Neuropsychologia 72 (2015) 52–58
Contents lists available at ScienceDirect
Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia
Congenital anosmia and emotion recognition: A case-control study Cédric Lemogne a,b,c,n, Julien Smadja a,b, El-Hadi Zerdazi b, Yaël Soudry b, Marion Robin d, Sylvie Berthoz d, Frédéric Limosin a,b,c, Silla M. Consoli a,b, Pierre Bonfils a,e,f a
Université Paris Descartes, Sorbonne Paris Cité, Faculté de médecine, 15 rue de l’Ecole de Médecine, 75015 Paris, France AP-HP, Hôpitaux Universitaires Paris Ouest, Service de Psychiatrie de l’adulte et du sujet âgé, 20 rue Leblanc, 75015 Paris, France c Inserm U894, Centre Psychiatrie et Neurosciences, 2 ter Rue d’Alésia, 75014 Paris, France d Inserm U1178 & Institut Mutualiste Montsouris, Universités Paris-Sud & Paris Descartes, 97 Bd Port Royal, 75014 Paris, France e AP-HP, Hôpitaux Universitaires Paris Ouest, Service d’ORL et de chirurgie cervico-faciale, 20 rue Leblanc, 75015 Paris, France f CNRS UMR MD 8257, Cognition and Action Group, Centre Universitaires des Saints-Pères, 45 rue des Saints Pères, 75006 Paris, France b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 February 2015 Received in revised form 15 April 2015 Accepted 24 April 2015 Available online 25 April 2015
Patients with anosmia are not able to detect volatile chemicals signaling the presence of infectious and non-infectious environmental hazards, which typically elicit disgust and fear, respectively. Social animals may compensate a loss of olfaction by taking advantage of signals of threat that are produced by their conspecifics. Among humans and other primates, body postures and facial expressions are powerful cues conveying emotional information, including fear and disgust. The aim of the present study was to examine whether humans with agenesis of the olfactory bulb, a rare disorder characterized by congenital anosmia, would be more accurate in recognizing facial expressions of fear and disgust. A total of 90 participants with no history of mental disorder or traumatic brain injury were recruited, including 17 patients with congenital anosmia (10 men, mean age7standard deviation: 36.5 7 14.8 years), 34 patients with acquired anosmia (18 men, mean age7 standard deviation: 57.27 11.8 years) and 39 healthy subjects (22 men, mean age7 standard deviation: 36.7 7 13.2 years). For each patient with congenital anosmia, the agenesis of the olfactory bulb was ascertained through magnetic resonance imaging. Emotion recognition abilities were examined with a dynamic paradigm in which a morphing technique allowed displaying emotional facial expressions increasing in intensity over time. Adjusting for age, education, depression and anxiety, patients with congenital anosmia required similar levels of intensity to correctly recognize fear and disgust than healthy subjects while they displayed decreased error rates for both fear (mean difference [95% confidence interval] ¼ 28.3% [ 46.3%, 10.2%], P ¼0.003) and disgust (mean difference [95% confidence interval] ¼ 15.8% [ 31.5%, 0.2%], P¼ 0.048). Furthermore, among patients with acquired anosmia, there was a negative correlation between duration of anosmia and the rate of errors for fearful (Spearman's ρ ¼ 0.531, P¼ 0.001) or disgust (Spearman's ρ ¼ 0.719, Po 0.001) faces recognition. No significant difference was observed for the other primary emotions. Overall, these results suggest that patients with congenital anosmia and long-lasting acquired anosmia may compensate their inability to detect environmental hazards through olfaction by an increased ability to detect fear or disgust as facially expressed by others. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Acquired anosmia Congenial anosmia Emotion Facial expressions Olfaction
1. Introduction Among several functions that olfaction deserves in mammals, a critical one is to drive behavioral avoidance from two kinds of environmental hazards signaled by volatile chemicals: non-infectious hazards, such as fire, predators or degraded air, and infectious hazards, such as feces, spoiled food or organic decay n Correspondence to: Unité de Psychologie et Psychiatrie de liaison et d’urgence, Hôpital Européen Georges Pompidou, 20 rue Leblanc, 75908 Paris Cedex 15, France. Fax: þ33 1 56 09 31 46. E-mail address: [email protected] (C. Lemogne).
http://dx.doi.org/10.1016/j.neuropsychologia.2015.04.028 0028-3932/& 2015 Elsevier Ltd. All rights reserved.
