Abstract

Recent theories of semantic memory suggest a subdivision into several separate domains of knowledge. The present study examined the structure of semantic person knowledge by analyzing both behavioral and ERP correlates of associative priming (via co-occurrence and/or shared semantic information) versus purely categorical priming (via shared occupational information). Participants performed familiarity decisions for target faces, which were preceded by sandwich-masked prime names at either short (33 msec) or long (1033 msec) SOAs. Although masking effectively prevented explicit prime recognition, faster RTs were generally observed for both associative and categorical priming (relative to an unrelated prime-target condition). At the short SOA, both associatively and categorically primed targets similarly elicited more positive going ERPs compared with unrelated targets in the N400 time range (N400 priming effect), suggesting a common initial mechanism mediating both forms of priming. By contrast, at the long SOA, a typical N400 priming effect was observed for associative priming only, whereas the corresponding effect for categorical priming was small and restricted to a left parietal site. This hitherto unreported interaction of relatedness, and SOA in the N400 suggests an initial fast spreading of activation to a wide range of related targets, which subsequently focuses to more closely related people at longer SOAs. This ability of ERPs to trace the neural dynamics of activation for different forms of prime/target relatedness can be exploited for testing different models of semantic priming.

INTRODUCTION

Human semantic memory stores a vast amount of information, not only about facts and world knowledge, word meaning, and objects but also about people. Although early theories on semantic memory assumed a unified system to store all these types of information, more recent findings suggest a division into several separate domains of knowledge (see, e.g., Kiefer, 2005; Caramazza & Shelton, 1998). In line with these approaches, evidence from neuropsychological (Thompson et al., 2004), ERP (Engst, Martin-Loeches, & Sommer, 2006), and functional brain imaging research (Mitchell, Heatherton, & Macrae, 2002) suggests distinct semantic systems for representing knowledge about people as compared with objects. However, within the domain of person-specific knowledge, it remains controversial how exactly semantic memory is structured.

According to one view of semantic person memory, although knowledge (e.g., about occupation) is stored within the representation of an individual person, it is not used to link representations of different people (Barry, Johnston, & Scanlan, 1998). Rather, direct links are established by episodic factors such as co-occurrence. Another view of semantic person memory stresses the role of semantic features, which may be shared by many people (Carson & Burton, 2001; Burton, Bruce, & Johnston, 1990). On that view, just like in other domains of semantic memory (McRae & Boisvert, 1998; McRae, deSa, & Seidenberg, 1997), the overarching organizing principle of person-related semantic memory is feature similarity. Studies investigating patients with local brain lesions (Tranel, Damasio, & Damasio, 1997; Ellis, Young, & Critchley, 1989; Hanley, Young, & Pearson, 1989) and studies using functional brain imaging (Gobbini & Haxby, 2007; Rotshtein, Henson, Treves, Driver, & Dolan, 2005; Gorno-Tempini & Price, 2001; Sergent, Ohta, & Macdonald, 1992) have implicated a role of anterior temporal lobe regions or fronto-temporal circuits (Elfgren et al., 2006; Leveroni et al., 2000; Gorno-Tempini et al., 1998) for semantic person memory. In addition, the medial-temporal lobe may retrieve person-related information (Elfgren et al., 2006; Douville et al., 2005; Haist, Gore, & Mao, 2001), perhaps particularly so under conditions of episodic processing (Denkova, Botzung, & Manning, 2006). Thus, although a picture of the brain structures involved in accessing person-related memory is therefore beginning to emerge, the organizing principles of semantic person memory remain controversial.

To date, the most influential account describing the organization of person-specific knowledge is the Interactive Activation and Competition (IAC) model (Burton et al., 1990; see Figure 1), which suggests a categorical structure of person-related information. According to this view, any known person is represented by an individual so-called person identity node (PIN), which becomes activated as a consequence of spreading activation from perceptual representations (e.g., so-called face or name recognition units) for this particular person. Perceptual entries to PINs are thought to be multimodal, such that for instance a person's face, a written or heard name, or a voice can trigger activation. If PIN activation exceeds a certain threshold, the person is classified as being familiar. Moreover, PINs act as gateways to semantic information (e.g., a person's occupation), which is stored in so-called semantic information units (SIUs). Importantly, SIUs are not specific to any single person but are connected to the PINs of other people sharing this particular semantic information. For instance, all “actors” will be connected via a common SIU.

Figure 1. 

Illustration of the IAC model (Burton et al., 1990). NRUs = name recognition units; FRUs = face recognition units; PINs = person identity nodes; SIUs = semantic information units.

Figure 1. 

Illustration of the IAC model (Burton et al., 1990). NRUs = name recognition units; FRUs = face recognition units; PINs = person identity nodes; SIUs = semantic information units.

An important experimental approach to examine both semantic person knowledge and semantic relations in other domains such as the processing of word meaning has been to study priming. In word recognition experiments, semantic priming refers to the finding that a stimulus is recognized faster as being a word (and not a nonword) in a lexical decision task when preceded by a related compared with an unrelated word (for a recent review, see McNamara, 2005). For instance, participants are faster to classify “butter” as a word when it is preceded by “bread” compared with “doctor.” In person recognition, semantic priming refers to the similar finding that a given person is classified faster to be familiar (and not unfamiliar) when preceded by a related person (see, e.g., Bruce, 1983). For instance, when presented with Angelina Jolie's face, participants are usually faster to decide that the person is familiar when preceded by the face (or written name) of Brad Pitt compared with, say, John Lennon. Importantly, the IAC model predicts that not only Brad Pitt but also any actor (e.g., Hugh Grant) would prime the recognition of Angelina Jolie. According to the IAC model, the only necessary prerequisite for priming is that both persons are connected via shared SIUs. Thus, priming is assumed to be mediated entirely via semantic (in this case occupational category) information (see Figure 1). By contrast, others suggested that only persons who usually co-occur (i.e., highly associated persons such as Brad Pitt, but not Hugh Grant) would effectively prime a familiarity decision (Barry et al., 1998). Whereas Barry et al. (1998) assume only associative priming to exist, the IAC model not only predicts purely categorical priming but also makes explicit that “associative” priming is exclusively mediated by shared semantic information (Carson & Burton, 2001).

Although earlier findings were interpreted to support the suggestion of priming for associated people only (Barry et al., 1998; Young, Flude, Hellawell, & Ellis, 1994), there is growing recent evidence for purely categorical priming (Stone, 2008; Stone & Valentine, 2007; Carson & Burton, 2001). However, associative priming has usually been found to be stronger, a phenomenon known as the “associative boost” (Lucas, 2000). Carson and Burton (2001) explained this phenomenon by suggesting that highly associated persons are linked not only via one SIU (e.g., occupation) but typically share more semantic information (e.g., about nationality, place of living etc.) compared with purely categorically related pairs. Although this is a plausible explanation for the associative boost, an alternative is that qualitatively different processes underlie associative and categorical priming. As detailed below, comparing ERP effects of associative and categorical priming is a promising technique to test between these competing (though not necessarily mutually exclusive) explanations.

