Facial Expression Decoding Deficits in Clinical Populations with Interpersonal Relationship Dysfunctions

Facial Expression Decoding Deficits
in Clinical Populations with Interpersonal Relationship Dysfunctions

The existentialist French philosopher, Jean-Paul Sartre is famous for his statement 'Hell is the others.' This pessimistic stance is actually an 'a contrario' claim that the secret of happiness rests, at least partly, in the way we relate to others. At an even more basic level in our social species, interactive adjustment to and coordination with others are central to our survival. Successful interaction, adjustment, and coor­dination with others depend upon many emotional processes, and more specifically on emotional communication and coordination. A failure to adequately communicate one's emotional and motivational state and/or to accurately perceive the internal state of others is likely to result in interpersonal and personal problems. This notion is sup-ported by theories and empirical data relating nonverbal social skills and more general social competence, or psychopathology (e.g., Perez & Riggio, 2003).

Indeed, several lines of research have demonstrated that the capacity to accurately decode facial expression is an acquired skill that develops

until adolescence (Gross & Ballif, 1991). Further, this skill is related to more general social skills in adults (Patterson, 1999) as well as in chil­dren (Philippot & Feldman, 1990). Poor skills in decoding emotional fa­cial expression have been related to clinical conditions as various as depression (Bouhuys, 2003), alcohol dependency (Philippot, Kornreich & Blairy, 2003), or schizophrenia (Kring & Earnst, 2003). However, the causal direction of this relation remains an open issue: Are some clini­cal conditions a consequence of a basic emotional deficit, such as a defi­cit in decoding nonverbal expression of emotion, or is this latter deficit the consequence of the clinical condition?

One can speculate that many interpersonal problems might result from a deficit in decoding facial expression, whatever the direction of causality with clinical conditions. The most obvious problem is the dif­ficulty in identifying the internal states of others: their desires, emo­tions, or intentions. Such information is essential for the under-standing of others, of the meaning of their behavior in general as well as during social interaction. Relating to someone whose intentions and emotions are obscure is virtually impossible. Further, such a decoding deficit may make more likely the occurrence of interpretation bias, that is, erroneously attributing a given emotion to someone. For in-stance, people fearing social rejection might erroneously attribute con-tempt to people with whom they are interacting; A neutral face in this case might be misperceived as expressing non-interest, rejection, or even despise. Such hypothetical bias might surely impact on the be­havior of socially anxious individuals. This resulting behavior is likely to be perceived as awkward by their interaction partners. A social dis­tance would thus be created and would result in effective social dis­tance, ultimately confirming the fear of the socially anxious.

Another, more subtle, problem might result from a deficit in decod­ing facial expression. According to the self perception theory of Bem (1972), the way we perceive ourselves depends to a significant extent on how we imagine that others see us. In other words, the perception of ourselves results in part from how others do react to us. It follows that misinterpreting others' behavior and attitude toward us, including misinterpreting their facial expression when they are interacting with us, might result in the construction of an inaccurate social self, and ul­timately in a biased view of ourselves. For instance, a socially anxious individual erroneously decoding contempt in the faces of people with whom they are interacting, might end up believing that they are con-tempt deserving persons.

Thus, a deficit or a systematic bias in decoding facial expression might result in personal as well as in interpersonal difficulties. Some clinical populations seem to be particularly exposed to such problems. FYom the examples given above, it appears that social phobics are likely candidates for presenting systematic biases in the way they process facial expres­sions of others. To the opposite of the continuum, psychopaths, charac­terized by a lack of empathy and perverse interpersonal relationships

(Hare, 1998), might be particularly insensitive to the affective signal com municated via nonverbal behaviors. People suffering from a dependency to alcohol are renown to present important difficulties in their social and intimate relationships, difficulties that are often related to the regulation of frustration and anger (Sferrazza, Philippot, Kornreich et al., 2002). They might thus also present difficulties in understanding accurately the desires and intentions of others toward them. Other clinical populations, such as individuals suffering from paranoia may also be suspected of presenting deficits in the decoding of emotional expression. Indeed, for all these populations, a sound theoretical rationale can be constructed to support the notion that a nonverbal deficit might constitute a mainte­nance factor for their clinical condition. Similarly, for all the clinical con ditions mentioned above, this notion is supported by a strong conviction of clinicians working in the field.

The aim of the present contribution is to examine the empirical evi­dence for a nonverbal deficit in three clinical populations that are espe­cially characterized by difficulties in interpersonal relationships: social phobia, psychopathy, and alcohol dependence. For each popula­tion, we will review experiments from our and others' laboratories with three questions in mind: What do we know about potential non-verbal deficits or bias in that population that could be applied and used by practitioners?; at are the myths that need to be dispelled?; What are the current limitations of the area?

Before addressing these questions, we need to distinguish among the different types of deficits and biases that might be encountered. First, one should differentiate between deficits in the evaluation of the inten­sity of the emotion conveyed by the face, and the accuracy of the emotion attributed. In other words, one can over- or under-estimate the intensity of an emotion that is present on the face of the interaction partner; For example, the psychopath can underestimate the intensity of the sadness or distress expressed by the 'interaction partner'. In this case, we wi11 speak of evaluative deficit in intensity. We will speak of evaluative deficit in accuracy in the case of a general poor performance in the identifica­tion of the emotion conveyed by the face. This situation must be distin­guished from the situation in which one wrongly and systematically attributes an emotion X to a face, while it is actually emotion Y that is ex-pressed. In this latter case, we will speak of evaluative bias. Finally, an attentional bias consists in the fact that the perception threshold for cer tain facial expression is lower than for others. For example, socially anx­ious individuals might have their attention more readily attracted to faces expressing rejection than to other faces.

SOCIAL ANXIETY

The study of biases and deficits in the processing of interpersonal in-formation has generated a wealth of research in anxiety in general (e.g., Williams, Watts, MacLeod & Mathews, 1999) and in social anxiety in

particular (Clark & McManus, 2002). Most of this research focuses on attentional biases. Surprisingly little research has been devoted to evaluative biases and deficits, despite a strong belief in the clinical world that socially anxious individuals over-estimate threat in social signals (e.g., Beck, Emery & Greenberg, 1985).

In their cognitive-motivational model of anxiety, Mogg and Bradley (1998) have articulated attentional and evaluative biases. Their model relies on two different systems: The Valence Evaluation System and the Goal Engagement System. The Valence Evaluation System assesses the stimulus threat value according to the relevance of the stimulus to the person's preoccupation and learning experiences. The Goal Engage­ment System orients allocation of attention as a function of the output of the former system. If a stimulus in the environment is evaluated as threatening, the Goal Engagement System interrupts ongoing activities and orients attention toward the threat stimulus. This model postulates that attentional biases in anxious individuals result from a negative and unbalanced appraisal of social situations (Mogg & Bradley, 2002).