(Stevenson, 2010). Once detected, non-infectious hazards typically elicit fear, whereas infectious hazards typically elicit disgust (Davey, 2011). Associations between specific odors and avoidance are both innate and learned in mammals, especially rodents (Kobayakawa et al., 2007), but learned associations prevail in humans (Stevenson et al., 2010). Among humans, individuals with acquired anosmia are no longer able to detect several environmental threats such as burning food, smoke or gas leak (Bonfils et al., 2008; Miwa et al., 2001; Santos et al., 2004; Temmel et al., 2002). In addition, individuals with anosmia are also concerned with the fear of eliciting disgust in others through bodily odors (Miwa et al., 2001; Temmel et al., 2002). Avoiding environmental hazards and social
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
rejection were arguably major adaptive problems that our ancestors had to deal with (Cosmides and Tooby, 2013; Eisenberger, 2012). Emotion expression, including facial expression, is likely to have been promoted by natural selection to help solving adaptive problems. This idea can be traced back at least to 1872 in ‘The Expression of the Emotions in Man and Animals’ by Charles Darwin (Darwin, 1965). In the context of natural selection, the most limiting resource is not food or safety or access to mates, but the information required for making adaptive behavioral choices (Cosmides and Tooby, 2013). When it comes to identify environmental hazards, social animals not only rely on their own senses, including olfaction, but may also take advantage of signals of threat that are produced by their conspecifics. Among humans and other primates, body postures and facial expressions are powerful means to convey emotional information, including fear and disgust (de Gelder et al., 2004; Ekman et al., 1969). Discrete facial expressions associated with the six primary emotions (i.e., happiness, sadness, anger, surprise, fear and disgust) are readily recognized regardless of cultural differences (Ekman et al., 1969; Mesquita and Frijda, 1992) and can be produced by congenitally blind individuals (Matsumoto and Willingham, 2009). Beside explicit recognition, facial expressions of fear or disgust may also elicit analogous emotional states in the observer through activation of shared neural representations (Davey, 2011; Whalen et al., 2004). Overall, there is strong evidence that the recognition of emotions expressed by others might help humans to make adaptive behavioral choices (Cosmides and Tooby, 2013). Building on evidence linking anosmia with impaired ability to identify environmental hazards (Bonfils et al., 2008; Miwa et al., 2001; Santos et al., 2004; Temmel et al., 2002), we hypothesized that humans may compensate a lack of olfaction by taking advantage of facial expressions of fear or disgust that are produced by their conspecifics when they face these hazards. Moreover, humans demonstrate an implicit association between disease and smell (Bulsing et al., 2009) and ostracism may indeed partially derive from disgust-related behaviors leading to the avoidance of individuals that may be sick (Curtis, 2011) or prone to moral transgressions (Chapman and Anderson, 2013). Fear of rejection by others is a strong drive in humans (Rotge et al., 2015) and may lead to increased monitoring of other facial expressions of avoidance, including disgust. Therefore, we hypothesized that individuals with long-lasting anosmia would be more accurate in recognizing both fear and disgust. Our objective was to examine this hypothesis in humans with congenital anosmia, a rare disorder characterized by a lack of olfaction at birth. In the present study, we assessed the ability to recognize facial emotional expressions of individuals with normal olfaction, acquired anosmia and congenital anosmia due to agenesis of the olfactory bulb. Agenesis of the olfactory bulb results from abnormal embryological development of the olfactory system and may be isolated or associated with more complex malformation syndromes (e.g. Kallman syndrome). To our knowledge, the association of anosmia with the recognition of fear or disgust has never been studied apart from patients with traumatic brain injury (Neumann et al., 2012) or mental disorders (Kohler et al., 2007). Since participants in the present study were free of such conditions, only subtle differences were expected. Therefore, we used a dynamic paradigm in which a morphing technique allowed displaying emotional facial expressions increasing in intensity over time, from neutrality to a fully expressed emotion (Domes et al., 2008; Lynch et al., 2006; Robin et al., 2012). We hypothesized that individuals with congenital anosmia or long-lasting acquired anosmia, relative to healthy individuals, would correctly identify fear and disgust portrayed at lower levels of intensity (i.e., a faster recognition) and/or would be more accurate in recognizing
53
expressions of fear and disgust (i.e, a lower rate of errors).
2. Materials and methods 2.1. Participants Patients presenting with either congenital or acquired anosmia were recruited from the Head and Neck Surgery department of the European Georges Pompidou Hospital (Paris, France). Inclusion criteria were: being aged 18–74; mastery over French language. Exclusion criteria were: history of mental disorder or traumatic brain injury and current use of psychotropic medication. Eligible patients were proposed to participate after a medical consultation. Healthy subjects with no history of olfactory disease were recruited from the community to constitute a control group. For each participant, the entire experimental procedure was conducted by the same investigator (i.e., JS, EZ or YS). Each participant was given a short questionnaire covering socio-demographic and psychological data (i.e. depressive symptoms and anxiety). Participants were then screened for Axis I disorders of the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, with the Mini-International Neuropsychiatric Interview (Sheehan et al., 1998). Finally, they underwent the emotion recognition test and the olfactory tests. Type of anosmia and duration of the disease were collected from the clinical records. Participants' consent was obtained according to the Declaration of Helsinki and the study has been approved by the local ethical committee. 2.2. Measures of depressive symptoms and anxiety Since depression and anxiety may affect emotion recognition, all participants were given the 21-item Beck Depression Inventory (BDI-21) to assess depressive symptoms (Beck et al., 1988) and the Y-A form of the State-Trait Anxiety Inventory (STAI) to assess trait anxiety (Spielberger et al., 1993). The BDI-21 is a brief, reliable, and valid questionnaire of 21 items mainly related to the cognitive features of depressive mood such as pessimism, sense of failure, or guilt. For each item, subjects endorse one of the four statements that are proposed according to the perception of their current state. Each item is subsequently scored from 0 to 3 according to the endorsed sentence, yielding a total score ranging from 0 to 39 (cronbach's alpha ¼0.90 in the present study). The trait scale of the STAI includes 20 sentences describing various psychological symptoms of anxiety that are rated by the subjects from 1 (not endorsed) to 4 (fully endorsed) yielding a total score ranging from 20 to 80 (cronbach's alpha ¼0.91 in the present study). 2.3. Emotion recognition test Emotion recognition was assessed with a task that comprised 36 trials presented in a random order (Robin et al., 2012). Each trial began with a neutral face gradually morphed into one of the six primary emotions (i.e., happiness, sadness, anger, surprise, fear and disgust) using forty 2.5% incremental stages (Fig. 1). Each picture was presented for 500 ms followed immediately by the next morphed face in the sequence. Each trial therefore consisted of a 20 s continuum. We used the pictures of three men and three women, who were each expressing the six emotions of interest. Each target emotion was thus portrayed six times. Participants were told that they would see faces appearing on the computer screen, whose expressions would slowly and continuously evolve from neutral to one of the six target emotions. They were asked to watch the changes in facial expression and report the emotion expressed whenever they thought they had
54
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
Fig. 1. Stimulus example for the expression of fear.
