Whole not hole: Expert face recognition requires holistic perception

Face recognition is an important ability of the human brain, yet its underlying mechanisms are still poorly understood. Two opposite views have been proposed to account for human face recognition expertise: the ability to extract the most diagnostic local information, feature-by feature (analytical view), or the ability to process all features at once over the whole face (holistic view). To help clarifying this debate, we used an original gaze-contingent stimulus presentation method to compare normal observers and a brain-damaged patient specifically impaired at face recognition (prosopagnosia). When a single central facial feature was revealed at a time through a gaze-contingent window, normal observers’ performance at an individual face matching task decreased to the patient level. However, when only the central feature was masked, forcing normal observers to rely on the whole face but the fixated feature, their performance was almost not affected. In contrast, the prosopagnosic patient's performance decreased dramatically in this latter condition. These results were independent of the absolute size of the face and window/mask. This dissociation indicates that expertise in face recognition does not rest on the ability to analyze diagnostic local detailed features sequentially but rather on the ability to see the individual features of a face all at once, a function that is critically impaired in acquired prosopagnosia.     

Neurodynamic Studies on Emotional and Inverted Faces in an Oddball Paradigm

The detection of a change in a face stimulus was studied in an oddball paradigm. Event-related potentials (ERPs) and MEG responses to face stimuli were recorded in four conditions: 1) happy standard, neutral deviant; 2) neutral standard, neutral deviant; 3) inverted happy standard, inverted neutral deviant; 4) inverted neutral standard, inverted neutral deviant. In all conditions, the target was a face with glasses. Neutral deviants elicited a negative deflection (with a maximum around 280 ms) in ERP and MEG responses, an effect similar to auditory mismatch negativity. Face inversion diminished deviance-related negativity, implying an important role of face recognition in the observed effect. Emotional content and larger physical differences between stimuli in conditions 1 and 3 compared to conditions 2 and 4 did not show statistically significant effect on the neutral-deviant-related negativity.     

The observer observed: Frontal EEG asymmetry and autonomic responses differentiate between another person's direct and averted gaze when the face is seen live

Recently, we showed that another person's gaze direction influenced the perceiver's frontal EEG asymmetry and autonomic arousal in response to freely viewed real faces, but not in response to face pictures. However, the lack of a task during the viewing may have resulted in less attention allocation to face pictures vs. live faces. In the present study, the participants performed two online tasks while viewing the faces presented live through an electronic shutter and as pictures on a computer screen. The results replicated those from our previous experiment showing that direct gaze elicited greater relative left-sided frontal EEG asymmetry and autonomic arousal than averted gaze but, again, only in the live condition. However, the results also showed that two live stimulus faces (male and female) elicited differential EEG asymmetry responses in our participants (all females), and the effects of gaze direction were observed only for the (live) female faces. The results suggest that the discriminative responses to live faces vs. pictures are likely to reflect the participants' enhanced mental-state attributions and self-awareness when looking at and being looked by live faces. Thus, the motivation- and affect-related psychophysiological responses to gaze direction are most discriminative in the presence of another person, regardless of whether the face/gaze is actively monitored or not.     

Configural processing of other-race faces is delayed but not decreased

The processing of other-race faces has been suggested to differ from own-race face processing with regard to the extent to which configural and/or holistic information is taken into account. We aimed at investigating the underlying mechanisms more precisely by applying event-related potentials (ERP). We presented upright and inverted own-race, other-race, and other-species faces (apes), as well as non-facial control stimuli (houses), and analysed the effects of stimulus type and inversion on the N170. Peak latencies for same-race, other-race, other-species, and non-face stimuli gradually increased, and inversion of all types of face stimuli led to an additional delay. Importantly, the two factors did not interact for face stimuli. Inversion increased N170 amplitudes for own- and other-race faces only. We conclude that early perceptual mechanisms of face processing are not qualitatively different for own- and other-race faces. Instead, the same underlying processes appear to work less efficiently for other-race faces.     

Event‐related potentials in response to cheating and cooperation in a social dilemma game

A sequential prisoner's dilemma game was combined with psychophysiological measures to examine the cognitive underpinnings of reciprocal exchange. Participants played four rounds of the game with partners who either cooperated or cheated. In a control condition, the partners’ faces were shown, but no interaction took place. The partners’ behaviors were consistent in the first three rounds of the game, but in the last round some of the partners unexpectedly changed strategies. In the first round of the game, the feedback about a partner's decision elicited a feedback P300, which was more pronounced for cooperation and cheating in comparison to the control condition, but did not vary as a function of feedback valence. In the last round, both the feedback negativity and the feedback P300 were sensitive to expectancy violations. There was no consistent evidence for a negativity bias, that is, enhanced allocation of attention to feedback about another person's cheating in comparison to feedback about another person's cooperation. Instead, participants focused on both positive and negative information, and flexibly adjusted their processing biases to the diagnosticity of the information. This conclusion was corroborated by the ERP correlates of memory retrieval. Successful retrieval of a partner's reputation was associated with an anterior positivity between 400 and 600 ms after face onset. This anterior positivity was more pronounced for both cooperator and cheater faces in comparison to control faces. The results suggest that it is not the negativity of social information, but rather its motivational and behavioral relevance that determines its processing.     