Attentional Biases

A wealth of studies has evidenced an attentional bias in the processing of threatening stimuli by socially anxious individuals (see Musa & Lepine, 2000, for review). However, the direction of these attentional biases is the object of a controversy. On one hand, several cognitive models of anxiety (e.g., Mogg & Bradley, 1998; Williams et al., 1999) propose that anxious individuals preferentially attend to threatening information (Beck, Emery, & Greenberg, 1985).

Different studies have demonstrated such a vigilance bias towards threat words by social phobics (Asmundson & Stein, 1994; Maidenberg Chen, Craske, & Bohn, 1996; Mattia, Heimberg, & Hope, 1993). Some authors have criticized the use of words to measure response to social cues (Chen, Ehlers, Clark, & Mansell, 2002): Responses to words would index worry rather than actual response to social stimuli. However, the same findings have been replicated with more ecological material­faces-by socials phobics (Gilboa-Schechtman, Foa & Amir, 1999) and in non-clinical samples with high fear of negative evaluation (FNE, Watson & Friend, 1969; Mogg & Bradley, 2002).

On the other hand, some researchers predict the opposite attentional bias. Clark (1999) has proposed that avoidance of threat­ening information may play an important role in the maintenance of social anxiety. For instance, actively avoiding social stimuli (e.g., faces) constitutes a form of cognitive escape from anxiety-provoking situa­tions for social phobics (e.g., avoiding looking at others' faces makes conversation less likely; Clark & Wells, 1995). Studies using probe de­tection tasks found that social phobics (Chen et al., 2002) and socially anxious individuals (Mansell, Clark, Ehlers, & Chen, 1999) avoid emotional (negative and positive) faces.

In an attempt to reconciliate these divergent findings, Amir, Fresh man and Foa (2002) have proposed a two-stage model of information processing. According to this view, anxious individuals would show an initial hypervigilance for threat-relevant stimuli. This hypervigilance would be the consequence of automatic processes, and it could be ob­served without conscious perception of threat-relevant information (Mogg & Bradley, 1999). However, at further and less automatic stages of information processing, people would actively turn away from threatening information. Thus, this model postulates a dynamic shift of attention allocation from initial threat hypervigilance to later threat avoidance. For instance, while speaking to other people, socially anx ious individuals would have their attention automatically attracted to frowns more readily than would non-anxious individuals. Because of this perception bias, socially anxious individuals are likely to automat ically over-activate a state of social anxiety. However, as soon as a frown was detected, they would turn their attention away from it-and, more generally, from others' faces-to avoid the threatening stimulus and the discomfort associated with it. Unfortunately, in doing so, they are likely to maintain their anxiety: Not only are they likely to behave so­cially inappropriately, but they will also be unable to determine whether the frowns were a sign of actual social threat or, for instance, simply a sign of perplexity.

Two studies (Amir, Freshman, & Foa, 2002; Amir, Mc Nally, Riemann, Burns, Lorenz & Mullen, 1996) manipulating strategic con trol in the Stroop task suggest that social phobics are able to modulate their attention to threat using strategic processes. However, this 'vigi lance-avoidance' hypothesis was not supported in a non-clinical sam ple of anxious individuals (Mogg, Bradley, de Bono, & Painter, 1997).

In an experiment using the dot prime paradigm (Mogg, Philippot, & Bradley, 2003), we examined the time course of attentional biases for faces in order to test the 'vigilance-avoidance' hypothesis. Social phobics and matched controls achieved a probe detection task with fa cial expressions as stimuli. In order to observe if the focus of attention changed over time, the stimulus duration was manipulated (either 500 ms or 1250 ms). It was predicted, and observed, that social phobics initially focus their attention on the threatening face, but that this attentional bias rapidly disappears. In contrast, non-phobics showed the opposite pattern. Similar results, using a different paradigm (ho­mograph paradigm) were reported by Amir, Foa, and Coles (1998). They fit nicely with our prediction of an initial automatic vigilance for threatening information, followed by a protective voluntary attempt to redirect attention away from the threatening stimulus.

Evaluative Deficits and Biases

In the previous section, we have stressed the strong belief that attentional biases result from evaluative biases. However, few studies

have tested this hypothesis. In a study by Merckelbach, Van Hout, Van den Hout, & Mersch (1989), social phobics and controls had to evaluate angry neutral, and joyful faces with respect to their pleasantness. Con­trary to the cognitive-motivational model's prediction (Mogg & Bradley, 1998), no differences were observed between the two populations.

We recently replicated this intriguing result (Douilliez & Philippot, 2003): Socially-anxious and control participants were asked to evalu­ate the threatening value of fearful, joyful, and neutral faces. In addi­tion, we extended the study to other types of stimuli: words and pictures, of which we manipulated valence and social relevance. Our rationale was that faces are potent innate stimuli (Ohman & Soares, 1993), and, as such, the processing of faces should not be influenced by social anxiety. In contrast, words and scenes depicted in the pic­tures require an interpretation and can therefore be affected by experi­ence, including social anxiety. As predicted, replicating Merckelbach et al. (1989), no differences between anxious individuals and controls were observed for the evaluation of faces. In contrast, anxious individ­uals evaluated negative pictures and words as more threatening, com­pared to evaluations by normal controls.

A possible explanation to the limitations of the study of Merckelbach et al. (1989) as well as our studies is that prototypical facial expressions were used, displaying full-blown emotions. These extreme stimuli have not only little ecological validity, but they are also easy to decode and the use of such a material is likely to produce ceiling effects (Hess, Blairy & Kleck, 1997). To avoid ceiling effects and to use a material reflecting real life expressions, we designed a study in which stimuli varied in emo­tional intensity (Philippot & Douilliez, 2003). Specifically, a series of emotional facial expressions constructed by Hess and Blairy (1995) was employed in which two actors portray five emotions (happiness, anger, sadness, disgust and fear) at four intensity levels (0%-i.e., neutral, 30%, 70%, 100%). These stimuli were presented in a random order on a computer screen. Finally, to increase the sensitivity of our measures, participants rated each facial expression on 7-point scales for a large profile of eight emotions (happiness, sadness, fear, anger, disgust, sur­prise, shame, and contempt).

This decoding task has been proposed to 17 out-patients diagnosed with social phobia according to DSM IV criteria, to 17 out-patients diag­nosed with another anxiety disorder (agoraphobia, general anxiety) ac-cording to DSM IV criteria and to 41 controls who were matched for sex, age, and level of education. The analysis of the data revealed no differ­ences among the three groups in terms of intensity ratings, accuracy or systematic biases, nor in their estimation of the difficulty of the task.

