations served as the local grammatical context for the pronouns rather than facilitating access to associative memory. Moreover, because the LPP is very large, it can obscure the N400 when present (Chwilla et al., 1998; Franklin et al., 2007) so it is possible that any N400 effect in Osterhout and colleagues’ findings (1997) was hidden by the large LPP. Together, these studies indicate that to observe the effect of stereotypes on the N400 a paradigm that forces participants to directly access associative memory may be important.

The primary purpose of this research was to introduce a potential tool that may be used in addition to behavioral measures for the study of stereotypes. In light of this purpose, we wish to avoid over-speculation with regards to the theoretical implications of the findings. A couple possible theoretical applications do, however, bear mentioning. The first is that the N400 ERP component in concert with behavioral data may provide an alternative means for independently investigating stereotypes. Because both stereotype and prejudice priming lead to response latency effects (Wittenbrink et al., 1997), it becomes difficult to separately identify the mechanisms that underlie these constructs without placing increasingly restrictive controls upon the paradigm (Judd et al., 2004; Amodio and Devine, 2006). The present study has demonstrated that the N400 is sensitive to stereotype incongruities, while also replicating response facilitation. Additionally, we recently conducted two studies in which word-pair evaluative incongruities modulated response latencies, but failed to modulate the N400 amplitude (Taylor et al., 2008). When combined with the findings of the present study, these results indicate that only semantic/associative priming modulates the N400 ERP component. Thus the N400 may potentially be developed as an additional tool to study the cognitive mechanisms underlying stereotypes.

The second application is that the N400 component's sensitivity to context may play a role in the progression of stereotype theory. Originally researchers thought of stereotypes as concrete, automatically activated, immutable knowledge structures. However, accumulating research indicates that the activation of stereotypes is rather malleable, depending upon factors such as context (for review, see Blair, 2002). These findings have typically been explained with theoretical footnotes, thereby preserving the original notion that stereotypes are rather impervious to change. Yet, there are those who would argue that this approach has placed a roadblock in the progression of theory, and that focusing upon the impact of context on stereotype activation would lead to far more theoretical progression (Smith and Semin, 2007). Because the N400 component is modulated by contextual variations and constraint (Kutas and Federmeier, 2000), it may be a promising choice for studying the impact of context upon stereotype activation.

In summation, the present study demonstrated that the N400 ERP component is modulated by violations of gender stereotypes in a sequential associative priming paradigm. This effect was interpreted as a reflection of facilitated lexical access and context integration. The authors argue that given its sensitivity to context, and responsiveness to semantic/associative relationships rather than evaluative relationships, the N400 may prove to be a useful tool for the progress of social cognitive theory.

Footnotes

1

An explicit matching task was utilized because previous research has indicated that using an explicit task has the effect of enlarging N400 amplitudes (Chwilla et al., 2000; Kreher et al., 2006).

2

An analysis was also run to explore hemisphere effects. N400 amplitudes were larger over the left hemisphere, F(1,10) = 5.616, P = 0.039. No other hemisphere effects were significant.

3

This effect was also significant when the analysis was conducted upon all trials (correct and incorrect), F(1,17) = 7.157, P = 0.016.

4

There was also a main effect for site, F(4,14) = 15.787, P < 0.001, and an interaction with congruency, F(4,14) = 4.519, P = 0.015, indicating that the congruency effect was largest over central–frontal sites (FCz and Cz). This pattern of distribution is somewhat unexpected and will require future investigation.

5

There was also a significant interaction between target word type and congruency, F(1,21) = 4.452, P = 0.047, which was qualified by an interaction with prime gender category, F(1,21) = 5.353, P = 0.031. Across congruency, prime gender category had a differential impact for traits, but not for non-traits. For congruent traits, participants were quicker to respond when the prime was Women than Men. For incongruent traits, participants were quicker to respond when the prime was Men than Women. This would imply that female stereotypical traits are more closely tied to their gender category than are male stereotypical traits.