The role of stimulus type in age-related changes of visual working memory

Aging is accompanied by increasing difficulty in working memory associated with the temporary storage and processing of goal-relevant information. Face recognition plays a preponderant role in human behavior, and one might therefore suggest that working memory for faces is spared from age-related decline compared to socially less important visual stimulus material. To test this hypothesis, we performed working memory (n-back) tasks with two different visual stimulus types, namely faces and doors, and compared them to tasks with primarily verbal material, namely letters. Age-related reaction time slowing was comparable for all three stimulus types, supporting hypotheses on general cognitive and motor slowing. In contrast, performance substantially declined with age for faces and doors, but little for letters. Working memory for faces resulted in significantly better performance than that for doors and was more sensitive to on-line manipulation errors such as the temporal order. All together, our results show that even though face perception might play a specific role in visual processing, visual working memory for faces undergoes the same age-related decline as it does for socially less relevant visual material. Moreover, these results suggest that working memory decline cannot be solely explained by increasing vulnerability in prefrontal cortex related to executive functioning, but indicate an age-related decrease in a visual short-term buffer, possibly located in the temporal cortex.     

Following the time course of face gender and expression processing: A task-dependent ERP study

The effects of task demands and the interaction between gender and expression in face perception were studied using event-related potentials (ERPs). Participants performed three different tasks with male and female faces that were emotionally inexpressive or that showed happy or angry expressions. In two of the tasks (gender and expression categorization) facial properties were task-relevant while in a third task (symbol discrimination) facial information was irrelevant. Effects of expression were observed on the visual P100 component under all task conditions, suggesting the operation of an automatic process that is not influenced by task demands. The earliest interaction between expression and gender was observed later in the face-sensitive N170 component. This component showed differential modulations by specific combinations of gender and expression (e.g., angry male vs. angry female faces). Main effects of expression and task were observed in a later occipito-temporal component peaking around 230 ms post-stimulus onset (EPN or early posterior negativity). Less positive amplitudes in the presence of angry faces and during performance of the gender and expression tasks were observed. Finally, task demands also modulated a positive component peaking around 400 ms (LPC, or late positive complex) that showed enhanced amplitude for the gender task. The pattern of results obtained here adds new evidence about the sequence of operations involved in face processing and the interaction of facial properties (gender and expression) in response to different task demands.     

Bipolar patients show mood‐congruent biases in sensitivity to facial expressions of emotion when exhibiting depressed symptoms,but not when exhibiting manic symptoms

Introduction?A number of studies have reported mood‐congruent biases in processing facial expressions of emotion in depression and mania. Most of them have failed to establish that mood reliably affects relevant more than irrelevant expressions, or that the effect is specifically mood‐related rather than due to resource or task difficulty artefacts. The aim was to examine, using appropriate statistical methods, whether depressed mood in bipolar patients decreases and manic mood increases sensitivity to facial expressions of happiness and vice versa for facial expressions of negative emotion.

Methods?Sensitivity to facial expression of six basic emotions in bipolar patients when depressed and when manic was compared to closely matched controls.

Results?Mood‐related biases in sensitivity to facial expressions of happiness and of negative affect in general operate in persons with bipolar disorder when depressed. There is little evidence of similar biases in persons with bipolar disorder when manic.

Conclusions?These data show a mood‐congruent bias in sensitivity to facial expressions in bipolar depressed patients.     

孤独症儿童面孔认知ERP的PrN170研究

目的：探讨孤独症儿童与普通儿童在面孔识别事件相关电位（ERP）的差异。方法：以中性表情的面孔以及物件照片为刺激材料,以儿童喜欢物件的照片作为靶刺激,采用32通道ERP仪,分别记录20名孤独症儿童与22名普通儿童的ERP,比较两类儿童在不同刺激下N170电位的前体（Pre—cursorN170,PrN170）成分。采用SPSS13．0进行数据处理分析。结果：孤独症儿童组按键任务反应时较普通儿童组长（P〈0．05）,正确率无差别（P〉0．05）;两组儿童在面孔刺激呈现后约220ms颞枕叶导联发现PrN170,两组问对于不同类型的图片刺激,PrN170电位波幅与潜伏期均未见差异（P〉0．05）。结论：孤独症儿童的面孔结构编码过程无显著异常。