identified it, by clicking with the mouse on one of the six corresponding boxes (located below the stimulus). They were asked to rather wait until they had recognized the expression than simply guessing. Participants were informed that they would not be provided with any information on the quality (correct or incorrect) of their response, and that the face would continue to change even after their response. They were also told that they could change their initial response at any time and as often as necessary by clicking again on one of the response buttons until the end of the trial. Participants were also instructed that, at the end of each trial (40th stage), they would have to indicate their final choice. A practice phase on each of the six emotions was first presented in order to familiarize the participant with the procedure. This practice phase involved two different actors – one man and one woman - than those used for the main task in order to avoid a familiarization bias with faces. For each emotion, three scores can be computed: (1) the final success rate (i.e., percentage of correct responses at the 40th stage), (2) the number of stages required for the accurate identification of the emotion (ranging from 1 to 40), and (3) the percentage of first responses that were incorrect. 2.4. Olfactory tests In addition to clinical records for patients and self-report for controls, olfaction was assessed with either the Biolfa test (Lecanu et al., 2002) or the Sniffin Test (Hummel et al., 1997), depending upon the period of recruitment. Only quantitative measures of olfactory thresholds will be reported here. The Biolfa test measures the olfactory threshold for three substances (i.e., 2-phenylethanol, eugenol and aldehyde C14) and yields 3 scores ranging from 1 to 9. The Sniffin test measures the olfactory threshold for either the 2-phenylethanol or the n-butanol and yields a score ranging from 1 to 16, with greater scores indicating better olfaction. Here, only the n-butanol was used. For both tests, a score of 1 was given to the participants who failed to detect the substance at the greatest concentration. 2.5. Statistical analyses Differences between the three groups regarding continuous and discrete variables were examined with one-way ANOVAs and χ² tests, respectively. To test our a priori hypotheses, we first computed 3 2 ANCOVAs with a three-level group factor (i.e., patients with congenital anosmia, acquired anosmia, healthy subjects) and a two-level emotion factor (i.e., fear and disgust versus happiness, sadness, anger and surprise), entering age, education, depression and anxiety as covariates. Thus the model included the group and emotion main effects, the group emotion interaction and the interactions of the emotion factor with each covariate. Should a significant group emotion interaction be found, post hoc pairwise comparisons across groups for each emotion type (i.e., fear and disgust versus happiness, sadness, anger and surprise) were based on estimated means computed
with one-way ANCOVAS with a three-level group factor and the same covariates. Exploratory analyses followed regarding each of the six primary emotions, separately. Since we had strong a priori hypotheses regarding the recognition of fear and disgust, specifically, significance level was set at Po 0.05. For each significant difference, mean difference and 95% confidence interval (95% CI) were computed. All the analyses were conducted with the PASW Statistics 18 software (Chicago: SPSS Inc.).
3. Results 3.1. Participants A total of 90 participants were recruited, including 17 patients with congenital anosmia, 34 patients with acquired anosmia and 39 healthy subjects. For each patient with congenital anosmia, the agenesis of the olfactory bulb was ascertained through magnetic resonance imaging (Fig. 2). The olfactory bulb agenesis was isolated for 13 patients and associated with a Kallman syndrome for 4 patients. Acquired anosmia resulted from post-viral anosmia (N ¼15), nasal polyposis (N ¼ 9), head trauma without brain injury (N ¼3) and undetermined causes (i.e., idiopathic anosmia) (N ¼7). The characteristics of the three groups are displayed in Table 1. The three groups differed regarding age, education and depression. There was also a trend regarding differences in anxiety (P ¼0.062). Therefore, subsequent analyses were adjusted for age, education, depression and anxiety by entering these variables as covariates. Because of limited availability of the olfactory tests, only a subsample of the healthy subjects (N ¼15) were tested for olfaction. None of the healthy subjects had olfactory complaint and those who were tested had normal olfaction. 3.2. Emotion recognition There was no significant group main effect or group emotion interaction regarding the rate of final correct recognition or the stage for correct recognition (all P Z0.248). However, there was a significant group main effect regarding the error rate at first response (F¼3.235; df ¼2, 83; P ¼0.044; η² ¼0.072), which was driven by a significant group emotion interaction (F¼ 5.088; df¼ 2, 83; P¼ 0.008; η² ¼0.109). The other interactions were not significant (all PZ 0.171). In line with our hypotheses, post hoc pairwise comparisons across groups based on estimated means found that patients with congenital anosmia displayed lower error rates at first response than healthy subjects for fear and disgust (Mean difference [95% CI] ¼ 21.8% [ 35.9%, 7.7%]; t¼3.077; df¼83; P ¼0.003) with no other significant between-group difference (all P Z0.067). There was no significant between-group difference regarding the other four primary emotions (all PZ 0.290). Likewise, exploratory analyses regarding each primary emotion separately found that patients with congenital anosmia displayed lower error rates at first response than healthy subjects for both fear (Mean difference [95% CI] ¼ 28.3% [ 46.3%,
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
55
Fig. 2. MRI coronal view of the olfactory tract of a healthy subject (A) and of a patient with congenital anosmia (B).