In conclusion, even if the 'vigilance-avoidance' model of anxiety is not fully supported in social anxiety, initial attentive biases toward threatening stimuli, including real life information such as facial ex­pressions, are supported by a wealth of empirical studies. However, socially anxious individuals do not seem to over or under-estimate the

intensity of an emotion present on the face, and they identify accurately the emotions conveyed by the face. Moreover, the evaluative biases are less likely to generate attentional biases than hypothesized by Mogg and Bradley (1998). Clearly, further research is needed to investigate the possibility of implicit as well as of explicit evaluative biases in the socially anxious and to examine the relationship between possible evaluative biases and attentional biases.

ALCOHOL DEPENDENCE

As suggested in the introduction, in their daily functioning, alcoholics are confronted with severe interpersonal problems (Duberstein, Conwell, & Caine, 1993), including the use of violence (Myers, 1984). Alcoholics seem to have difficulties dealing with negative emotions, and especially with anger (Marlatt, 1979). This observation has in-spired clinicians to design and evaluate communication training pro grams in the treatment of alcoholism. For instance, Monti et al. (1990) have compared the effectiveness of different treatment groups for alco­holic men. In a Communication Skills Training (CST) condition, par­ticipants were taught communication skills and interpersonal problem solving skills. In a Cognitive Behavioral Mood Management Training (CBMMT) condition, participants were taught how to control their desires to consume alcohol in difficult situations. The results showed that all treatments had a positive impact on social skills and on reducing anxiety in participants. CST was somewhat superior to CBMMT in this respect, attesting to the importance of communication deficit in alcoholics' problems. Moreover, participants in the CST con dition drank less alcohol up to six months after treatment than partici­pants in the CBMMT condition. In sum, this study suggests that emotion communication plays a very important role in the problems to which alcoholics are confronted.

We directly addressed the question of communication problems in alcoholism in a study focusing on emotion communication within cou ples with an alcoholic member (Sferrazza et al., 2002). Both wife and husband independently completed a questionnaire addressing the type, intensity, rumination about, and control of emotion, first for themselves, then for their spouses, and finally for what they believed their spouses were perceiving about their own (respondent's) emotion. Both partners from twenty-five alcoholic couples and twenty-five matched control couples participated in this study. Overall, the results showed marked differences in emotional experiences and expression between alcoholic couples and control couples. Interestingly, there were very few differences between the alcoholic member of the couple and his or her spouse. Specifically, alcoholic couples reported experi encing more intense emotions in general, and in particular for anger, guilt, sadness, anxiety, shame, and disgust. Interestingly, while alco­holics and their spouses reported to feel more guilt, they attributed

more anger to their partner. Alcoholic couples also reported less emo­tional control. When they spoke about their emotion, they felt more discomfort, they did not know how to react and how to express them-selves, and did not feel understood. They also attributed more negative and less positive effects to their emotional expression. Thus, com­pared to matched controls, both members of couples with an alcoholic member reported more intense and negative emotions, difficulties in expressing and controlling their emotions, and negative consequences of their emotional expression.

These observations are suggestive of an important deficit in emotion communication in alcoholics' families. The importance of communi­cative aspects in alcohol problems is further documented by the effec­tiveness of treatments focusing on communication training. Based on these observations, we developed the hypothesis that alcoholics suffer from deficits in nonverbal communication. There are several empiri­cal arguments suggesting such deficits. Some arguments pertain to the immediate effects of alcohol, while others are related to the effect of alcohol dependency.

Regarding the immediate effects of alcohol, it has been well demon­strated that alcohol impairs higher cognitive functioning and that this impairment impacts on several emotional processes (Lang et al., 1999). For instance, emotional appraisal appears to be impaired. This produces consequences both for the type of emotion that is experi­enced and expressed, and for the way nonverbal cues of emotion are decoded. Quite obviously, evaluative deficits in accuracy are expected when under the influence of alcohol. Second, alcohol changes expecta­tions and self perception (Cooper, Frone, Russell, & Mudar, 1995). When intoxicated, men are likely to behave more aggressively (Keane & Lisman, 1980), to express more anger nonverbally, and to interpret others' nonverbal cues as indicating provocation or threat (evaluative bias). Other reasons to suspect a nonverbal deficit are related to conse­quences of alcohol dependency. Alcoholics have difficulties dealing with negative emotions, especially with anger and frustration (Marlatt, 1996). They report more problems expressing their emotions and more negative consequences of such expression. A large part of emo­tion communication relies on nonverbal cues, and as social compe­tence and harmonious functioning require the mastery of nonverbal communication, the problems of alcoholics in solving interpersonal conflicts and in communicating their emotions are suggestive of a nonverbal deficit.

Based on these considerations, we propose that alcoholics are char­acterized by specific deficits in the decoding of nonverbal cues of emo­tion: They should over-perceive negative displays in others, especially those related to anger and frustration (evaluative bias). They should also be less accurate in general (evaluative deficit in accuracy). We fur­ther propose that this nonverbal deficit impairs alcoholics' social com­petence. They would be more likely to find themselves in interpersonal

conflicts, and more importantly, in such situations they would mis­attribute anger and hostile feelings to their partners. This would di­minish alcoholics' capacities to react efficiently and to find a constructive solution to the conflict that would remain unresolved. Al­coholics would then turn to alcohol as a coping strategy (although a faulty one).

The use of alcohol as an avoidant coping strategy is likely to main­tain interpersonal problems and even to increase them. A first positive feedback loop would be created: increased interpersonal tension would result in increased alcohol consumption, feeding back into the interpersonal tension. Further, as alcohol intoxication diminishes nonverbal decoding capacity, a second feedback loop would be cre­ated: alcohol intoxication would lead to more nonverbal impairments, the latter nourishing interpersonal tension, which then results in more alcohol consumption. This process is illustrated in Fig. 2.1.

We now turn to empirical evidences pertaining to facial expression decoding in alcoholics. Indeed, despite the importance of the question both from a clinical and from a theoretical perspective, few empirical studies have investigated nonverbal decoding skills in alcoholics. To our knowledge, the first experimental investigation of facial expression in alcoholics has been conducted by Oscar-Berman and colleagues (Oscar-Berman et al., 1990). They compared alcoholic Korsakoff pa­tients, non-Korsakoff alcoholics, and non-alcoholic controls regarding their ability to identify and recognize emotional material, including fa­cial expressions. They observed that alcoholic Korsakoff patients and non-Korsakoff alcoholics attributed more emotional intensity to facial expressions than controls (evaluative deficit in intensity). Further, the ability to match facial expressions with written labels was determined by the interaction between experimental group and age of the subject. Unfortunately, Oscar-Berman and colleagues did not specify nor inter­pret this interaction. Similarly, they did not explore alcoholics' accuracy in the decoding of facial expression.