The most prominent ERP correlate of semantic processing is the N400 (Kutas & Hillyard, 1980), a negative deflection approximately 400 msec after stimulus onset. In semantic word priming experiments, unrelated target items have been found to elicit more negative going ERP waveforms compared with related targets over central and parietal regions (Bentin, Mccarthy, & Wood, 1985). Although the negative deflection is often superimposed on a large positive component and is therefore not necessarily negative in absolute terms, this phenomenon has been termed the N400 priming effect.

In the domain of person recognition, several studies demonstrated N400 priming effects for highly associated target stimuli (Schweinberger, Pfutze, & Sommer, 1995), independent of whether the prime and target stimuli depicted faces or written names (Schweinberger, 1996). More recently, we demonstrated ERP correlates of priming not only for highly associated but also for purely categorically related prime/target pairs (Wiese & Schweinberger, 2008). Importantly, the N400 effect was clear for associative priming but much reduced for categorical priming. However, categorical priming caused a somewhat earlier and more posterior effect, which started in the time range of the P2 component. From these findings, we concluded that associative and categorical priming in person recognition are mediated by at least partly different mechanisms.

However, semantic priming may be mediated not only by passive activation processes but also by strategic processes such as expectancy and retrospective semantic matching (for a recent review, see McNamara, 2005). Thus, although different processes may underlie categorical and associative priming in person recognition (Wiese & Schweinberger, 2008), it is possible that the results of that study were affected by strategic processes. This may be critical because (i) it appears questionable whether the structure of semantic memory can be inferred from experiments in which participants are able to arrange semantic information according to task demands, and (ii) brain processes mediating spreading activation and strategic processing are likely to differ.

Taking the above considerations into account, it is of importance whether nonstrategic ERP priming effects can be demonstrated in the N400 or in other ERP components. Because visual masking largely reduces conscious identification of the prime stimulus, this technique dramatically constrains strategic processing of that prime. A first study examining ERP correlates of semantic word priming with both masked and unmasked primes found N400 effects only in the unmasked condition (Brown & Hagoort, 1993), arguing in favor of a strategic process to drive the effect. Subsequent studies, however, demonstrated N400 priming effects regardless of whether primes were masked or unmasked (Kiefer, 2002; Deacon, Hewitt, Yang, & Nagata, 2000). Importantly, the N400 masked semantic priming effect depends on SOA (Kiefer & Spitzer, 2000): Although clear effects emerged for both masked and unmasked primes at an SOA of 67 msec, a similar effect was only found for unmasked primes in a longer SOA condition (200 msec). Subsequent studies additionally underlined the important role of SOA. On the one hand, studies with long prime/target SOAs found no conclusive evidence for masked N400 priming effects (Holcomb, Reder, Misra, & Grainger, 2005; Ruz, Madrid, Lupianez, & Tudela, 2003). On the other hand, Grossi (2006) demonstrated N400 effects for masked primes at a short prime-target SOA (50 msec), which was independent of the relatedness proportion (see also Kiefer & Brendel, 2006).

All of the above-cited studies examined word priming. To the best of our knowledge, no prior study examined ERP correlates of masked semantic priming in person recognition. However, a recent study on masked priming demonstrated an N400-like effect for repeated compared with unrepeated famous but not unfamiliar faces (Henson, Mouchlianitis, Matthews, & Kouider, 2008). Although relevant for present purposes, this study investigated repetition priming rather than semantic priming. Thus, as pointed out by the authors, facilitated access to perceptual rather than semantic representations could underlie this effect.

In sum, although semantic N400 priming effects from masked word priming have been studied extensively (for a recent review, see Kouider & Dehaene, 2007), no previous study examined ERP correlates of masked semantic priming in person recognition. Therefore, it is unclear whether ERP correlates of categorical and associative priming in person recognition relate to strategic or nonstrategic processes.

The Present Study

As detailed above, our previous results (Wiese & Schweinberger, 2008) pointed to distinct neural mechanisms mediating categorical and associative priming. Although this may pose a challenge for the IAC model (Burton et al., 1990), it may be premature to draw strong conclusions about the structure of semantic person memory from these data alone. In particular, prime stimuli were clearly visible which, together with the long SOA used in that study, may have encouraged participants to make strategic use of the prime with respect to the forthcoming target.

In the present experiment, we therefore used masked prime presentation throughout to severely reduce conscious recognition of the primes and thus strategic processing. In addition, we varied prime/target SOA to compare the time course of nonstrategic categorical versus associative priming. The experiment thus consisted of a two-factorial design with the factors (i) prime-target relatedness or prime type and (ii) SOA (33 vs. 1033 msec). We were particularly interested in whether our previously observed results of a widespread N400 effect for associative but not categorical priming and of a P2 effect for categorical but not associative priming (see Wiese & Schweinberger, 2008) would occur in the absence of strategic processing. Such N400 and P2 effects under conditions of nonstrategic processing would strengthen the position that categorical and associative priming in person recognition are mediated by qualitatively different processes. However, as N400 effects of masked priming may be highly sensitive to SOA (Kiefer & Spitzer, 2000), we also considered the possibility that the above distinctive mechanisms of categorical versus associative priming take time to develop. On that perspective, a common mechanism for categorical and associative priming (Carson & Burton, 2001) might still characterize initial activation of related representations, and the rejection of such a mechanism may be premature.

METHODS

Participants

Thirty-six undergraduate students from the University of Jena (28 women) with a mean age of 21.8 years (SD = 2.4 years) participated in the experiment, which were divided into two groups of equal size (between-subjects factor SOA, see below). These groups did not differ in age or sex distribution; age, t(34) = 0.4, p > .05; sex, χ2 = 0.6, p > .05. Participants either received course credits or were paid €5 per hour. All were right-handed according to a modified version of the Edinburgh Handedness Inventory (Oldfield, 1971) and reported normal or corrected-to-normal vision. The study was approved by the local ethics committee, and all participants gave written informed consent.

Stimuli

Stimuli were identical to those in Wiese and Schweinberger (2008; see that article for a complete list). Primes consisted of 120 famous names presented in Arial 18 at approximately 90 cm viewing difference. Surnames were centered below their corresponding first names, resulting in a vertical viewing angle of 1.25° and a maximum horizontal angle of 6.5°. All names were presented in capital letters. Primes were preceded and followed by two rows of capital “X” characters (total viewing angle 1.25° × 6.5°, thus covering all letters of the primes). Targets were 120 familiar and 120 unfamiliar faces. All pictures were edited using Adobe Photoshop to remove all information apart from the face (e.g., clothing, background). Images were converted to gray scale with black background and framed within an area of 170 × 216 pixels (6.0 × 7.6 cm), corresponding to 3.8° × 4.8° viewing angle. We chose a cross-domain design to exclude potential effects at the level of perceptual representations (Schweinberger, 1996).