10.2%]; t ¼3.111; df¼83; P¼ 0.003) and disgust (Mean difference [95% CI] ¼ 15.8% [ 31.5%, 0.2%]; t ¼2.013; df¼ 83; P ¼0.048) (Table 2, Fig. 3). In contrast, there was no significant betweengroup difference regarding the other four primary emotions (Table 2). There was no interaction between group and emotion type when considering fear versus disgust only (P ¼0.376), suggesting that between-group differences were similar for these two emotions. There was no indication of a greater tendency for patients with congenital anosmia, compared to healthy subjects, to falsely recognize fear (P ¼0.681) or disgust (P ¼0.344) when presented with other emotions. Contrary to patients with congenital anosmia, those with acquired anosmia did not display lower error rate at first response than healthy subjects for fear (P ¼0.259) or disgust (P ¼0.971)
(Table 2, Fig. 3). However, post hoc analyses among patients with acquired anosmia found a negative correlation between the duration of anosmia and the error rate at first response for both fear (Spearman’s ρ ¼ 0.531, P ¼0.001) and disgust (Spearman's ρ ¼ 0.719, P o0.001) (Fig. 4). In other words, the longer the duration of anosmia was, the lower the rate of errors for fear and disgust. This issue was further explored using between-group comparisons (i.e., long-lasting versus recent acquired anosmia) based on a median split of the duration of anosmia (i.e. 3 years). Comparing the two groups of patients with acquired anosmia with healthy controls, there was a significant group main effect regarding the error rate at first response (F¼ 8.409; df¼2, 66; P¼ 0.001; η² ¼0.203), which was driven by a significant group emotion interaction (F ¼4.032; df¼ 2, 66; P¼ 0.022; η²¼ 0.109).
Table 1 Characteristics of the participants. Healthy subjects (N ¼39)
Patients with congenital anosmia (N¼ 17)
Patients with acquired anosmia (N¼ 34)
Χ²
P
Discrete variables Male gender
N 22
% 56.4
N 10
% 58.8
N 18
% 52.9
0.179
0.914
Continuous variables Age Education (years) STAI-Trait BDI-21 Anosmia duration (years) Biolfa Testa 2-Phenylethanol (1–9) Eugenol (1–-9) Aldehyde C14 (1–9) Sniffin Testb n-Butanol (1–16)
Mean 36.7 16.8 35.7 5.2
SD 13.2 3.4 9.3 6.0
Mean 36.5 14.5 42.8 10.3
SD 14.8 3.0 9.4 8.6
Mean 57.2 15.1 38.6 8.8 5.5
SD 11.8 4.0 11.6 8.7 6.2
F 26.562 3.267 2.880 3.354
P o 0.001 0.043 0.062 0.040
7.3 7.7 7.5
0.9 1.3 1.5
1.6 1.4 1.4
1.5 1.5 1.3
1.5 1.6 1.5
1.3 1.8 1.5
111.54 82.485 93.876
o 0.001 o 0.001 o 0.001
1
1.4
0.8
a b
The Biolfa Test was used for 15 healthy subjects, 16 patients with congenital anosmia and 22 patients with acquired anosmia. The Sniffin Test was used for one patient with congenital anosmia and 12 patients with acquired anosmia.
56
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
The other interactions were not significant (all P Z0.200). Patients with long-lasting acquired anosmia displayed lower error rates for fear and disgust than both healthy subjects (Mean difference [95% CI] ¼ 22.5% [ 37.9%, 7.1%]; t¼ 2.913; df ¼66; P¼ 0.005) and patients with more recent acquired anosmia (Mean difference [95% CI]¼ 30.4% [ 46.1%, 14.6%]; t¼ 3.845; df ¼66; P o0.001). Table 2 Emotion recognition performances. Healthy subject Patients with con(N ¼ 39) genital anosmia (N ¼ 17)
Patients with acquired anosmia (N¼ 34)
Mean
SD
Mean
SD
Mean
SD
Correct final recognition rate (%) Happiness Sadness Anger Surprise Fear Disgust
100.0 83.8 80.3 87.6 85.5 76.1
0.0 17.7 17.9 15.2 21.0 16.1
100.0 86.3 77.5 81.4 85.3 75.5
0.0 16.9 22.0 23.5 14.3 14.6
98.0 86.7 78.9 77.4 73.5 78.9
9.0 15.8 16.0 21.3 20.5 18.0
Stage of correct recognition (1–40) Happiness Sadness Anger Surprise Fear Disgust
19.2 27.1 28.6 24.4 29.3 26.6
3.7 4.1 3.5 4.6 3.6 4.9
21.4 29.0 30.0 25.3 30.9 27.3
5.7 3.8 4.1 4.5 3.6 4.3
21.9 28.6 29.6 27.2 31.1 29.0
5.6 4.3 4.3 4.6 3.6 4.4
Error rate at first response (%) Happiness Sadness Anger Surprise Fear Disgust
2.6 11.9 12.1 9.8 33.5a 21.4b
6.2 23.0 18.6 14.7 29.4 23.7
2.0 7.8 11.4 6.6 10.4a 11.7b
5.5 18.6 21.8 13.9 18.0 24.2
7.2 13.6 19.5 12.2 34.1 24.5
14.1 23.8 26.9 23.4 35.3 29.0
a P ¼0.003 for the comparison of patients with congenital anosmia versus healthy controls, adjusting for age, education, depression and anxiety in general linear model. b P¼0.048 for the comparison of patients with congenital anosmia versus healthy controls adjusting for age, education, depression and anxiety in general linear model.