Fig. 2. I . The cycle of nonverbal deficits, social competence deficits, and alcohol consumption.

In order to further document possible biases or impairments in the way alcoholics interpret emotional facial expression, we started a sys­tematic research program in our laboratories. In the first study (Philippot et al., 1999), we addressed three questions. First, we won­dered whether we could replicate the observation of Oscar-Berman et al. (1990) that alcoholics over-attribute emotional intensity to facial stimuli. Second, we examined whether alcoholics are less accurate than non-alcoholics in recognizing the type of emotion portrayed by a facial expression. Third, we asked if alcoholics show systematic biases in interpreting facial expression. In other words, do they tend to mis­attribute some types of emotion more than others?

We used exactly the same procedure as the one described for the study on social phobia in the preceding section (Philippot & Douilliez, 2003). The decoding task was proposed to 25 inpatients diagnosed with alcohol dependence according to DSM III-R criteria and to 25 con trols who were matched for sex, age, and level of education. Inpatients were in their third week of detoxification process and were not receiv­ing any psychotropic drugs at the time of assessment. The results demonstrated that alcoholics suffer from several deficits in the inter pretation of emotional facial expressions. First, compared to controls, they overestimated the intensity of the emotion conveyed by facial ex pressions, thereby replicating the observation of Oscar-Berman et al. (1990) with full-blown expressions and extending it to expressions of moderate and weak intensity and even to neutral faces: Alcoholics tend to perceive more intense emotion than controls in the faces of their interaction partners, even if no emotion is expressed.

Second, alcoholic participants misinterpreted facial expressions more than controls: They were more likely to believe that someone pre­senting a happy face was actually in a negative mood. They further tended to misattribute negative expressions (except for fear). For dis gust, they presented a systematic bias, attributing to their interaction partners' emotions of anger and contempt, two emotions typical of in terpersonal conflict. Finally, despite their poor performance, alcohol­ics did not report more difficulties with the decoding task than controls. It is thus likely that they do not perceive their deficit in the de coding of emotional facial expression. In sum, this first study portrays alcoholics as living in a world in which they perceive more emotional signals from their interaction partners, emotional signals that they tend to misinterpret with a negative and hostile bias, without noticing their deficits in this domain.

Alcoholic participants in this first study were inpatients at the end of the detoxification process. We do not know whether they already pre sented a facial expression decoding deficit before they became depend­ent on alcohol nor do we know whether the deficit is maintained in long-term abstinent alcoholics. Indeed, two interpretations of the defi­cits in the decoding of emotional facial expressions observed by Philippot et al. (1999) can be made. On the one hand, the deficits might

be the consequence of a general neurocognitive deterioration caused by alcohol that is known to impair multiple functions in chronic alco­holics. As most of these cognitive impairments remit with long-term abstinence, one would expect the deficits in the decoding of facial ex pression to be alleviated with long-term abstinence (Mann, Gunther, Selter, & Ackerman, 1999). On the other hand, emotional decoding deficits in alcoholics might be related to fundamental impairments that would precede the onset of alcohol dependency. Indeed, social skills deficits in alcoholics seem to be present before the onset of alcoholism (Rosenthal-Gaffney et al., 1998).

Following this question, we designed a second study (Kornreich et al., 2001b) in which we compared the performance of abstainers (for mer alcoholics, abstinent for at least several months) with the perfor­mance of recently detoxified alcoholics in the facial expression decoding task. If it could be shown that there are no differences be­tween these two populations, such an observation would rule out the possibility that the deficits are a consequence of a general cognitive de­terioration alleviating with abstinence. The analysis of the data re­vealed that, while some nonverbal impairments were no longer present in abstainers, others persisted. Specifically, the over-attribu tion of emotional intensity to facial expression was not observed in ab­stainers. Similarly, the mis-interpretation of happy and sad faces shown by recently detoxified alcoholics was not present in abstainers. However, their decoding accuracy deficit still persisted for anger and disgust facial expressions: For these emotions, there were no differences between recently detoxified and abstinent alcoholics.

Overall, this pattern of results suggests that different facets of alco holics' nonverbal impairments are determined by different processes. Some decline with time. Others seem to persist long after alcohol de­toxification, like the misinterpretation of some negative emotions. However, it remains to establish whether these deficits were present before the onset of alcohol dependency. Indeed, the fact that they re main, even years after the recession of alcohol abuse, does not imply that they are pre-existing or independent from alcohol abuse.

Another question that needs to be addressed is whether the precise nonverbal deficits that we have observed in alcoholics are specific to the alcoholic population (Sher, Trull, Bartholow, & Vieth, 1999). To partially answer this question, we replicated our first study, with two non-alcoholic control groups, one with psychopathology (i.e., obses­sive compulsive disorder, OCD) and one with no psychopathology (Kornreich et al., 2001a). We chose an OCD control group because al coholism and OCD display symptomatic similarities but do not share common etiologies. Indeed, several investigators have noted similari­ties between urges and desires to drink heavily and obsessive-compul­sive disorders (Anton, Moak, & Latham, 1995; Caetano, 1985; Edwards & Gross, 1976; Modell, Glaser, Cyr, & Montz, 1992). Fur­thermore, the life-time risk for obsessive compulsive disorder among

close relatives of alcoholics is 1.4 percent, which does not support the existence of a common genotype for the two disorders (Schuckit et al., 1995). It seemed therefore relevant to use an obsessive-compulsive sample as a control group with psychopathology.

We used the same procedure as in our former studies, but with a re­stricted set of stimuli, given the (obsessively) long response time of participants with OCD. Twenty-two outpatients suffering from obses­sive-compulsive disorder according to the DSM IV were recruited in a general hospital out-patient department. They were matched for age, sex, and educational level with 22 volunteers with no psychiatric re-cord and 22 inpatients diagnosed with alcohol dependence according to DSM IV criteria who were at the end of their detoxification process. The results of Study 1 were replicated: Recently detoxified alcoholics attributed more emotional intensity to facial stimuli, were less accu­rate in identifying the emotion portrayed, and did not report more dif­ficulties in the decoding task. The patients with OCD, however, did not differ from the normal controls. This observation supports the conclu­sions of our study comparing controls, anxious, and socially anxious outpatients-a study that had observed no differences among the three groups. Thus, the facial expression decoding deficits observed in alco­holics could not be found in OCD patients, or in another clinically anx­ious population. The social isolation and stigmatization shared by these conditions is thus unlikely to account for the nonverbal deficits observed in alcoholics.