Famous target faces were preceded by (i) highly associated primes (“associated”), (ii) categorically related primes that belonged to the same occupational categories (actors, musicians, politicians, sports or TV celebrities) but were not associated to their respective target (“same category”), or (iii) nonassociated primes that belonged to a different occupational category (“different category”). Pairings were identical to those used in a previous study, and selection was based on a pilot study described in detail elsewhere (Wiese & Schweinberger, 2008). In short, in a free association task, the targets selected for the associated condition were produced with high probability to their respective primes (by a mean of 16/20 raters). By contrast, targets in the same and different category conditions were not only nonassociated to their respective primes (0/20 raters) but also showed only weak associations to any specific celebrity (with a mean maximum association of 6/20 raters for both same and different category pairs). Stimulus pairs for the three conditions were matched for mean familiarity as assessed in the pilot study.

Please note that the procedure chosen in this experiment did not allow to balance the target stimuli across the different conditions (associated, same category, different category) for two reasons: (i) the type of relatedness (purely categorical vs. associative) between prime and target was the main factor of interest in this experiment. As noted above, mean maximum association was kept low for prime/target pairs in both same category and different category conditions. This was because presenting prime stimuli that are highly associated to one specific person before a different target person that is categorically related may lead to interferences between the processes mediating categorical and associative relatedness (for a similar argument, see Stone, 2008). (ii) To fully balance items over conditions, it would have been necessary to repeat target stimuli. This, however, may lead to a situation in which the response to the target is influenced by explicit memories of a previous encounter, which may weaken processes of semantic memory (see McNamara, 2005).

To create the task demand, an equal number of unfamiliar targets were presented. To ensure that primes were nonpredictive for target familiarity, the same primes were used for famous and unfamiliar targets. Note that potential effects of prime repetition were equivalent in all experimental conditions.

Procedure

All participants were seated in a dimly lit, electrically shielded, and sound-attenuated chamber (400-A-CT-Special; Industrial Acoustics, Niederkrüchten, Germany) with their heads in a chin rest. Each experimental session began with a series of practice trials on different stimuli, which were excluded from data analysis.

In the main experiment, all primes were masked. At 33 msec SOA, trials started with a forward mask (500 msec), followed by the prime (16.67 msec) and a backward mask (16.67 msec), which was directly followed by the target (1200 msec). At 1033 msec SOA, an additional blank screen was presented for 1000 msec between backward mask and target. All trials ended with a blank screen (2000 msec).

Stimuli were presented in four blocks with 60 trials each. Each block consisted of 10 associated, 10 same category, and 10 different category pairs as well as 30 trials with unfamiliar targets. The task was to decide as fast and correctly as possible whether the target represented a famous face. Participants were told that letters would precede the presentation of the target. They were instructed to closely attend to the monitor during the whole trial. Neither the existence of the masked primes nor the potential relatedness of prime and target stimuli was explained to participants. Responses were made via left and right index finger button presses and were scored as correct if the respective key was pressed between 200 and 1200 msec after target onset. Trials with unfamiliar faces were not further analyzed.

Following the main experiment, a test of the masking procedure was carried out. For that purpose, 20 famous and 20 unfamiliar names were presented randomly intermixed in a forced-choice recognition task. All stimuli were masked, and presentation times for forward mask, name stimulus, and backward mask were identical to those in the main part of the experiment. Here, participants were informed about the presentation of the masked name and were asked to decide whether that name was famous or not. Participants were explicitly encouraged to guess when no conscious information about the name was available.

Please note that in contrast to the main part of the experiment, no face stimuli were presented after the masked names. It has been reported that related targets improved accuracy of lexical decisions on very briefly presented primes (e.g., Dark, 1988). However, free associations to such targets are likely to influence the reports of the prime, which in turn may substantially distort the measurement of masked stimulus recognition (Carr & Dagenbach, 1990). Thus, we decided not to present targets during the test of the masking procedure. From the resulting responses, sensitivity measures (d′) were calculated.

ERP Recording and Analysis

Thrity-two-channel EEG was recorded in the main part of the experiment with a BioSemi Active II system (BioSemi, Amsterdam, Netherlands). The active sintered Ag/Ag-Cl electrodes were mounted in an elastic cap. Recording sites corresponded to Fz, Cz, Pz, Iz, FP1, FP2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T7, T8, P7, P8, F9, F10, FT9, FT10, TP9, TP10, P9, P10, PO9, PO10, I1, and I2. EEG was recorded continuously with a 512-Hz sample rate from DC to 120 Hz. Note that BioSemi systems work with a “zero-Ref” setup with ground and reference electrodes replaced by a so-called CMS/DRL circuit (for further information, see www.biosemi.com/faq/cms&drl.htm).

Contributions of blink artifacts were corrected using the algorithm implemented in BESA 5.1 (Berg & Scherg, 1994). In the 33-msec SOA condition, EEG was segmented from −200 relative to forward mask onset to 1000 msec relative to target stimulus onset, with the first 200 msec as baseline. In the 1033-msec SOA condition, a time segment from −200 to 1000 msec relative to target onset was chosen, again with the first 200 msec as baseline. Only trials with correct responses entered the analysis. Trials with incorrect responses, nonocular artifacts, and saccades were rejected from further analysis. Artifact rejection was carried out using the BESA 5.1 tool, with an amplitude threshold of 100 μV as well as a gradient criterion rejecting all trials with more than a 75-μV difference between two consecutive data points. Remaining trials were recalculated to average reference, digitally low-pass filtered at 40 Hz (12 db/oct, zero phase shift), and averaged according to experimental condition. This procedure resulted in a mean of 32.4 trials (SD = 5.2 ) in the associated condition, of 30.0 trials (SD = 5.0) in the same category, and of 28.8 trials (SD = 5.8) in the different category condition. The minimum number of trials was 20 in each condition.

In the resulting waveforms, mean amplitudes for N170 and P2 were calculated at electrodes P9 and P10. N400 was measured at left, midline, and right frontal, central, and parietal positions (F3, Fz, F4, C3, Cz, C4, P3, Pz, P4). Statistical analysis involved ANOVAs, with degrees of freedom corrected according to Huynh–Feldt where appropriate.

RESULTS

Performance

An ANOVA on correct RTs (see Table 1) with the within-subject factor “prime type” (associated, same category, different category) and the between-subjects factor SOA (33 msec, 1033 msec) revealed significant main effects for both Prime Type, F(2, 68) = 26.3, p < .001, and SOA, F(1, 34) = 10.6, p < .01. The interaction was not significant (F < 1). Contrast analysis on the factor prime type revealed significantly faster RTs both in the associated compared with the different category condition, F(1, 34) = 50.2, p < .001, as well as in the same category condition compared with the different category condition, F(1, 34) = 21.6, p < .001. Associated and same category conditions did not differ significantly, F(1, 34) = 1.9, p > .05.

Table 1. 