4. Discussion The aim of the present study was to examine whether individuals with congenital or long-lasting acquired anosmia may display enhanced ability to recognize facial expressions of fear and disgust. We found mixed results in favor of this hypothesis. Indeed, patients with congenital anosmia presented no advantage for detecting fear and disgust at lower levels of emotional intensity than healthy subjects while they displayed a decreased rate of errors for these two emotions. Furthermore, these results seem to be specific to fear and disgust, as suggested by a significant interaction between group and emotion type (i.e. fear and disgust versus the other four primary emotions). Finally, as regards patients with acquired anosmia, we found a negative correlation between duration of anosmia and the error rates for the recognition of fear and disgust. In other words, the longer the illness duration was, the greater the accuracy for recognizing fear and disgust. Overall, these results suggest that patients with congenital anosmia and long-lasting acquired anosmia may compensate their inability to detect environmental hazards through olfaction by increasing their ability to detect fear or disgust as expressed by others. One may argue that these results could be explained by either a more careful appraisal of the stimuli by patients with congenital anosmia, or by a greater tendency to consider facial expressions as expressing fear or disgust, regardless of the actual emotion that is portrayed. Nevertheless, patients with congenital anosmia did not display longer delays for correctly identifying fear and disgust than healthy subjects, providing no support for the first hypothesis. In addition, patients with congenital anosmia were not more likely to be biased towards misrecognition of fear or disgust than healthy subjects when presented with other emotions, ruling out the second hypothesis as well. Of particular note, the association between congenital anosmia and the recognition of fear appeared to be somewhat stronger than with disgust, though there was no significant interaction between group and emotion type when considering only fear and disgust. Apart from phenomenological aspects, several features distinguish fear from disgust on a physiological level, including facial mimicry and their neural bases. For instance, facial expressions of fear activates the amygdala, but not the insula, whereas facial expressions of disgust activates the insula, but not the amygdala (Fusar-Poli et al., 2009). More specific to the present
Fig. 3. Mean error rate at first response and upper limit of the 95% confidence interval across groups and emotions. (a) P¼ 0.003 for the comparison of patients with congenital anosmia versus healthy controls, adjusting for age, education, depression and anxiety in general linear model. (b) P¼ 0.048 for the comparison of patients with congenital anosmia versus healthy controls adjusting for age, education, depression and anxiety in general linear model.
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
57
Fig. 4. Error rate at first response for fear and disgust as a function of duration of anosmia among patients with acquired anosmia. Fit lines were drawn with locally weighted scatterplot smoothing based on an Epanechnikov kernel.
study is the fact that the recognition of facial expression of disgust occurs later in development than the recognition of fear. Although 5-year-old children are able to produce facial and verbal expressions of disgust and to infer disgust in others from a situation or a behavior, the ability to infer disgust in others from facial mimicry only gradually improves until the late teens (Widen and Russell, 2013). In contrast, 5-year old children are able to readily recognize fear in others, even from partial facial expressions (Gagnon et al., 2014). In other words, the association between odors and disgust might be learned several years before the association between facial expressions of disgust in others and this emotion. Should the first association facilitate the learning of the second one (e.g., by perceiving disgusting odors and facial expressions of disgust in others, simultaneously), the absence of olfaction at birth may impede the development of disgust recognition among patients with congenital anosmia. This might account for the somewhat smaller effect observed with disgust. Strength of the present study includes its novelty, the use of a sensitive protocol of emotion recognition and the recruitment of patients with congenital anosmia, a rare disorder characterized by the absence of olfaction at birth. These patients offer the unique opportunity to examine the association between a complete lack of olfaction and the ability to recognize facial expressions, particularly of fear and disgust. In addition, the participants were carefully screened for potential confounding effects, such as a history of mental disorders, with a standardized interview and assessed for subclinical levels of depression and anxiety with validated measures. Previous studies that addressed the association of anosmia with the recognition of facial emotional expressions included patients with traumatic brain injury (Neumann et al., 2012) or mental disorders (Kohler et al., 2007). Given that olfaction and emotion share some neural substrates (Soudry et al., 2011), any brain disorder may confound their association. Abnormal olfaction has indeed been proposed as a marker of psychopathology (Atanasova et al., 2008; Naudin et al., 2012) and traumatic brain injury may be associated with impaired recognition of facial expressions of emotion (Biszak and Babbage, 2014). Finally, psychotropic medications also influence emotion recognition abilities (Harmer, 2013). It is thus noteworthy that the present results were observed in participants free of brain disorders and psychotropic medications, and adjusted for subclinical levels of depression and anxiety. Some limitations should be noted. First, one may argue that differences regarding fear and disgust resulted from chance because of multiple comparisons. However, these results were a priori expected as exposed in the introduction. In addition, there was no significant difference as regards the other four primary
emotions (i.e. happiness, sadness, anger and surprise), whatever the measure. Furthermore, the interaction between group and emotion type (i.e., fear and disgust versus happiness, sadness, anger and surprise) was significant, suggesting that the results may indeed be specific to fear and disgust. However, we cannot rule out that potential ceiling effects might account for null findings regarding the other four primary emotions. Further studies are warranted to confirm the specificity of our results regarding fear and disgust. Increased attention to the facial expression of others to detect fear and disgust may indeed have resulted in better emotion recognition regardless of emotion type. Second, our experimental design did not allow testing the implicit recognition of emotions. Future studies might use subliminal stimuli and physiological measures to address this issue. Finally, dynamic facial recognition paradigms might be viewed as more ecological than static paradigms, but duration of 20 s is very far from the duration of expression of emotion by humans and constitutes a limit of morphing-based studies. In conclusion, we used a sensitive measure of emotion recognition in a population of patients with congenital and acquired anosmia to test the role of olfaction in the recognition of facial expressions of fear and disgust. Our results suggest that individuals with long-standing anosmia, either congenital or acquired, may develop enhanced ability to recognize these two emotions, possibly to compensate their inabilities to detect olfactory signals of both infectious and non-infectious environmental hazards.