This procedure was replicated in a study in which we compared post-cure groups presenting a dependency to opiate, to both opiate and alcohol, to alcohol only, and controls. The results indicated that partici­pants who presented a dependency to alcohol only, and to both opiate and alcohol had the worst accuracy scores. Opiate only dependent par­ticipants were more accurate that the latter, but still not as accurate as controls. It thus seems that alcohol dependency has a particularly pro­nounced effect on the accuracy of facial expression decoding. At the least, these results demonstrate that the deficits we evidenced in alco­holics are not ubiquitous in psychopathological populations. Still, more investigations are needed to establish how specific these deficits are, and how they relate to conditions of substance dependence and to the social exclusion often accompanying these conditions.

Above, we defended the notion that the impairments shown by alco­holics in the recognition of emotion from nonverbal cues might gener­ate interpersonal difficulties. These conflictive social relations might increase the probability of alcohol abuse. Alcohol intoxication might in turn impair the capacity of alcoholics to accurately interpret others' in­ternal states from their nonverbal behavior. They would then fall in a vicious circle, leading to more interpersonal conflict and to more alcohol consumption.

If this hypothesis is correct, the deficit in nonverbal decoding ob­served in alcoholics should be accounted for by their deficit in inter-

personal relations. To examine this possibility, we conducted a fourth study in which we replicated the procedure of Study 3 with 29 recently detoxified alcoholics and 29 controls matched for age, sex and educa tional level. In addition, we administrated to all participants the Inter-personal Problem Inventory of Horowitz et al. (1988). This scale comprises 127 items assessing six domains of potential interpersonal difficulties: assertiveness, sociability, submissiveness, intimacy, ex­cessive self-control, and excessive self-responsibility. Once again, the results indicated that alcoholics were less accurate in decoding facial expression, that they attributed more emotional intensity to the facial stimuli, but that they did not report more difficulties with the task than the control participants. As expected, alcoholics reported more inter-personal difficulties for all domains (excepted for self-control). We then examined whether the nonverbal decoding deficits of alcoholics were still statistically observable after partialling out the variance ac counted for by their interpersonal difficulties. The ANCOVAs revealed that alcoholics and controls were no more different in terms of nonver bal decoding accuracy, after partialling out the variance accounted for by interpersonal difficulties. This latter observation suggests that the relationship between nonverbal deficit and alcoholism is mediated by interpersonal problems and tension.

In conclusion, it appears that chronic alcoholics present three defi cits in the interpretation of facial expression. First, they over-estimate the intensity of the emotion felt by their interactant. Second, they de code facial expression less accurately than controls; they might also present a systematic bias in the over-attribution of anger and con-tempt, but we did not replicate this finding in all our studies. Third, al­coholics are not aware of their nonverbal deficits. This pattern of deficits seems specific to alcoholics, although more research is needed regarding this point. These deficits are enduring, as abstinent alcohol ics present the same pattern of deficits with the exception that they no longer over-estimate emotional intensity. Finally, these nonverbal defi­cits are related to interpersonal difficulties, which act as a mediator between nonverbal deficits and alcohol abuse.

Introducing this section, we have demonstrated that alcoholics tend to generate tension and conflict when interacting with others, includ ing their close relatives and family members. Furthermore, alcoholics present special difficulties in dealing with anger and frustration, two feelings that are often generated by interpersonal tension and con­flicts. Difficulties in dealing with and expressing these feelings are the best predictors of relapse (Marlatt, 1979). In other words, relapse pre­vention programs should focus on teaching alcoholics appropriate coping strategies and expression modes in situations in which they feel angry and/or frustrated.

Some research suggests that communication deficits, especially those relating to emotion, might be central in the deficient coping strat egies used by alcoholics. The mechanism that we propose is that be-

cause of their inability to correctly read others' emotional states, alcoholics generate interpersonal tensions and are less well-armed to solve these tensions constructively. Further, to avoid feelings of help­lessness generated by their inability to solve these situations, alcohol­ics turn to alcohol consumption as a coping strategy. They thus initiate two positive feedback loops. Alcohol intoxication first aggravates inter-personal tensions and second depletes the already limited nonverbal skills. This suggests that training programs, aimed at developing non-verbal sensitivity in alcoholics, should decrease interpersonal tension, increase appropriate coping skills, and consequently decrease alcohol consumption and relapse. Such training programs should especially focus on emotional intensity, and expression of emotion related to interpersonal tension such as anger, contempt, and disgust.

PSYCHOPATHY

One of the defining characteristics of psychopathy is the lack of empa­thy. One may thus suspect psychopaths of paying little attention to the emotional state of others, especially to states of distress, fear, or sad­ness. In this perspective, one might expect a poor performance for the decoding of facial expression of emotion in psychopaths, especially for distress, fear, or sadness. On the other hand, psychopaths have also been portrayed as having a 'superficial charm' and as being skilled manipulators. In this perspective, one would expect better perfor­mance in the decoding of facial expression in psychopaths than in con­trols. Which of these two plausible but contradictory hypotheses is best supported in the literature?

While several studies have provided consistent evidence of deficits for psychopaths in processing verbal emotional material (Williamson, Harpur, & He, 1990, 1991), thereby supporting the former hypothe­sis, evidences are much less consistent regarding nonverbal material. Actually, most studies that examined the meaning attributed to facial features did not observe differences between psychopaths and non psy­chopaths. For instance, Day and Wong (1996) observed hemispheric asymmetric differences between the two groups in a tachitoscopic task involving emotional words, but not in a similar task involving emotional faces. Likewise, Richell, Mitchell, Newman, Leonard, Baron-Cohen, and Blair (2003) did not observe differences between psychopaths and non psychopaths in a task requiring the identification of mental states from photographs of the eye region alone. In contrast, Stevens, Chap-man and Blair (2001) reported that children with psychopathic tenden­cies were impaired in the recognition of sad and fearful faces, but not of angry and happy faces. However, the samples were small (n = 9 for each group). Further, this observation was not replicated in adults samples by Kosson, Suchy, Mayer, and Libby (2002) who reported that psycho-paths' deficits were specific to the classification of disgust faces only when participants were required to use their left hand (i.e., in condi-

tons designed to minimize the involvement of left-hemisphere mecha nisms). Further, in that study, psychopaths were unexpectedly observed to be better at decoding anger when relying on left-hemisphere re-sources (when using their right hand).