Mean ± SEM Behavioral and Event-related Potential Measures for the Three Types of Prime/Target Relatedness in the 33- and 1033-msec SOA Conditions


33-msec SOA
1033-msec SOA
Associated
Same Category
Different Category
Associated
Same Category
Different Category
RT (msec) 667.3 ± 14.9 671.6 ± 15.1 702.5 ± 17.5 598.2 ± 14.1 604.1 ± 16.4 625.1 ± 16.7 
ACC .91 ± 0.1 .84 ± 0.2 80 ± 0.3 .88 ± 0.3 .82 ± 0.3 .78 ± 0.4 
 
N170 (μV) 
P9 −2.4 ± 0.6 −2.8 ± 0.6 −2.2 ± 0.8 −1.8 ± 0.8 −2.1 ± 0.9 −2.1 ± 0.8 
P10 −4.9 ± 0.9 −5.1 ± 0.9 −5.0 ± 1.0 −2.6 ± 1.0 −3.0 ± 1.0 −3.0 ± 1.0 
 
P2 (μV) 
P9 −1.0 ± 0.8 −1.2 ± 0.8 −0.5 ± 0.8 0.4 ± 0.8 1.1 ± 0.8 0.7 ± 0.8 
P10 −3.4 ± 0.8 −3.6 ± 0.9 −2.9 ± 0.8 0.4 ± 0.8 1.0 ± 0.9 0.7 ± 0.8 
 
N400 (μV) 
F3 −0.5 ± 0.8 −0.8 ± 0.8 −1.3 ± 0.8 −2.0 ± 0.8 −3.1 ± 0.8 −2.7 ± 0.8 
Fz −0.6 ± 0.9 −0.4 ± 0.9 −1.1 ± 0.8 −2.2 ± 0.9 −2.9 ± 0.9 −2.8 ± 0.8 
F4 −0.3 ± 0.7 −0.7 ± 0.7 −0.6 ± 0.7 −2.2 ± 0.7 −2.9 ± 0.7 −2.4 ± 0.7 
C3 0 ± 0.8 0.4 ± 0.8 −0.8 ± 0.8 1.0 ± 0.8 −0.1 ± 0.8 −0.4 ± 0.8 
Cz −0.7 ± 1.1 0 ± 0.9 −1.6 ± 1.1 1.0 ± 1.1 −0.3 ± 0.9 −0.3 ± 1.1 
C4 0.3 ± 0.7 0.4 ± 0.6 −0.4 ± 0.7 0.5 ± 0.7 −0.2 ± 0.6 −0.6 ± 0.7 
P3 4.4 ± 0.6 3.9 ± 0.6 3.7 ± 0.6 6.2 ± 0.6 6.0 ± 0.6 5.2 ± 0.7 
Pz 5.5 ± 0.8 6.1 ± 0.8 4.3 ± 0.7 7.2 ± 0.8 6.2 ± 0.8 6.0 ± 0.7 
P4 4.9 ± 0.8 5.6 ± 0.8 4.2 ± 0.7 7.2 ± 0.8 6.3 ± 0.8 5.8 ± 0.7 

33-msec SOA
1033-msec SOA
Associated
Same Category
Different Category
Associated
Same Category
Different Category
RT (msec) 667.3 ± 14.9 671.6 ± 15.1 702.5 ± 17.5 598.2 ± 14.1 604.1 ± 16.4 625.1 ± 16.7 
ACC .91 ± 0.1 .84 ± 0.2 80 ± 0.3 .88 ± 0.3 .82 ± 0.3 .78 ± 0.4 
 
N170 (μV) 
P9 −2.4 ± 0.6 −2.8 ± 0.6 −2.2 ± 0.8 −1.8 ± 0.8 −2.1 ± 0.9 −2.1 ± 0.8 
P10 −4.9 ± 0.9 −5.1 ± 0.9 −5.0 ± 1.0 −2.6 ± 1.0 −3.0 ± 1.0 −3.0 ± 1.0 
 
P2 (μV) 
P9 −1.0 ± 0.8 −1.2 ± 0.8 −0.5 ± 0.8 0.4 ± 0.8 1.1 ± 0.8 0.7 ± 0.8 
P10 −3.4 ± 0.8 −3.6 ± 0.9 −2.9 ± 0.8 0.4 ± 0.8 1.0 ± 0.9 0.7 ± 0.8 
 
N400 (μV) 
F3 −0.5 ± 0.8 −0.8 ± 0.8 −1.3 ± 0.8 −2.0 ± 0.8 −3.1 ± 0.8 −2.7 ± 0.8 
Fz −0.6 ± 0.9 −0.4 ± 0.9 −1.1 ± 0.8 −2.2 ± 0.9 −2.9 ± 0.9 −2.8 ± 0.8 
F4 −0.3 ± 0.7 −0.7 ± 0.7 −0.6 ± 0.7 −2.2 ± 0.7 −2.9 ± 0.7 −2.4 ± 0.7 
C3 0 ± 0.8 0.4 ± 0.8 −0.8 ± 0.8 1.0 ± 0.8 −0.1 ± 0.8 −0.4 ± 0.8 
Cz −0.7 ± 1.1 0 ± 0.9 −1.6 ± 1.1 1.0 ± 1.1 −0.3 ± 0.9 −0.3 ± 1.1 
C4 0.3 ± 0.7 0.4 ± 0.6 −0.4 ± 0.7 0.5 ± 0.7 −0.2 ± 0.6 −0.6 ± 0.7 
P3 4.4 ± 0.6 3.9 ± 0.6 3.7 ± 0.6 6.2 ± 0.6 6.0 ± 0.6 5.2 ± 0.7 
Pz 5.5 ± 0.8 6.1 ± 0.8 4.3 ± 0.7 7.2 ± 0.8 6.2 ± 0.8 6.0 ± 0.7 
P4 4.9 ± 0.8 5.6 ± 0.8 4.2 ± 0.7 7.2 ± 0.8 6.3 ± 0.8 5.8 ± 0.7 

ACC = accuracy.

An analogous ANOVA on accuracies (see Table 1) revealed a significant main effect of Prime Type, F(2, 68) = 24.9, p < .001. Neither the SOA factor nor the interaction were significant (both F < 1). Accuracy on associated targets was significantly enhanced compared with both the same category, F(1, 34) = 31.9, p < .001, and the different category condition, F(1, 34) = 32.0, p < .001, and again for the same compared with the different category condition, F(1, 34) = 7.5, p < .05.

The test for explicit perception of the masked names yielded a mean d′ of 0.04 (SD = 0.32) in the 33-msec SOA group and of 0.01 (SD = 0.57) in the 1033-msec SOA group. Neither value differed significantly from zero; short SOA, t(17) = 0.6, p > .05; long SOA, t(17) = 0.1, p > .05. In addition, individual d′ did not correlate significantly with RT measures of associative or categorical priming as calculated by subtracting the associated and same category condition from the different category condition, respectively, neither in the 33-msec SOA group (associative priming, r = −.26; categorical priming, r = −.18; both p > .05) nor in the 1033-msec SOA group (associative priming, r = .17; categorical priming, r = .05; both p > .05).