Conflict of interests None of the authors have any conflict of interests to report.
Role of funding sources None.
Acknowledgments The authors thank Patrick Faulcon for his help in recruiting the participants and Laurence Laccourreye for having reviewed the manuscript.
58
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
References Atanasova, B., Graux, J., El Hage, W., Hommet, C., Camus, V., Belzung, C., 2008. Olfaction: a potential cognitive marker of psychiatric disorders. Neurosci. Biobehav. Rev. 32, 1315–1325. Beck, A., Steer, R., Garbin, M., 1988. Psychometric properties of the Beck depression inventory-25 years of evaluation. Clin. Psychol. Rev. 8, 77–100. Biszak, A.M., Babbage, D.R., 2014. Facial affect recognition difficulties in traumatic brain injury rehabilitation services. Brain Injury 28, 97–104. Bonfils, P., Faulcon, P., Tavernier, L., Bonfils, N.A., Malinvaud, D., 2009. Home Accidents Associated With Anosmia. vol. 37. Presse Médicale, Paris, France, pp. 742–745. Bulsing, P.J., Smeets, M.A.M., Van den Hout, M.A., 2009. The implicit association between odors and illness. Chem. Senses 34, 111–119. Chapman, H.A., Anderson, A.K., 2013. Things rank and gross in nature: a review and synthesis of moral disgust. Psychol. Bull. 139, 300–327. Cosmides, L., Tooby, J., 2013. Evolutionary psychology: new perspectives on cognition and motivation. Annu. Rev. Psychol. 64, 201–229. Curtis, V., 2011. Why disgust matters. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 366, 3478–3490. Darwin, C., 1965. The Expression of the Emotions in Man and Animals. University of Chicago Press. Davey, G.C.L., 2011. Disgust: the disease-avoidance emotion and its dysfunctions. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 366, 3453–3465. De Gelder, B., Snyder, J., Greve, D., Gerard, G., Hadjikhani, N., 2004. Fear fosters flight: a mechanism for fear contagion when perceiving emotion expressed by a whole body. Proc. Natl. Acad. Sci. USA 101, 16701–16706. Domes, G., Czieschnek, D., Weidler, F., Berger, C., Fast, K., Herpertz, S.C., 2008. Recognition of facial affect in Borderline Personality Disorder. J. Pers. Disord 22, 135–147. Eisenberger, N.I., 2012. The neural bases of social pain: evidence for shared representations with physical pain. Psychosom. Med. 74, 126–135. Ekman, P., Sorenson, E.R., Friesen, W.V., 1969. Pan-cultural elements in facial displays of emotion. Science 164, 86–88. Fusar-Poli, P., Placentino, A., Carletti, F., Landi, P., Allen, P., Surguladze, S., Benedetti, F., Abbamonte, M., Gasparotti, R., Barale, F., Perez, J., McGuire, P., Politi, P., 2009. Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional magnetic resonance imaging studies. J. Psychiatry Neurosci. 34, 418–432. Gagnon, M., Gosselin, P., Maassarani, R., 2014. Children's ability to recognize emotions from partial and complete facial expressions. J. Genet. Psychol. 175, 416–430. Harmer, C.J., 2013. Emotional processing and antidepressant action. Curr. Top. Behav. Neurosci. 14, 209–222. Hummel, T., Sekinger, B., Wolf, S.R., Pauli, E., Kobal, G., 1997. “Sniffin” sticks': olfactory performance assessed by the combined testing of odor identification, odor discrimination and olfactory threshold. Chem. Senses 22, 39–52. Kobayakawa, K., Kobayakawa, R., Matsumoto, H., Oka, Y., Imai, T., Ikawa, M., Okabe, M., Ikeda, T., Itohara, S., Kikusui, T., Mori, K., Sakano, H., 2007. Innate versus learned odour processing in the mouse olfactory bulb. Nature 450, 503–508. Kohler, C.G., Barrett, F.S., Gur, R.C., Turetsky, B.I., Moberg, P.J., 2007. Association between facial emotion recognition and odor identification in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 19, 128–131.
