Markus F. Neumann

Perceiving age and gender in unfamiliar faces: brain potential evidence for implicit and explicit person categorization.

We used repetition priming to investigate implicit and explicit processes of unfamiliar face categorization. During prime and test phases, participants categorized unfamiliar faces according to either age or gender. Faces presented at test were either new or primed in a task-congruent (same task during priming and test) or incongruent (different tasks) condition. During age categorization, reaction times revealed significant priming for both priming conditions, and event-related potentials yielded an increased N170 over the left hemisphere as a result of priming. During gender categorization, congruent faces elicited priming and a latency decrease in the right N170. Accordingly, information about age is extracted irrespective of processing demands, and priming facilitates the extraction of feature information reflected in the left N170 effect. By contrast, priming of gender categorization may depend on whether the task at initial presentation requires configural processing.

Perceptual load manipulation reveals sensitivity of the face-selective N170 to attention.

It has been controversial whether the face-sensitive N170 is affected by selective attention. We manipulated attention sensu Lavies perceptual load theory to short (200 ms) presentations of task-irrelevant unfamiliar faces or houses, while participants identified superimposed target letters ‘X’ versus ‘N’. These targets were strings of either six identical (low load) or six different letters (high load). Under low load, we found a prominent face-selective N170 response. Under high load, however, we not only observed a dramatic reduction of the face N170 but also an unexpected enhancement of the house N170, such that face selectivity was almost completely lost. We conclude that the early stages of face processing indexed by the N170 strongly depend on selective attention.

Repetition effects in human ERPs to faces.

In the present paper, we review research conducted over the past 25 years addressing the effects of repeating various kinds of information in faces (e.g., pictorial, spatial configural, identity, semantic) on different components in human event-related brain potentials (ERPs). This body of evidence suggests that several ERP components are systematically linked to different functional components of face identity processing. Specifically, we argue (1) that repetition of the category of faces (categorical adaptation) strongly affects the occipitotemporal N170 amplitude, which is systematically suppressed when a face is preceded by another face, irrespective of its identity, whereas (2) the prototypicality of a faces second order spatial configuration has a prominent effect on the subsequent occipitotemporal P200. Longer-latency repetition effects are related to the processing of individual facial identities. These include (3) an ERP correlate of the transient activation of individual representations of repeated faces in the form of an enhanced occipitotemporal N250r as seen in repetition priming experiments, and (4) a correlate of the acquisition of individual face identity representations during learning as seen in a topographically similar long-lasting N250 effect. Finally, (5) the repetition of semantic information in familiar person recognition elicits a central-parietal N400 ERP effect. We hope that this overview will encourage researchers to further exploit the potential of ERPs to provide a continuous time window to neuronal correlates of multiple processes in face perception under comparatively natural viewing conditions.

N250r ERP repetition effects from distractor faces when attending to another face under load: Evidence for a face attention resource.

Recently, evidence for a face-specific attentional resource was suggested, which limits simultaneous processing to only one face. In the present Experiment 1, we manipulated perceptual load using two central item types (CITs: small central buildings or unfamiliar faces). To test whether distractor face processing is effectively prevented by face targets, CITs were superimposed on large famous distractor faces. ERPs were measured to subsequent faces, which could be a repetition or non-repetition of the previous distractor face. In Experiment 2, we used famous and unfamiliar faces as CITs under high load. For building CITs, we found common N250r repetition effects both under high and low load. For face CITs, N250r was reduced (Experiment 1) or even eliminated (Experiment 2) under high load. These findings support notions of a face-specific attentional resource which, at least under high demands, may limit processing to only one face at a time.

Reduced set averaging of face identity in children and adolescents with autism

Individuals with autism have difficulty abstracting and updating average representations from their diet of faces. These averages function as perceptual norms for coding faces, and poorly calibrated norms may contribute to face recognition difficulties in autism. Another kind of average, known as an ensemble representation, can be abstracted from briefly glimpsed sets of faces. Here we show for the first time that children and adolescents with autism also have difficulty abstracting ensemble representations from sets of faces. On each trial, participants saw a study set of four identities and then indicated whether a test face was present. The test face could be a set average or a set identity, from either the study set or another set. Recognition of set averages was reduced in participants with autism, relative to age- and ability-matched typically developing participants. This difference, which actually represents more accurate responding, indicates weaker set averaging and thus weaker ensemble representations of face identity in autism. Our finding adds to the growing evidence for atypical abstraction of average face representations from experience in autism. Weak ensemble representations may have negative consequences for face processing in autism, given the importance of ensemble representations in dealing with processing capacity limitations.