Event-related Potentials

N170

N170 peaked approximately 170 msec after target onset in the 33-msec SOA group and 160 msec in the 1033-msec SOA group at electrode P10. Thus, mean amplitudes were calculated from 150 to 190 msec and from 140 to 180 msec, respectively (see Table 1). A repeated measures ANOVA on electrodes P9 and P10 with the within-subject factors hemisphere and prime type as well as the between-subjects factor SOA revealed a main effect of hemisphere, F(1, 34) = 13.4, p < .001, reflecting a large N170 over the right hemisphere. In addition, a trend toward a significant interaction of Hemisphere × SOA, F(1, 34) = 3.0, p = .09, was observed. No other main effect or interaction was significant (all p > .1).

P2

Mean amplitudes from 210 to 290 msec were taken in both SOA groups (see Table 1, Figure 2). The ANOVA on electrodes P9 and P10 yielded significant main effects of Hemisphere, F(1, 34) = 6.2, p < .05, and SOA, F(1, 34) = 8.3, p < .01, as well as significant interactions of Hemisphere × SOA, F(1, 34) = 5.5, p < .05, and Prime Type × SOA, F(2, 68) = 3.3, p < .05. To follow-up these interactions, separate ANOVAs for both SOA conditions were calculated. In the 33-msec SOA group, a significant main effect of Hemisphere, F(1, 17) = 12.9, p < .001, reflected more negative amplitudes over the right hemisphere. No significant effect of prime type was detected (p > .1). In the 1033-msec SOA group, a trend for a significant main effect of prime type was evident, F(2, 34) = 2.8, p = .07, reflecting more positive amplitudes for the same category compared with both different category and associated conditions. No effect of hemisphere was observed (F < 1).

Figure 2. 

Grand mean ERPs of electrodes P9/P10. Note the inversed amplitudes of same and different category conditions in the P2 (dashed lines) in the 33- and 1033-msec SOA conditions. Also note that the 200 msec before target onset do not correspond to the baseline at 33 msec SOA (see Methods section).

Figure 2. 

Grand mean ERPs of electrodes P9/P10. Note the inversed amplitudes of same and different category conditions in the P2 (dashed lines) in the 33- and 1033-msec SOA conditions. Also note that the 200 msec before target onset do not correspond to the baseline at 33 msec SOA (see Methods section).

N400

N400 effects were analyzed by calculating mean amplitudes in a time window from 300 to 600 msec at electrodes F3, Fz, F4, C3, Cz, C4, P3, Pz, and P4 (see Table 1, Figures 3 and 4). An ANOVA with repeated measures on anterior/posterior electrode position (frontal, central, parietal), laterality (left, midline, right), prime type (associated, same category, different category), and between-subjects factor SOA was performed. This revealed significant main effects of Anterior/Posterior, F(2, 68) = 116.4, p < .001; ɛ = .70, and Prime Type, F(2, 68) = 9.0, p < .001, as well as significant interactions of Anterior/Posterior × SOA, F(2, 68) = 6.2, p < .001, Prime Type × SOA, F(2, 68) = 116.4, p < .001, Anterior/Posterior × Laterality, F(4, 136) = 3.3, p < .05; ɛ = .90, and Anterior/Posterior × Prime Type, F(4, 136) = 2.6, p < .05; ɛ = .84, the latter effect indicating larger N400 priming effects over more posterior electrodes.

Figure 3. 

Grand mean ERPs from the 33-msec SOA condition. The first and the second dashed lines mark the onset of the forward mask and the target, respectively. Note the increased amplitudes in the N400 time range (continuous lines) for both associated and same category targets.

Figure 3. 

Grand mean ERPs from the 33-msec SOA condition. The first and the second dashed lines mark the onset of the forward mask and the target, respectively. Note the increased amplitudes in the N400 time range (continuous lines) for both associated and same category targets.

Figure 4. 

Grand mean ERPs from the 1033-msec SOA condition. The dashed line marks the onset of the target. Note the increased amplitudes in the N400 time range (continuous lines) for associated targets only at Pz and for both associated and same category targets at P3.

Figure 4. 

Grand mean ERPs from the 1033-msec SOA condition. The dashed line marks the onset of the target. Note the increased amplitudes in the N400 time range (continuous lines) for associated targets only at Pz and for both associated and same category targets at P3.

To follow-up the Prime Type × SOA interaction, which was of primary interest for the present study, separate ANOVAs were calculated for the 33- and the 1033-msec SOA groups, respectively. In the 33-msec SOA group, a significant effect of prime type was detected, F(2, 34) = 5.0, p < .01, reflecting significant differences between both the associated and the different category condition, F(1, 17) = 5.4, p < .05, as well as between the same category and the different category condition, F(1, 17) = 7.2, p < .05. In the 1033-msec SOA group, again a significant effect of prime type was observed, F(2, 34) = 7.4, p < .01. In this group, a significant difference between the associated and the different category condition was found, F(1, 17) = 11.8, p < .01, whereas the same category condition did not differ significantly from the different category condition (F < 1, see also Figure 5).

Figure 5. 

Scalp topographical voltage maps (spherical spline interpolation, 90° equidistant projection) from 300 to 600 msec for associative priming (associated–different category) and categorical priming (same category–different category) at both short (33 msec) and long (1033 msec) SOA. Note the similar priming effects in the 33-msec SOA condition (upper part) and the widespread associative but largely reduced categorical priming effect in the long SOA condition (lower part).

Figure 5. 

Scalp topographical voltage maps (spherical spline interpolation, 90° equidistant projection) from 300 to 600 msec for associative priming (associated–different category) and categorical priming (same category–different category) at both short (33 msec) and long (1033 msec) SOA. Note the similar priming effects in the 33-msec SOA condition (upper part) and the widespread associative but largely reduced categorical priming effect in the long SOA condition (lower part).

DISCUSSION

The present study exemplifies how the temporal dynamics of nonstrategic categorical and associative priming in person recognition can be traced using event-related brain potentials. Although behavioral results yielded significant and similar associative and categorical RT priming in both short and long SOA conditions, ERP priming effects differed considerably as a function of SOA. Specifically, whereas categorical and associative priming both elicited an N400 priming effect of similar amplitude and topography at the short SOA, a different pattern emerged at the long SOA: Here, a large and widespread N400 effect was found for associative priming, whereas an N400 effect was small and topographically restricted to a left parietal position for categorical priming. In addition, categorical but not associative priming elicited a trend for an earlier and more posterior P2 effect. These findings are discussed in detail below.

Our finding of both associative and categorical priming with masked prime names and target faces is in line with the IAC account, which explicitly describes the spreading of activation from PIN to SIU levels to occur in the absence of conscious prime identification (Burton, Young, Bruce, Johnston, & Ellis, 1991). However, associative priming is often substantially stronger than categorical priming (see e.g., Stone, 2008), a phenomenon known as the “associative boost” (Lucas, 2000). In person recognition, this finding has been explained by a larger amount of overlapping SIUs for associated compared with purely categorically related pairs (Carson & Burton, 2001). Although the present study found no significant difference in RT priming for purely categorical and associative priming, in our previous study significantly stronger associative priming was observed with the identical set of stimuli (Wiese & Schweinberger, 2008). These differences could suggest that the associative boost is largely due to expectancy and therefore strategic processes. When expectancy-based processes are unlikely to occur in both short and long SOA conditions because of masked primes, as in the present study, no significant associative boost is observed.