Lecanu, J.B., Faulcon, P., Werner, A., Bonfils, P., 2002. [Normative data of the Biolfa (R) olfactory test]. Ann. Oto-Laryngol. Chir. Cervico Faciale Bull. Soc. Oto-Laryngol. Hôp. Paris 119, 164–169. Lynch, T.R., Rosenthal, M.Z., Kosson, D.S., Cheavens, J.S., Lejuez, C.W., Blair, R.J.R., 2006. Heightened sensitivity to facial expressions of emotion in borderline personality disorder. Emot. Wash. DC 6, 647–655. Matsumoto, D., Willingham, B., 2009. Spontaneous facial expressions of emotion of congenitally and noncongenitally blind individuals. J. Pers. Soc. Psychol. 96, 1–10. Mesquita, B., Frijda, N.H., 1992. Cultural variations in emotions: a review. Psychol. Bull 112, 179–204. Miwa, T., Furukawa, M., Tsukatani, T., Costanzo, R.M., DiNardo, L.J., Reiter, E.R., 2001. Impact of olfactory impairment on quality of life and disability. Arch. Otolaryngol. Head Neck Surg. 127, 497–503. Naudin, M., El-Hage, W., Gomes, M., Gaillard, P., Belzung, C., Atanasova, B., 2012. State and trait olfactory markers of major depression. PloS One 7, e46938. Neumann, D., Zupan, B., Babbage, D.R., Radnovich, A.J., Tomita, M., Hammond, F., Willer, B., 2012. Affect recognition, empathy, and dysosmia after traumatic brain injury. Arch. Phys. Med. Rehab. 93, 1414–1420. Robin, M., Pham-Scottez, A., Curt, F., Dugre-Le Bigre, C., Speranza, M., Sapinho, D., Corcos, M., Berthoz, S., Kedia, G., 2012. Decreased sensitivity to facial emotions in adolescents with Borderline Personality Disorder. Psychiatry Res. 200, 417–421. Rotge, J.-Y., Lemogne, C., Hinfray, S., Huguet, P., Grynszpan, O., Tartour, E., George, N., Fossati, P., 2015. A meta-analysis of the anterior cingulate contribution to social pain. Soc. Cogn. Affect. Neurosci. 10, 19–27. Santos, D.V., Reiter, E.R., DiNardo, L.J., Costanzo, R.M., 2004. Hazardous events associated with impaired olfactory function. Arch. Otolaryngol. Head Neck Surg. 130, 317–319. Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., Hergueta, T., Baker, R., Dunbar, G.C., 1998. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J. Clin. Psychiatry 59 (Suppl. 20), S22–S33 ;quiz 34–57. Soudry, Y., Lemogne, C., Malinvaud, D., Consoli, S.-M., Bonfils, P., 2011. Olfactory system and emotion: common substrates. Eur. Ann. Otorhinolaryngol. Head Neck Dis. 128, 18–23. Spielberger, C.D., Bruchon-Schweitzer, M., Paulhan, I., 1993. Inventaire d’anxiété état-trait forme Y (STAI-Y). du Centre de psychologie appliquée, Paris, France. Stevenson, R.J., 2010. An initial evaluation of the functions of human olfaction. Chem. Senses 35, 3–20. Stevenson, R.J., Oaten, M.J., Case, T.I., Repacholi, B.M., Wagland, P., 2010. Children's response to adult disgust elicitors: development and acquisition. Dev. Psychol. 46, 165–177. Temmel, A.F.P., Quint, C., Schickinger-Fischer, B., Klimek, L., Stoller, E., Hummel, T., 2002. Characteristics of olfactory disorders in relation to major causes of olfactory loss. Arch. Otolaryngol. Head Neck Surg. 128, 635–641. Whalen, P.J., Kagan, J., Cook, R.G., Davis, F.C., Kim, H., Polis, S., McLaren, D.G., Somerville, L.H., McLean, A.A., Maxwell, J.S., Johnstone, T., 2004. Human amygdala responsivity to masked fearful eye whites. Science 306, 2061. Widen, S.C., Russell, J.A., 2013. Children's recognition of disgust in others. Psychol. Bull 139, 271–299.