REFERENCES

  1. Amodio DM, Devine PG. Stereotyping and evaluation in implicit race bias: evidence for independent constructs and unique effects on behavior. Journal of Personality and Social Psychology. 2006;91(4):652–61. doi: 10.1037/0022-3514.91.4.652. [DOI] [PubMed] [Google Scholar]
  2. Anderson JE, Holcomb PJ. Auditory and visual semantic priming using different stimulus onset asynchronies: an event-related brain potential study. Psychophysiology. 1995;32(2):177–90. doi: 10.1111/j.1469-8986.1995.tb03310.x. [DOI] [PubMed] [Google Scholar]
  3. Babad E. Wishful thinking among voters: motivational and cognitive influences. International Journal of Public Opinion Research. 1997;9(2):105–25. [Google Scholar]
  4. Banaji MR, Hardin CD. Automatic stereotyping. Psychological Science. 1996;7(3):136–41. [Google Scholar]
  5. Bartholow BD, Fabiana M, Gratton G, Battencourt BA. A psychophysiological examination of cognitive processing of and affective responses to social expectancy violations. Psychological Science. 2001;12(3):197–204. doi: 10.1111/1467-9280.00336. [DOI] [PubMed] [Google Scholar]
  6. Bartholow BD, Pearson MA, Gratton G, Fabiani M. Effects of alcohol on person perception: a social cognitive neuroscience approach. Journal of Personality and Social Psychology. 2003;85(4):627–38. doi: 10.1037/0022-3514.85.4.627. [DOI] [PubMed] [Google Scholar]
  7. Bem SL. The measurement of psychological androgyny. Journal of Consulting and Clinical Psychology. 1974;42(2):155–62. [PubMed] [Google Scholar]
  8. Bentin S, McCarthy G, Wood CC. Event-related potentials, lexical decision and semantic priming. Electroencephalography and Clinical Neurophysiology. 1985;60:343–55. doi: 10.1016/0013-4694(85)90008-2. [DOI] [PubMed] [Google Scholar]
  9. Blair IV. The malleability of automatic stereotypes and prejudice. Personality and Social Psychology Review. 2002;6(3):242–61. [Google Scholar]
  10. Blair IV, Banaji MR. Automatic and controlled processes in stereotype priming. Journal of Personality and Social Psychology. 1996;70(6):1142–63. [Google Scholar]
  11. Boddy J. Event-related potentials in chronometric analysis of primed word recognition with different stimulus onset asynchronies. Psychophysiology. 1986;23(2):232–45. doi: 10.1111/j.1469-8986.1986.tb00624.x. [DOI] [PubMed] [Google Scholar]
  12. Brown C, Hagoort P. The processing nature of the N400: evidence from masked priming. Journal of Cognitive Neuroscience. 1993;5:34–44. doi: 10.1162/jocn.1993.5.1.34. [DOI] [PubMed] [Google Scholar]
  13. Chwilla DJ, Hagoort P, Brown CM. The mechanism underlying backward priming in a lexical decision task: spreading activation versus semantic matching. The Quarterly Journal of Experimental Psychology A: Human Experimental Psychology. 1998;51(3):531–60. [Google Scholar]
  14. Chwilla DJ, Kolk HHJ, Mulder G. Mediated priming in the lexical decision task: evidence from event-related potentials and reaction time. Journal of Memory and Language. 2000;42(3):314–41. [Google Scholar]
  15. Coles MGH. Modern mind-brain reading: psychophysiology, physiology, and cognition. Psychophysiology. 1989;26:251–69. doi: 10.1111/j.1469-8986.1989.tb01916.x. [DOI] [PubMed] [Google Scholar]
  16. Coles MGH, Gratton G, Bashore TR, Eriksen CW, Donchin E. A psychophysiological investigation of the continuous flow model of human information processing. Journal of Experimental Psychology: Human Perception and Performance. 1985;11:529–53. doi: 10.1037//0096-1523.11.5.529. [DOI] [PubMed] [Google Scholar]
  17. Collins AM, Loftus EF. A spreading-activation theory of semantic processing. Psychological Review. 1975;82(6):407–28. [Google Scholar]
  18. Damasio AR. Time-locked multiregional retroactivation: a systems-level proposal for the neural substrates of recall and recognition. Cognition. 1989;33:25–62. doi: 10.1016/0010-0277(89)90005-x. [DOI] [PubMed] [Google Scholar]
  19. Davis M, Whalen PJ. The amygdala: vigilance and emotion. Molecular Psychiatry. 2001;6(1):13–34. doi: 10.1038/sj.mp.4000812. [DOI] [PubMed] [Google Scholar]