As a qualification, a small and nonsignificant RT advantage for associated compared with same category targets was present in both SOA conditions. It thus remains possible that a small associative boost is present even when strategic influences can be largely excluded. Such an interpretation would be again well in line with the IAC account (Carson & Burton, 2001).

The present study observed topographically similar N400 priming effects for categorical and associative priming at a short SOA. By contrast, in the long SOA condition, a widespread N400 effect was detected for associative priming, whereas no significant effect for categorical priming was observed. As can be seen from Figure 5, however, same category compared with different category targets elicited somewhat more positive going waveforms in the N400 time window at parietal sites, particularly over the left hemisphere. It may be noted that this finding is strikingly similar to our previous study, which demonstrated only local and small left parietal N400 effects for categorically related face targets (Wiese & Schweinberger, 2008), whereas widespread N400 effects for associative priming have repeatedly been reported (Schweinberger, 1996; Schweinberger et al., 1995). Overall, a consistent pattern emerges across these studies, suggesting that masking of the primes had little effect per se. Thus, N400 effects in the previous and the present study are unlikely to be much influenced by strategic processes. A novel finding of the present study is the similar N400 effect for categorical and associative priming in the short SOA condition. It thus appears that the N400 effect for categorical priming is short lived and largely decays within a few hundred milliseconds, whereas the N400 effect for associative priming is more durable and even extends in topography at longer SOAs.

In the word priming literature, it has been a matter of controversy whether the N400 priming effect is driven by an automatic spreading-activation mechanism or by strategic processes (for a review, see Kouider & Dehaene, 2007). As detailed in the Introduction, one critical aspect in this debate is prime/target SOA (Kiefer & Brendel, 2006; Kiefer & Spitzer, 2000). Several studies observed masked priming effects on the N400 at short SOAs (Grossi, 2006; Kiefer, 2002; Deacon et al., 2000), whereas others did not find comparable effects at longer SOAs (Ruz et al., 2003; Brown & Hagoort, 1993). Broadly in line with these studies, N400 for categorical priming was much reduced at a longer SOA, and its topography was restricted to a left parietal region. In strong contrast, N400 for associative priming was not reduced but even topographically more widespread at long compared with short SOA. This latter finding appears to be in remarkable contrast to word priming experiments and may thus reflect an exceptional relevance of associative relationships for semantic memory for people.

It might be argued that the N400 effects in the long SOA condition may be related to partial recognition of the primes. Specifically, a minority of recognized primes may have been processed strategically, which might in turn have increased the N400 effect. Two arguments render this possibility unlikely: First, if strategic processing had affected the results of the long SOA condition, a larger associative RT priming effect would have been predicted. This was obviously not the case. Second, the measure of prime recognizability at long SOA was very close to 0 and did not differ significantly from this value, rendering the possibility of strategic usage of the primes highly unlikely.

In our view, the N400 effects in the present study reflect an “early broad–later narrow” mechanism: Immediately after prime presentation, a large variety of relatively loosely related target persons become activated. As time elapses, this early broad activation decreases at the benefit of a more narrow activation of more closely related persons. As a consequence, although N400 effects were similar at the short SOA, at the long SOA categorically related targets only elicited a weak residual N400 effect, whereas associated targets exhibited a large and widespread N400. These findings are potentially related to mechanisms described in center–surround theory (see Carr & Dagenbach, 1990) or distributed feature models of semantic priming (McRae & Boisvert, 1998; McRae et al., 1997), which show that effects of relatedness or feature correlation change over time. Our results therefore provide important information with respect to the relative time courses of associative and categorical priming, although they may be less conclusive with respect to precise mechanisms of “associative” priming at long SOA. However, the present findings do provide evidence for a semantic nature of “associative” priming at short SOA, as neural correlates for associative and categorical priming were comparable in that condition, suggesting a shared mechanism underlying both phenomena. Because categorically related stimuli are highly likely to elicit priming through semantic information, it appears reasonable to assume that shared semantic information also underlies “associative” priming in this situation. As a qualification, further research will be needed to determine whether priming is mediated by nonstrategic access to information about co-occurrences and thus purely associative relatedness (Barry et al., 1998), or by more extensive semantic overlap (Carson & Burton, 2001) at longer SOAs.

Wiese and Schweinberger (2008) observed a posterior ERP correlate of categorical but not associative priming in the P2. In the present study, a similar effect was observed only as a statistical trend and only at long SOA. Thus, although at present it is not entirely clear what processes exactly are reflected in this P2 effect, it would appear that these processes benefit from strategic processing and take some time to emerge. It may also be noted that potentially similar effects in a comparable time range have been observed in word priming experiments, especially for categorically related targets (Hagoort, Brown, & Swaab, 1996) and at longer SOAs (Kiefer & Spitzer, 2000).

In sum, we suggest that the present N400 results can be explained by the following mechanisms: (i) Whenever a prime is presented, a fast increase in activation occurs for a large set of semantically related PINs. Thus, both “associated” (which are usually also semantically related, see Carson & Burton, 2001) and categorically related primes elicit similar N400 effects on processing target persons at very short SOAs. This mechanism largely corresponds to the one described in the IAC model (Burton et al., 1990). (ii) If prime/target SOA is sufficiently long, activation focuses to a smaller subset of those PINs, which are densely connected to the prime. At the same time, initial preactivations of only weakly connected PINs decay. As a result, the N400 priming effect increases over time for associated targets, whereas it largely decays for targets from the same category condition, which are related to the prime via occupation only.

It needs to be noted that this account is somewhat preliminary at present. Most importantly, although the present findings suggest that overlapping semantic information drives the N400 effect at short SOA for both categorical and associative priming, it is unclear what type of relatedness (overlapping semantic information, visual co-occurrence, or both) drives the focusing observed in the N400 priming effect over longer intervals. In addition, the specific nature of the P2 effect observed in the present and previous studies needs to be explored in more detail. Future studies are planned to address these issues.

A potential restriction for the present experimental approach is that categorical and associative relatedness cannot be completely disentangled when using celebrities as stimuli. Although it is possible to find categorically related prime/target pairs that are not associatively related (such as the same category pairs in the present study), most (if not all) associatively related people are also categorically related. A possible solution may be the use of learning paradigms, in which participants learn categorical information for, or associations between, preexperimentally unfamiliar faces in a learning phase. In principle, it should then be possible to disentangle categorical and associative priming effects for these newly learned stimuli. However, perhaps because of the difficulty to learn many faces, the only study on semantic priming in person recognition with such a design (Vladeanu, Lewis, & Ellis, 2006) used considerably fewer stimuli than would be required for an ERP design. Vladeanu et al. (2006) also used a long prime/target SOA and unmasked prime stimuli. Thus, with respect to present purposes, it can be assumed that their observed effects in RTs may not exclusively represent semantic memory phenomena but may have been confounded by episodic memory. Studies from the domain of word priming also suggest that very extensive learning is needed to obtain behavioral effects on the basis of newly established representations in semantic memory (Pecher & Raaijmakers, 1999; Dagenbach, Horst, & Carr, 1990). Nevertheless, future studies may make use of such paradigms to further our knowledge not only about the structure of semantic person knowledge but also about the mechanisms of incorporating new information into it.