To account for these discrepancies, authors often evoke the lack of sensitivity of the nonverbal tasks used (relying on full-blown facial ex­pression) as well as the fact that some studies did not distinguish among the emotion tones of facial expressions presented (Kosson et al., 2002). To overcome these weaknesses, we recently conducted a study in which we compared criminal psychopaths and non psycho-paths among inmates of a Belgian state prison and of a high security fo­rensic treatment facility. We used our highly sensitive nonverbal decoding task described in the preceding sections of this chapter (e.g., Philippot & Douilliez, in prep.; Philippot et at, 1999) and we further compared these criminal groups to men with no history of psychiatric disorder and with no criminal history. The analyses revealed that there were no differences between groups regarding the intensity of any emo tion they attributed to the facial stimuli (no evaluative deficit in inten sity), nor regarding the type of emotion they attributed to neutral faces. However, both criminal groups were less accurate than the controls, especially for fear and disgust. Further, psychopaths were less accu rate than controls for anger, and non psychopathic criminals were less accurate than controls for sadness. These effects were not affected by entering Education as a covariate (the controls were significantly more educated than the criminals). Finally, differences appeared in the diffi culty participants reported for the decoding task. Controls tended to report more difficulties than psychopaths, especially for weak inten­sity stimuli. Further, both criminal groups reported fewer difficulties than controls for decoding angry and sad faces. In sum, although both criminal groups reported fewer difficulties in decoding facial expres­sion of emotion, they were less accurate, especially for negative emotions. Importantly, no differences were observed between the two criminal groups in any dimension of the decoding task.

In conclusion, the psychopathic deficit consistently observed in the processing of emotional verbal material is not replicated with nonver­bal material. From the available evidence, one can conclude that if such a deficit exists, it should be highly specific and/or of weak inten­sity. Indeed, psychopathy is a rare syndrome resulting from a very complex combination of subtle deficits (Halle, Hodgins, & Roussy, 2000). We suggest that faces are basic social stimuli that require little to no reflexive processes to be decoded. However, the understanding of verbal material requires more elaborated processes. The latter would be impaired in psychopaths, but not the former. It is intriguing to no tice that the same pattern of results and rationale does apply to social phobics, whose social impairment (lack of assertiveness and irratio­nal social fear) is in many respects opposite to the social impairment of psychopaths (narcissism and total absence of social fear).

CONCLUSIONS AND DIRECTION FOR FUTURE RESEARCH

In this chapter, we have presented several lines of research investigat­ing how the interpretation of facial expression of emotion might be al­tered by clinical conditions characterized by difficulties in interpersonal regulation. Specifically, we studied social anxiety, alco­hol dependency, and psychopathy. For each condition, a logical ratio­nal could be formulated to support the notion that deficits in facial expression decoding were to be expected. Further, for each of these pa­thologies, there exists a strong clinical belief in the existence of bias or of deficit in the interpretation of others' emotional signals, and fore-most of facial expression. As we have demonstrated, these beliefs were in some cases myths, but in other cases reality.

The more surprising case is the one of social phobia. For this anxi­ety disorder, there is strong evidence of attentional biases in general, and of evaluative biases for affective words specifically. However, de-spite a strong belief and theoretical claim of the opposite, the evaluative bias for facial expression in social phobics turned out to be a myth. Although one cannot capitalize on the null hypothesis and ab­sence of differences, the lack of any differences between socially anx­ious individuals and non anxious in many different studies strongly suggests that if an evaluative bias exists, it should be at the least mod­est and of very little clinical significance. Similar conclusions could be made as regards to psychopaths. However, studies focusing on possi­ble attentional biases are still to be done to obtain a fuller picture of how this specific population processes nonverbal emotional information.

Finally, the case of alcoholism is much better documented regarding emotion facial expression recognition. We have reviewed ample and consistent evidence of interpretation biases and deficits in accuracy and evaluation of the intensity of facial expression. More research is needed to determine whether this profile of nonverbal decoding defi­cits is proper to alcoholism, or if it might also be found in other condi­tions related to substance dependence. We also need to investigate the possibility of attentional bias in alcoholics, especially for faces ex-pressing rejection and contempt. Still, the corpus of research pre­sented in this chapter is suggestive of many paths that can be directly exploited by clinicians, as the ones specified at the end of that section.

Still, in closing this chapter, we want to stress that the future of the field is not only dependent upon 'more research,' a common conclu­sion in many chapters! In our view, this area suffers from an important limitation of a basic theoretical nature: We urgently need a model speci­fying the processes involved in the decoding of facial expression of emo­tion. Indeed, despite three decades of intense empirical research on facial expression decoding, we are still uncertain about the processes implicated in the perception of emotional facial expression, and we ig­nore how affective meaning is attributed to faces. The little theoretical

basis we have is that different processes are active in the recognition of faces as compared to those implicated in the interpretation of facial ex­pression. Having a theoretical model of the processes involved in the de-coding of facial expression would offer a basis to infer hypotheses regarding which processes should be preserved and which one would be impaired in given pathologies. On this basis, one could predict the profile of decoding performances for given clinical conditions.

As in many areas of psychology, clinical research may help in the elaboration of such a fundamental model. Indeed, showing consisten­cies in the association or dissociation of some performances and defi­cits in given clinical conditions suggests that these performances are underlined by similar or different processes. For instance, we demon­strated that alcoholics' over-estimation of emotional intensity in faces disappears with long term abstinence, while the interpretation biases remain. This observation is suggestive that these two facets of the in­terpretation of emotional facial expression are supported by different processes. Similar inferences can be made from the observation that social phobics and psychopaths present biases in the processing of emotional verbal material, but not of nonverbal material.

It is our hope that the clinical work presented in this chapter, and the one that is forthcoming, will prove useful in the endeavor of build­ing a basic model of the decoding of facial expression of emotion.

ACKNOWLEDGMENTS

The writing of this chapter has been facilitated by grants from the 'Fonds National de la Recherche Scientifique de Belgique' 8.4512.98 and 3.4609.01. The authors appreciate helpful comments of Christopher Long on earlier drafts of this paper. Correspondence regarding this chap-ter should be addressed to Pierre Philippot who is at Faculte de Psycho-loge, Universite de Louvain, place du Cardinal Mercier, 10, B-1348 Louvain-la-Neuve, Belgique. Electronic mail may be sent via Internet to Pierre.Philippot@psp.ucl.ac.be Recent updates of our research can be found on Internet : www.ecsa.ucl.ac.bpersonnel/philippot/

REFERENCES

Amir, N., Freshman, M., & Foa, E. (2002). Enhanced Stroop interference for threat in social phobia. Journal of Anxiety Disorder, 16, 1-9.