Attentional spread in deaf and hearing participants: Face and object distractor processing under perceptual load

The case of human deafness constitutes a unique opportunity to examine possible consequences for perceptual processing due to altered sensory experiences. We tested whether deaf—in contrast to hearing—individuals are more susceptible to visual distraction from peripheral than from central face versus object stimuli. The participants were required to classify the gender of a target male or female symbol presented either alone (low perceptual load) or together with three filler symbols (high perceptual load), while ignoring gender-congruent or -incongruent face versus object distractors presented at central or peripheral positions. The gender classifications were affected by distractor gender under low, but not under high, perceptual load in hearing participants. In contrast, the responses of deaf participants were similarly influenced by distractor gender under both levels of perceptual load. There was no evidence for generally enhanced attention to the visual periphery in deaf individuals. Our results indicate that auditory deprivation may result in enhanced attentional capacities under high perceptual load.

Combined effects of attention and inversion on event-related potentials to human bodies and faces.

We investigated effects of attentional load and inversion on event-related potentials to body or face distractors. Participants performed demanding (high load) or less demanding (low load) unrelated letter-search tasks. Bodies and faces were intact (Experiment 1) or without heads or eyes (Experiment 2). We measured prominent P100, N170, and late occipito-temporal negative (LNC) components. N170 to bodies had smaller and more anterior maxima than faces. N170 to intact bodies and faces was increased by inversion, relatively independently of load. Inversion effects were dramatically reduced for headless bodies, and even reversed for eyeless faces. Load effects were most prominent in LNC, with enhanced negativity under low load. We suggest that N170 reflects mandatory, category-specific initial distractor encoding in body- or face-sensitive cortical areas, a process which may depend on interactive encoding of hierarchical cues (bodies, heads, eyes). By contrast, LNC mainly reflects residual capacity allocated to extended processing of task-irrelevant distractors.

Event-related potential correlates of repetition priming for ignored faces.

An attentional capacity limit was recently suggested for faces, such that only one face can be processed at a time. We measured interference and repetition priming caused by irrelevant distractor faces. Participants initially performed male/female judgments for central faces or symbols flanked by distractor faces. Interference (slower responses for sex-incongruent target–distractor pairs) occurred for central symbols but was absent for central faces. In subsequent fame judgements, previously presented distractor faces had no repetition priming effect on response times. Relative to new faces, event-related brain potentials revealed a right occipitotemporal negativity ∼400–600 ms for faces previously shown as distractors flanking central symbols (but not distractors flanking faces). These findings support a face-specific attentional capacity limit, showing that event-related brain potential priming effects can reveal covert distractor processing.

The own-age bias in face memory is unrelated to differences in attention—Evidence from event-related potentials

Participants are more accurate at remembering faces from their own relative to a different age group (the own-age bias, or OAB). A recent socio-cognitive account has suggested that differential allocation of attention to old versus young faces underlies this phenomenon. Critically, empirical evidence for a direct relationship between attention to own- versus other-age faces and the OAB in memory is lacking. To fill this gap, we tested the roles of attention in three different experimental paradigms, and additionally analyzed event-related brain potentials (ERPs). In Experiment 1, we compared the learning of old and young faces during focused versus divided attention, but revealed similar OABs in subsequent memory for both attention conditions. Similarly, manipulating attention during learning did not differentially affect the ERPs elicited by young versus old faces. In Experiment 2, we examined the repetition effects from task-irrelevant old and young faces presented under varying attentional loads on the N250r ERP component as an index of face recognition. Independent of load, the N250r effects were comparable for both age categories. Finally, in Experiment 3 we measured the N2pc as an index of attentional selection of old versus young target faces in a visual search task. The N2pc was not significantly different for the young versus the old target search conditions, suggesting similar orientations of attention to either face age group. Overall, we propose that the OAB in memory is largely unrelated to early attentional processes. Our findings therefore contrast with the predictions from socio-cognitive accounts on own-group biases in recognition memory, and are more easily reconciled with expertise-based models.

Ensemble coding of face identity is not independent of the coding of individual identity

Information about a group of similar objects can be summarized into a compressed code, known as ensemble coding. Ensemble coding of simple stimuli (e.g., groups of circles) can occur in the absence of detailed exemplar coding, suggesting dissociable processes. Here, we investigate whether a dissociation would still be apparent when coding facial identity, where individual exemplar information is much more important. We examined whether ensemble coding can occur when exemplar coding is difficult, as a result of large sets or short viewing times, or whether the two types of coding are positively associated. We found a positive association, whereby both ensemble and exemplar coding were reduced for larger groups and shorter viewing times. There was no evidence for ensemble coding in the absence of exemplar coding. At longer presentation times, there was an unexpected dissociation, where exemplar coding increased yet ensemble coding decreased, suggesting that robust information about face identity might suppress ensemble coding. Thus, for face identity, we did not find the classic dissociation—of access to ensemble information in the absence of detailed exemplar information—that has been used to support claims of distinct mechanisms for ensemble and exemplar coding.