Overall, the present study has revealed novel temporal information about mechanisms mediating associative and categorical priming in person recognition. Specifically, whereas both kinds of priming initially elicit comparable N400 effects at a very short SOA, diverging neural mechanisms for both types of priming emerged at a longer SOA: A prominent N400 effect was seen for priming of highly associated persons only, suggesting that initial activation may focus to more closely related persons as time elapses after the prime. Our results thus demonstrate the ability of ERPs to trace the neural dynamics of activation in different forms of priming and exemplify how such studies may be used to test different models of semantic priming.

Acknowledgments

This work was supported by a grant from the Deutsche Forschungsgemeinschaft (DFG Wi3219/2-1) to H. W. and S. R. S. The authors are grateful to Christine Seibt, Dorit Grundmann, Christin Knorr, Patrick Lorenz, André Preis, and Ina Udhardt for help during data acquisition.

Reprint requests should be sent to Dr. Holger Wiese, Department of General Psychology, Institute of Psychology, Friedrich Schiller University of Jena, Am Steiger 3, Haus 1, 07743 Jena, Germany, or via e-mail: holger.wiese@uni-jena.de, Web site: http://www2.uni-jena.de/svw/allgpsy/holger.htm.

REFERENCES

Barry
,
C.
,
Johnston
,
R. A.
, &
Scanlan
,
L. C.
(
1998
).
Are faces “special” objects? Associative and semantic priming of face and object recognition and naming.
Quarterly Journal of Experimental Psychology, Section A, Human Experimental Psychology
,
51
,
853
882
.
Bentin
,
S.
,
Mccarthy
,
G.
, &
Wood
,
C. C.
(
1985
).
Event-related potentials, lexical decision and semantic priming.
Electroencephalography and Clinical Neurophysiology
,
60
,
343
355
.
Berg
,
P.
, &
Scherg
,
M.
(
1994
).
A multiple source approach to the correction of eye artifacts.
Electroencephalography and Clinical Neurophysiology
,
90
,
229
241
.
Brown
,
C.
, &
Hagoort
,
P.
(
1993
).
The processing nature of the N400—Evidence from masked priming.
Journal of Cognitive Neuroscience
,
5
,
34
44
.
Bruce
,
V.
(
1983
).
Recognizing faces.
Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences
,
302
,
423
436
.
Burton
,
A. M.
,
Bruce
,
V.
, &
Johnston
,
R. A.
(
1990
).
Understanding face recognition with an interactive activation model.
British Journal of Psychology
,
81
,
361
380
.
Burton
,
A. M.
,
Young
,
A. W.
,
Bruce
,
V.
,
Johnston
,
R. A.
, &
Ellis
,
A. W.
(
1991
).
Understanding covert recognition.
Cognition
,
39
,
129
166
.
Caramazza
,
A.
, &
Shelton
,
J. R.
(
1998
).
Domain-specific knowledge systems in the brain: The animate-inanimate distinction.
Journal of Cognitive Neuroscience
,
10
,
1
34
.
Carr
,
T. H.
, &
Dagenbach
,
D.
(
1990
).
Semantic priming and repetition priming from masked words—Evidence for a center–surround attentional mechanism in perceptual recognition.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
16
,
341
350
.
Carson
,
D. R.
, &
Burton
,
A. M.
(
2001
).
Semantic priming of person recognition: Categorical priming may be a weaker form of the associative priming effect.
Quarterly Journal of Experimental Psychology, Section A, Human Experimental Psychology
,
54
,
1155
1179
.
Dagenbach
,
D.
,
Horst
,
S.
, &
Carr
,
T. H.
(
1990
).
Adding new information to semantic memory—How much learning is enough to produce automatic priming.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
16
,
581
591
.
Dark
,
V. J.
(
1988
).
Semantic priming, prime reportability, and retroactive priming are interdependent.
Memory & Cognition
,
16
,
299
308
.
Deacon
,
D.
,
Hewitt
,
S.
,
Yang
,
C. M.
, &
Nagata
,
M.
(
2000
).
Event-related potential indices of semantic priming using masked and unmasked words: Evidence that the N400 does not reflect a post-lexical process.
Cognitive Brain Research
,
9
,
137
146
.
Denkova
,
E.
,
Botzung
,
A.
, &
Manning
,
L.
(
2006
).
Neural correlates of remembering/knowing famous people: An event-related fMRI study.
Neuropsychologia
,
44
,
2783
2791
.
Douville
,
K.
,
Woodard
,
J. L.
,
Seidenberg
,
M.
,
Miller
,
S. K.
,
Leveroni
,
C. L.
,
Nielson
,
K. A.
,
et al
(
2005
).
Medial temporal lobe activity for recognition of recent and remote famous names: An event-related fMRI study.
Neuropsychologia
,
43
,
693
703
.
Elfgren
,
C.
,
van Westen
,
D.
,
Passant
,
U.
,
Larsson
,
E. M.
,
Mannfolk
,
P.
, &
Fransson
,
P.
(
2006
).
fMRI activity in the medial temporal lobe during famous face processing.
Neuroimage
,
30
,
609
616
.
Ellis
,
A. W.
,
Young
,
A. W.
, &
Critchley
,
E. M.
(
1989
).
Loss of memory for people following temporal lobe damage.
Brain
,
112
,
1469
1483
.
Engst
,
F. M.
,
Martin-Loeches
,
M.
, &
Sommer
,
W.
(
2006
).
Memory systems for structural and semantic knowledge of faces and buildings.
Brain Research
,
1124
,
70
80
.
Gobbini
,
M. I.
, &
Haxby
,
J. V.
(
2007
).
Neural systems for recognition of familiar faces.
Neuropsychologia
,
45
,
32
41
.
Gorno-Tempini
,
M. L.
, &
Price
,
C. J.
(
2001
).
Identification of famous faces and buildings—A functional neuroimaging study of semantically unique items.
Brain
,
124
,
2087
2097
.
Gorno-Tempini
,
M. L.
,
Price
,
C. J.
,
Josephs
,
O.
,
Vandenberghe
,
R.
,
Cappa
,
S. F.
,
Kapur
,
N.
,
et al
(
1998
).
The neural systems sustaining face and proper-name processing.
Brain
,
121
,
2103
2118
.
Grossi
,
G.
(
2006
).
Relatedness proportion effects on masked associative priming: An ERP study.
Psychophysiology
,
43
,
21
30
.
Hagoort
,
P.
,
Brown
,
C. M.
, &
Swaab
,
T. Y.
(
1996
).
Lexical-semantic event-related potential effects in patients with left hemisphere lesions and aphasia, and patients with right hemisphere lesions without aphasia.