  20. Deacon D, Hewitt S, Yang CM, Nagata M. Event-related potentials indices of semantic priming using masked and unmasked words: evidence that the N400 does not reflect a post-lexical process. Cognitive Brain Research. 2000;9:137–46. doi: 10.1016/s0926-6410(99)00050-6. [DOI] [PubMed] [Google Scholar]
  21. Dovidio JF, Evans N, Tyler RB. Racial stereotypes: the contents of their cognitive representations. Journal of Experimental Social Psychology. 1986;22(1):22–37. [Google Scholar]
  22. Fabiani M, Gratton G, Coles MGH. Event-related brain potentials methods, theory, and applications. In: Cacioppo JT, Tassinary LG, Berntson GG, editors. Handbook of Psychophysiology. 2nd. New York: Cambridge University Press; 2007. pp. 85–119. [Google Scholar]
  23. Fazio RH, Jackson JR, Dunton BC, Williams CJ. Variability in automatic activation as an unobstrusive measure of racial attitudes: a bona fide pipeline? Journal of Personality & Social Psychology. 1995;69(6):1013–27. doi: 10.1037//0022-3514.69.6.1013. [DOI] [PubMed] [Google Scholar]
  24. Folstein JR, Van Petten C. Influence of cognitive control and mismatch on the N2 component of the ERP: a review. Psychophysiology. 2008;45(1):152–70. doi: 10.1111/j.1469-8986.2007.00602.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Franklin MS, Dien J, Neely JH, Huber E, Waterson LD. Semantic priming modulates the N400, N300, and N400RP. Clinical Neurophysiology. 2007;118(5):1053–68. doi: 10.1016/j.clinph.2007.01.012. [DOI] [PubMed] [Google Scholar]
  26. Gaertner SL, McLaughlin JP. Racial stereotypes: associations and ascriptions of positive and negative characteristics. Social Psychology Quarterly. 1983;46(1):23–30. [Google Scholar]
  27. Gawronski B, Bodenhausen GV. Accessibility effects on implicit social cognition: the role of knowledge activation and retrieval experiences. Journal of Personality and Social Psychology. 2005;89(5):672–85. doi: 10.1037/0022-3514.89.5.672. [DOI] [PubMed] [Google Scholar]
  28. Hill H, Strube M, Roesch-Ely D, Weisbrod M. Automatic vs. controlled processes in semantic priming—differentiation by event-related potentials. International Journal of Psychophysiology. 2002;44(3):197–218. doi: 10.1016/s0167-8760(01)00202-1. [DOI] [PubMed] [Google Scholar]
  29. Holcomb PJ, Anderson JE. Cross-modal semantic priming: a time-course analysis using event-related brain potentials. Language and Cognitive Processes. 1993;8(4):379–411. [Google Scholar]
  30. Holcomb PJ, Grainger J. Exploring the temporal dynamics of visual word recognition in the masked repetition priming paradigm using event-related potentials. Brain Research. 2007;1180:39–58. doi: 10.1016/j.brainres.2007.06.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ito TA, Cacioppo JT. Electrophysiological evidence of implicit and explicit categorization processes. Journal of Experimental Social Psychology. 2000;36(6):660–76. [Google Scholar]
  32. Johnson R, Jr., Barnhardt J, Zhu J. The contribution of executive processes to deceptive responding. Neuropsychologia. 2004;42(7):878–901. doi: 10.1016/j.neuropsychologia.2003.12.005. [DOI] [PubMed] [Google Scholar]
  33. Judd CM, Blair IV, Chapleau KM. Automatic stereotypes vs. automatic prejudice: sorting out the possibilities in the Payne (2001) weapon paradigm. Journal of Experimental Social Psychology. 2004;40(1):75–81. [Google Scholar]
  34. Kawakami K, Dion KL, Dovidio JF. Racial prejudice and stereotype activation. Personality & Social Psychology Bulletin. 1998;24(4):407–16. [Google Scholar]
  35. Kiefer M. Top-down modulation of unconscious ‘automatic’ processes: a gating framework. Advances in Cognitive Psychology. 2007;3(1):289–306. doi: 10.2478/v10053-008-0032-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kiefer M, Weisbrod M, Kern I, Maier S, Spitzer M. Right hemisphere activation during indirect semantic priming: evidence from event-related potentials. Brain and Language. 1998;64(3):377–408. doi: 10.1006/brln.1998.1979. [DOI] [PubMed] [Google Scholar]
  37. Klauer KC, Teige-Mocigemba S. Controllability and resource dependence in automatic evaluation. Journal of Experimental Social Psychology. 2007;43(4):648–55. [Google Scholar]