Amir, N., McNally, R. J., Riemann, B. C., Burns, J., Lorenz, M., & Mullen, J. T. (1996). Suppression of the emotional Stroop effect by increased anxiety in patients with social phobia. Behaviour Research & Therapy, 34, 945-948.

Anton, R. F., Moak, D. H., & Latham, P (1995). The Obsessive Compulsive Drinking Scale: A self-rated instrument for the quantification of thoughts about alcohol and drinking behavior. Alcoholism: Clinical and Experimen­tal Research, 19, 92-99.

Asmundson, G. J. G., & Stein, M. B. (1994). Selective processing of social threat in patients with generalized social phobia: Evaluation using a dot-probe paradigm. Journal of Anxiety Disorders, 8, 107-117.

Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Cognitive structures and anxiogenic rules. In A. T. Beck, G. Emery, & R. L. Greenberg (Eds.), Anxiety disorders and phobias (pp. 54-66). New York: Basic books

Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 1-62). New York: Academic Press.

Bouhuys, A. L. (2003). Ethology and depression. In P Philippot, E. J. Coats & R. S. Feldman (Eds.). Nonverbal behavior in clinical context. New York: Oxford University Press.

Caetano, R. (1985). Alcohol dependence and the need to drink: A compulsion. Psychosomatic Medicine, 15, 463-469.

Chen, Y. P, Ehlers, A., Clark, D. M., & Mansell, W. (2002). Patients with social phobia direct their attention away from faces. Behaviour Research & Ther­apy, 40, 677-687.

Clark, D. M. (1999). Anxiety disorders: Why do they persist and how to treat them. Behavior Research and Therapy, 37, 5-27.

Clark, D. M., & MacManus, F. (2002). Information processing in social phobia. Biological Psychiatry, 51, 92-100.

Clark, D. M. & Wells, A. (1995). A Cognitive model of social phobia. In R. Heimberg, M. Liebowitz, D. A. Hope, & F. R. Schneier. Social phobia: diag­nosis, assessment and treatment (pp. 69-93). New York: Guilford Press.

Cooper, M. L., Frone, M. R., Russell, M., & Mudar, P. (1995). Drinking to regu late positive and negative moods: A motivational model of alcohol use. Jour­nal of Personality and Social Psychology, 69, 990-1005.

Day, R., & Wong, S. (1996). Anomalous perceptual asymmetries for negative emotional stimuli in the psychopath. Journal of Abnormal Psychology. 105, 648-652.

Douilliez, C., & Philippot, P. (2003). Biais dans l'evaluation volontaire de stim uli verbaux et non-verbaux: Effet de 1'anxiete sociale. Revue FYancophone de Clintque Comportementale et Cognitive.

Duberstein, P. R., Conwell, Y., & Caine, E. D. (1993). Interpersonal strcssors, substance abuse, and suicide. Journal of Nervous and Mental Disorders, 181. 80-85.

Edwards, G., & Gross, M. M. (1976). Alcohol dependence: Provisional de­scription of a clinical syndrome. British Medical Journal, 1, 1058-1061.

Gilboa-Schechtman, E., Fba, E. B., & Amir, N. (1999). Attentional biases for fa­cial expressions in social phobia: The face-in-the-crowd paradigm. Cogni­tion and Emotion, 13, 305-318.

Gross, A. L.. & Ballif, B. (1991). Children's understanding of emotion from facial expressions and situations: A review. Developmental Review, 11, 368-398.

Halle, P, Hodgins, S., & Roussy, S. (2000). Revue critique des etudes expert­mentales aupres de detenus adultes: Precision du syndrome de la psychopathie et hypotheses developpementales. Septentrion: Presses Universitaires de Lille.

Hare, R. D. (1998). Psychopathy, affect and behavior. In D. J. Cooke, A. E. Forth, & R. D. Hare (Eds.), Psychopathy: Theory, research and implica­tions for society (pp. 105-137). Boston: Kluwer Academic.

Hess, U., & Blairy, S. (1995). Set of emotional facial stimuli. Montreal, Can­ada: Department of Psychology, University of Quebec at Montreal.

Hess, U., Blairy, S., & Kleck, R. E. (1997). Intensity of emotional facial expres­sion and decoding accuracy. Journal of Nonverbal Behavior, 21, 241-257.

Horowitz, L. M., Rosenberg, S. E., Baer, B. A., Ureno, G., & Villasenor, V. S. (1988). Inventory of Interpersonal problems: Psychometric properties and clinical applications. Journal of Consulting and Clinical Psychology, 56, 885-892.

Keane, T. M., & Lisman, S. A. (1980). Alcohol and social anxiety in males: Be havioral, cognitive and physiological effects. Journal of Abnormal Psychol­ogy, 89, 213-223.

Kornreich, C., Blairy S., Philippot, P., Dan, B., Foisy, M. L., Le Bon, O., Pelc, I., & Verbanck, P (2001a). Impaired emotional facial expression recognition in alcoholism compared to obsessive compulsive disorder and normal con­trols. Psychiatric Research, 102, 235-248.

Kornreich, C., Blairy, S., Philippot, P, Hess, U., Noel, X., Streel, E., Le Bon, O., Dan, B., Pelc, I., & Verbanck, P (2001b). Deficits in Recognition of emotional facial expression are still present after mid to long-term abstinence in alco­holics. Journal of Studies on Alcohol, 62, 533-542.

Kosson, D. S., Suchy, Y., Mayer, A. R., & Libby, J. (2002). Facial affect recogni­tion in criminal psychopaths. Emotion, 2, 398-411.

Kring, A. M., & Earnst, K. S. (2003). Nonverbal behavior in schizophrenia. In P. Philippot, E. J. Coats, & R. S. Feldman (Eds.). Nonverbal behavior in clini­cal context. New York: Oxford University Press.

Lang, A. R., Patrick, C. J., & Strike, W. G. K. (1999). Alcohol and emotional response: A multidimensional-multilevel analysis. In K. E. Leonard & H. T. Blane (Eds.), Psychological theories of drinking and alcoholism (2nd. ed., pp. 328-371). New York: Guilford Press.

Maidenberg, E., Chen, E., Craske, M., Bohn, P et al., (1996). Specificity of attentional bias in panic disorder and social phobia. Journal of Anxiety Disorders, 10, 529-541.

Mann K., Gunther, A., Stetter, F., & Ackerman, K. (1999). Rapid recovery from cognitive deficits in abstinent alcoholics: A controlled test-retest study. Al­cohol and Alcoholics, 34, 567-574.