Brain
,
119
,
627
649
.
Haist
,
F.
,
Gore
,
J. B.
, &
Mao
,
H.
(
2001
).
Consolidation of human memory over decades revealed by functional magnetic resonance imaging.
Nature Neuroscience
,
4
,
1139
1145
.
Hanley
,
J. R.
,
Young
,
A. W.
, &
Pearson
,
N. A.
(
1989
).
Defective recognition of familiar people.
Cognitive Neuropsychology
,
6
,
179
210
.
Henson
,
R. N.
,
Mouchlianitis
,
E.
,
Matthews
,
W. J.
, &
Kouider
,
S.
(
2008
).
Electrophysiological correlates of masked face priming.
Neuroimage
,
40
,
884
895
.
Holcomb
,
P. J.
,
Reder
,
L.
,
Misra
,
M.
, &
Grainger
,
J.
(
2005
).
The effects of prime visibility on ERP measures of masked priming.
Cognitive Brain Research
,
24
,
155
172
.
Kiefer
,
M.
(
2002
).
The N400 is modulated by unconsciously perceived masked words: Further evidence for an automatic spreading activation account of N400 priming effects.
Cognitive Brain Research
,
13
,
27
39
.
Kiefer
,
M.
(
2005
).
Repetition-priming modulates category-related effects on event-related potentials: Further evidence for multiple cortical semantic systems.
Journal of Cognitive Neuroscience
,
17
,
199
211
.
Kiefer
,
M.
, &
Brendel
,
D.
(
2006
).
Attentional modulation of unconscious “automatic” processes: Evidence from event-related potentials in a masked priming paradigm.
Journal of Cognitive Neuroscience
,
18
,
184
198
.
Kiefer
,
M.
, &
Spitzer
,
M.
(
2000
).
Time course of conscious and unconscious semantic brain activation.
NeuroReport
,
11
,
2401
2407
.
Kouider
,
S.
, &
Dehaene
,
S.
(
2007
).
Levels of processing during nonconscious perception: A critical review of visual masking.
Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences
,
362
,
857
875
.
Kutas
,
M.
, &
Hillyard
,
S. A.
(
1980
).
Reading senseless sentences—Brain potentials reflect semantic incongruity.
Science
,
207
,
203
205
.
Leveroni
,
C. L.
,
Seidenberg
,
M.
,
Mayer
,
A. R.
,
Mead
,
L. A.
,
Binder
,
J. R.
, &
Rao
,
S. M.
(
2000
).
Neural systems underlying the recognition of familiar and newly learned faces.
Journal of Neuroscience
,
20
,
878
886
.
Lucas
,
M.
(
2000
).
Semantic priming without association: A meta-analytic review.
Psychonomic Bulletin and Review
,
7
,
618
630
.
McNamara
,
T. P.
(
2005
).
Semantic priming: Perspectives from memory and word recognition.
New York
:
Psychology Press
.
McRae
,
K.
, &
Boisvert
,
S.
(
1998
).
Automatic semantic similarity priming.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
24
,
558
572
.
McRae
,
K.
,
deSa
,
V. R.
, &
Seidenberg
,
M. S.
(
1997
).
On the nature and scope of featural representations of word meaning.
Journal of Experimental Psychology: General
,
126
,
99
130
.
Mitchell
,
J. P.
,
Heatherton
,
T. F.
, &
Macrae
,
C. N.
(
2002
).
Distinct neural systems subserve person and object knowledge.
Proceedings of the National Academy of Sciences, U.S.A.
,
99
,
15238
15243
.
Oldfield
,
R. C.
(
1971
).
The assessment and analysis of handedness: The Edinburgh Inventory.
Neuropsychologia
,
9
,
97
113
.
Pecher
,
D.
, &
Raaijmakers
,
J. G. W.
(
1999
).
Automatic priming effects for new associations in lexical decision and perceptual identification.
Quarterly Journal of Experimental Psychology, Section A, Human Experimental Psychology
,
52
,
593
614
.
Rotshtein
,
P.
,
Henson
,
R. N. A.
,
Treves
,
A.
,
Driver
,
J.
, &
Dolan
,
R. J.
(
2005
).
Morphing Marilyn into Maggie dissociates physical and identity face representations in the brain.
Nature Neuroscience
,
8
,
107
113
.
Ruz
,
M.
,
Madrid
,
E.
,
Lupianez
,
J.
, &
Tudela
,
P.
(
2003
).
High density ERP indices of conscious and unconscious semantic priming.
Cognitive Brain Research
,
17
,
719
731
.
Schweinberger
,
S. R.
(
1996
).
How Gorbachev primed Yeltsin: Analyses of associative priming in person recognition by means of reaction times and event-related brain potentials.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
22
,
1383
1407
.
Schweinberger
,
S. R.
,
Pfutze
,
E. M.
, &
Sommer
,
W.
(
1995
).
Repetition priming and associative priming of face recognition—Evidence from event-related potentials.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
21
,
722
736
.
Sergent
,
J.
,
Ohta
,
S.
, &
Macdonald
,
B.
(
1992
).
Functional neuroanatomy of face and object processing—A positron emission tomography study.
Brain
,
115
,
15
36
.
Stone
,
A.
(
2008
).
Categorical priming of famous person recognition: A hitherto overlooked methodological factor can resolve a long-standing debate.
Cognition
,
108
,
874
880
.
Stone
,
A.
, &
Valentine
,
T.
(
2007
).
The categorical structure of knowledge for famous people (and a novel application of the centre–surround theory).
Cognition
,
104
,
535
564
.
Thompson
,
S. A.
,
Graham
,
K. S.
,
Williams
,
G.
,
Patterson
,
K.
,
Kapur
,
N.
, &
Hodges
,
J. R.
(
2004
).
Dissociating person-specific from general semantic knowledge: Roles of the left and right temporal lobes.
Neuropsychologia
,
42
,
359
370
.
Tranel
,
D.
,
Damasio
,
H.
, &
Damasio
,
A. R.
(
1997
).
A neural basis for the retrieval of conceptual knowledge.
Neuropsychologia
,
35
,
1319
1327
.
Vladeanu
,
M.
,
Lewis
,
M.
, &
Ellis
,
H.
(
2006
).
Associative priming in faces: Semantic relatedness or simple co-occurrence?
Memory & Cognition
,
34
,
1091
1101
.
Wiese
,
H.
, &
Schweinberger
,
S. R.
(
2008
).
Event-related potentials indicate different processes to mediate categorical and associative priming in person recognition.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
34
,
1246
1263
.
Young
,
A. W.
,
Flude
,
B. M.
,
Hellawell
,
D. J.
, &
Ellis
,
A. W.
(
1994
).
The nature of semantic priming effects in the recognition of familiar people.
British Journal of Psychology
,
85
,
393
411
.