  38. Koivisto M, Revonsuo A. Cognitive representations underlying the N400 priming effect. Cognitive Brain Research. 2001;12(3):487–90. doi: 10.1016/s0926-6410(01)00069-6. [DOI] [PubMed] [Google Scholar]
  39. Kreher DA, Holcomb PJ, Kuperberg GR. An electrophysiological investigation of indirect semantic priming. Psychophysiology. 2006;43(6):550–63. doi: 10.1111/j.1469-8986.2006.00460.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Kutas M, Federmeier KD. Electrophysiology reveals semantic memory use in language comprehension. Trends in Cognitive Sciences. 2000;4(12):463–70. doi: 10.1016/s1364-6613(00)01560-6. [DOI] [PubMed] [Google Scholar]
  41. Kutas M, Hillyard SA. Event-related brain potentials to semantically inappropriate and surprisingly large words. Biological Psychology. 1980;11:99–116. doi: 10.1016/0301-0511(80)90046-0. [DOI] [PubMed] [Google Scholar]
  42. Kutas M, Hillyard SA. Brain potentials during reading reflect word expectancy and semantic association. Nature. 1984;307:161–3. doi: 10.1038/307161a0. [DOI] [PubMed] [Google Scholar]
  43. Mackie DM, Smith ER. Intergroup relations: Insights from a theoretically integrative approach. Psychological Review. 1998;105:499–529. doi: 10.1037/0033-295x.105.3.499. [DOI] [PubMed] [Google Scholar]
  44. Mari-Beffa P, Valdes B, Cullen DJD, Catena A, Houghton G. ERP analyses of task effects on semantic processing from words. Cognitive Brain Research. 2005;23(2–3):293–305. doi: 10.1016/j.cogbrainres.2004.10.016. [DOI] [PubMed] [Google Scholar]
  45. Neely JH. Semantic priming and retrieval from lexical memory: roles of inhibitionless spreading activation and limited-capacity attention. Journal of Experimental Psychology: General. 1977;106(3):226–54. [Google Scholar]
  46. Nieuwenhuis S, Aston-Jones G, Cohen JD. Decision making, the P3, and the locus coeruleus-norepinephrine system. Psychological Bulletin. 2005;131(4):510–32. doi: 10.1037/0033-2909.131.4.510. [DOI] [PubMed] [Google Scholar]
  47. Oakhill J, Garnham A, Reynolds D. Immediate activation of stereotypical gender information. Memory and Cognition. 2005;33:972–83. doi: 10.3758/bf03193206. [DOI] [PubMed] [Google Scholar]
  48. Osterhout L, Bersick M, McLaughlin J. Brain potentials reflect violations of gender stereotypes. Memory & Cognition. 1997;25:273–85. doi: 10.3758/bf03211283. [DOI] [PubMed] [Google Scholar]
  49. Rhodes SM, Donaldson DI. Association and not semantic relationships elicit the N400 effect: electrophysiological evidence from an explicit language comprehension task. Psychophysiology. 2008;45(1):50–9. doi: 10.1111/j.1469-8986.2007.00598.x. [DOI] [PubMed] [Google Scholar]
  50. Roehm D, Bornkessel-Schlesewsky I, Rosler F, Schlesewsky M. To predict or not to predict: influences of task and strategy on the processing of semantic relations. Journal of Cognitive Neuroscience. 2007;19(8):1259–74. doi: 10.1162/jocn.2007.19.8.1259. [DOI] [PubMed] [Google Scholar]
  51. Rugg MD. The effects of semantic priming and word repetition on event-related potentials. Psychophysiology. 1985;22:642–7. doi: 10.1111/j.1469-8986.1985.tb01661.x. [DOI] [PubMed] [Google Scholar]
  52. Smith ER, Semin GNR. Situated social cognition. Current Directions in Psychological Science. 2007;16(3):132–5. doi: 10.1111/j.1467-8721.2007.00518.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Taylor JH, Crites SL, Corral G, White KR. The N400 effect: Semantic vs. evaluative incongruity; Poster session presented at the annual meeting of the Society for Psychophysiological Research; Austin, TX. 2008. Oct, [Google Scholar]
  54. Wittenbrink B, Judd CM, Park B. Evidence for racial prejudice at the implicit level and its relationship with questionnaire measures. Journal of Personality and Social Psychology. 1997;72(2):262–74. doi: 10.1037//0022-3514.72.2.262. [DOI] [PubMed] [Google Scholar]
  55. Wittenbrink B, Judd CM, Park B. Spontaneous prejudice in context: variability in automatically activated attitudes. Journal of Personality and Social Psychology. 2001;81(5):815–27. doi: 10.1037//0022-3514.81.5.815. [DOI] [PubMed] [Google Scholar]

Articles from Social Cognitive and Affective Neuroscience are provided here courtesy of Oxford University Press

RESOURCES