Mansell, W, Clark, D. M., Ehlers, A., & Chen, Y. P (1999). Social anxiety and at‑

tention away from emotional faces. Cognition and Emotion, 13, 673-690.
Marlatt, C. A. (1979). Alcohol use and problem drinking: a cognitive behav‑

ioral analysis. New York: Academic Press.

Marlatt, G. A. (1996). Taxonomy of high-risk situations for alcohol relapse: Evolution and development of a cognitive behavioral model. Addiction, 91 (suppl.), S37-S49.

Matta, J. I., Heimberg, R. G., & Hope. D. A. (1993). The revised Stroop color-naming task in socials phobics. Behavior Research and Therapy, 31, 305-313.

Merckelbach, H., Van Hout, W, Van den Hout, M. A., & Mersch, P P (1989). Psychophysiological and subjective reactions of social phobics and normals to facial stimuli. Behaviour Research and Therapy, 27, 289-294.

Modell, J. G., Glaser, F. B., Cyr, L., Montz, J. M. (1992). Obsessive and compul­sive characteristics of craving for alcohol in alcohol abuse and dependence. Alcoholism: Clinical and Experimental Research, 16, 272-274.

Mogg, K., & Bradley, B. P. (1999). Some methodological issues in assessing attentional biases for threatening faces in anxiety: A replication study using

a modified version of the probe detection task. Behaviour Research and Therapy, 37, 595-604.

Mogg, K., & Bradley, B. P. (2002). Selective orienting of attention to masked threat

faces in social anxiety. Behaviour Research and Therapy, 40, 1403-1414. Mogg, K., & Bradley, B. P. (1998). A cognitive-motivational analysis of anxiety.

Behavior Research and Therapy, 36, 809-848.

Mogg, K., Bradley, B. B, de Bono, J., & Painter, M. (1997) Time course of attentional bias for threat information in non-clinical anxiety. Behaviour Research and Therapy, 35, 297-303.

Mogg, K., Philippot, P, & Bradley, B. (2003). Selective attention to angry faces in a clinical sample with social phobia. Journal of Abnormal Psychology, 113, 160-165.

Musa, C. Z., & Lepine, J. P. (2000). Cognitive aspects of social phobia: a review of theories and experimental research. European Psychiatry, I5, 59-66.

Myers, T. (1984). Alcohol and violence: Self-reported alcohol consumption among violent and nonviolent male prisoners. In N. Karsner, J. S. Madden, & R. J. Walker (Eds.), Alcohol-related problems. Chichester: Wiley.

Ohman, A., & Soares, J. J. F. (1993). On the automatic nature of phobic fear: Conditioned electrodermal responses to masked fear-relevant stimuli. Journal of Abnormal Psychology, 102, 121-132.

Oscar-Berman, M., Hancock, M., Mildwordf, B., Hutner, N., & Altman-Weber, D. (1990). Emotional perception and memory in alcoholism and aging. Al­coholism: Clinical and Experimental Research, 14, 384-393.

Patterson, M. L. (1999). The evolution of a parallel process model of nonverbal communication. In P Philippot, R. S. Feldman, & E. J. Coats (Eds.), Non-verbal behavior in social context. New York: Cambridge University Press.

Perez, J. E., & Riggio, R. E. (2003). Nonverbal social skills and psychopath­ology. In P. Philippot, E. J. Coats, & R. S. Feldman (Eds. ), Nonverbal behav ior in clinical context. New York: Oxford University Press.

Philippot, P., & Feldman, R. S. (1990). Age and social competence in pre-schoolers' decoding of facial expression. British Journal of Social Psy­chology, 29, 43-54.

Philippot, B, & Douilliez, C. (2003). The decoding offacial expression of emo­tion in social phobics is not impaired. Manuscript in preparation.

Pppot, B, Baeyens, C., Douilliez, C., & FYancart, B. (2003). Cognitive regulation of emotion: Application to ccal disorders. In P Philippot & R. S. Feldman (Eds.), The regulation of emotion. New York: Lawrence ErlbaumAssociates.

Philippot, P. Kornreich, C., & Blairy, S. (2003). Nonverbal deficits and inter-personal regulation in alcoholics. In P Philippot, E. J. Coats, & R. S. Feldman (Eds.), Nonverbal behavior in clinical context. New York: Oxford University Press.

Philippot, P, Kornreich, C., Blairy, S., Baert, Y., Den Dulk, A., Le Bon, O., Verbanck, P., Hess, U., & Pelc, I. (1999). Alcoholics' deficits in the decoding of emotional facial expression. Alcoholism: Clinical and Experimental Re-search, 23, 1031-1038.

Richell, R. A., Mitchell, D. G. V., Newman, C., Leonard, A., Baron-Cohen, S., & Blair, R. J. R. (2003). Theory of mind and psychopathy: Can psychopathic in dividuals read the 'language of the eyes'? Neuropsychologia, 41, 523-526.

Rosenthal-Gaffney, L., Thorpe, K., Young, R., Colett, R., & Occhipinti, S. (1998). Social skills, expectancies, and drinking in adolescents. Addictive Behaviors, 23, 587-599.

Schuckit, M. A., Hesselbrock, V. M., Tipp, J., Nurnberger, J. I., Anthenelli, R. M., & Crowe R. R. (1995). The prevalence of major anxiety disorders in rela tives of alcohol dependent men and women. Journal of Studies on Alcohol, 56, 309-317.

Sferrazza, R., Philippot, P, Kornreich, C., Tang, C., Noel, X., Pelc, I., & Verbanck, P. (2002). Dysfonctionnement relationnel au sein des couples alcooliques. Alcoologie et Addictologie, 24, 117-125.

Sher, K. J., Trull, T. J., Bartholow, B. D., Vieth, A. (1999). Personality and co holism issues methods and etiological processes. In K. E. Leonard & H. T. Blane (Eds.), Psychological Theories of Drinking and Alcoholism (2nd ed.). New York: Guilford press.

Stevens, D., Chapman, T., & Blair, R. J. R. (2001). Recognition of emotion in fa­cial expressions and vocal tones in children with psychopathic tendencies. Journal of Genetic Psychology, 162, 201-211.

Watson, D., & Friend, R. (1969). Measurement of social-evaluative anxiety.

Journal of Consulting and Clinical Psychology, 33, 448-457.

Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1999). Cognitive

psychology and emotional disorders (2nd ed.). Chichester, UK: Wiley. Williamson, S., Harpur, T. T., & Hare, R. D. (1990). Sensitivity to emotional

polarity in psychopaths. Paper presented at the 98th Annual Convention of

the American Psychological Association, Boston.

Williamson, S., Harpur, T. T., & Hare, R. D. (1991). Abnormal processing of af­fective words by psychopaths. Psychophysiology, 28, 260-273.