Contents lists available at ScienceDirect
Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia
Congenital anosmia and emotion recognition: A case-control study Cédric Lemogne a,b,c,n, Julien Smadja a,b, El-Hadi Zerdazi b, Yaël Soudry b, Marion Robin d, Sylvie Berthoz d, Frédéric Limosin a,b,c, Silla M. Consoli a,b, Pierre Bonfils a,e,f a
Université Paris Descartes, Sorbonne Paris Cité, Faculté de médecine, 15 rue de l’Ecole de Médecine, 75015 Paris, France AP-HP, Hôpitaux Universitaires Paris Ouest, Service de Psychiatrie de l’adulte et du sujet âgé, 20 rue Leblanc, 75015 Paris, France c Inserm U894, Centre Psychiatrie et Neurosciences, 2 ter Rue d’Alésia, 75014 Paris, France d Inserm U1178 & Institut Mutualiste Montsouris, Universités Paris-Sud & Paris Descartes, 97 Bd Port Royal, 75014 Paris, France e AP-HP, Hôpitaux Universitaires Paris Ouest, Service d’ORL et de chirurgie cervico-faciale, 20 rue Leblanc, 75015 Paris, France f CNRS UMR MD 8257, Cognition and Action Group, Centre Universitaires des Saints-Pères, 45 rue des Saints Pères, 75006 Paris, France b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 February 2015 Received in revised form 15 April 2015 Accepted 24 April 2015 Available online 25 April 2015
Patients with anosmia are not able to detect volatile chemicals signaling the presence of infectious and non-infectious environmental hazards, which typically elicit disgust and fear, respectively. Social animals may compensate a loss of olfaction by taking advantage of signals of threat that are produced by their conspecifics. Among humans and other primates, body postures and facial expressions are powerful cues conveying emotional information, including fear and disgust. The aim of the present study was to examine whether humans with agenesis of the olfactory bulb, a rare disorder characterized by congenital anosmia, would be more accurate in recognizing facial expressions of fear and disgust. A total of 90 participants with no history of mental disorder or traumatic brain injury were recruited, including 17 patients with congenital anosmia (10 men, mean age7standard deviation: 36.5 7 14.8 years), 34 patients with acquired anosmia (18 men, mean age7 standard deviation: 57.27 11.8 years) and 39 healthy subjects (22 men, mean age7 standard deviation: 36.7 7 13.2 years). For each patient with congenital anosmia, the agenesis of the olfactory bulb was ascertained through magnetic resonance imaging. Emotion recognition abilities were examined with a dynamic paradigm in which a morphing technique allowed displaying emotional facial expressions increasing in intensity over time. Adjusting for age, education, depression and anxiety, patients with congenital anosmia required similar levels of intensity to correctly recognize fear and disgust than healthy subjects while they displayed decreased error rates for both fear (mean difference [95% confidence interval] ¼ 28.3% [ 46.3%, 10.2%], P ¼0.003) and disgust (mean difference [95% confidence interval] ¼ 15.8% [ 31.5%, 0.2%], P¼ 0.048). Furthermore, among patients with acquired anosmia, there was a negative correlation between duration of anosmia and the rate of errors for fearful (Spearman's ρ ¼ 0.531, P¼ 0.001) or disgust (Spearman's ρ ¼ 0.719, Po 0.001) faces recognition. No significant difference was observed for the other primary emotions. Overall, these results suggest that patients with congenital anosmia and long-lasting acquired anosmia may compensate their inability to detect environmental hazards through olfaction by an increased ability to detect fear or disgust as facially expressed by others. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Acquired anosmia Congenial anosmia Emotion Facial expressions Olfaction
1. Introduction Among several functions that olfaction deserves in mammals, a critical one is to drive behavioral avoidance from two kinds of environmental hazards signaled by volatile chemicals: non-infectious hazards, such as fire, predators or degraded air, and infectious hazards, such as feces, spoiled food or organic decay n Correspondence to: Unité de Psychologie et Psychiatrie de liaison et d’urgence, Hôpital Européen Georges Pompidou, 20 rue Leblanc, 75908 Paris Cedex 15, France. Fax: þ33 1 56 09 31 46. E-mail address: [email protected] (C. Lemogne).
http://dx.doi.org/10.1016/j.neuropsychologia.2015.04.028 0028-3932/& 2015 Elsevier Ltd. All rights reserved.
(Stevenson, 2010). Once detected, non-infectious hazards typically elicit fear, whereas infectious hazards typically elicit disgust (Davey, 2011). Associations between specific odors and avoidance are both innate and learned in mammals, especially rodents (Kobayakawa et al., 2007), but learned associations prevail in humans (Stevenson et al., 2010). Among humans, individuals with acquired anosmia are no longer able to detect several environmental threats such as burning food, smoke or gas leak (Bonfils et al., 2008; Miwa et al., 2001; Santos et al., 2004; Temmel et al., 2002). In addition, individuals with anosmia are also concerned with the fear of eliciting disgust in others through bodily odors (Miwa et al., 2001; Temmel et al., 2002). Avoiding environmental hazards and social
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
rejection were arguably major adaptive problems that our ancestors had to deal with (Cosmides and Tooby, 2013; Eisenberger, 2012). Emotion expression, including facial expression, is likely to have been promoted by natural selection to help solving adaptive problems. This idea can be traced back at least to 1872 in ‘The Expression of the Emotions in Man and Animals’ by Charles Darwin (Darwin, 1965). In the context of natural selection, the most limiting resource is not food or safety or access to mates, but the information required for making adaptive behavioral choices (Cosmides and Tooby, 2013). When it comes to identify environmental hazards, social animals not only rely on their own senses, including olfaction, but may also take advantage of signals of threat that are produced by their conspecifics. Among humans and other primates, body postures and facial expressions are powerful means to convey emotional information, including fear and disgust (de Gelder et al., 2004; Ekman et al., 1969). Discrete facial expressions associated with the six primary emotions (i.e., happiness, sadness, anger, surprise, fear and disgust) are readily recognized regardless of cultural differences (Ekman et al., 1969; Mesquita and Frijda, 1992) and can be produced by congenitally blind individuals (Matsumoto and Willingham, 2009). Beside explicit recognition, facial expressions of fear or disgust may also elicit analogous emotional states in the observer through activation of shared neural representations (Davey, 2011; Whalen et al., 2004). Overall, there is strong evidence that the recognition of emotions expressed by others might help humans to make adaptive behavioral choices (Cosmides and Tooby, 2013). Building on evidence linking anosmia with impaired ability to identify environmental hazards (Bonfils et al., 2008; Miwa et al., 2001; Santos et al., 2004; Temmel et al., 2002), we hypothesized that humans may compensate a lack of olfaction by taking advantage of facial expressions of fear or disgust that are produced by their conspecifics when they face these hazards. Moreover, humans demonstrate an implicit association between disease and smell (Bulsing et al., 2009) and ostracism may indeed partially derive from disgust-related behaviors leading to the avoidance of individuals that may be sick (Curtis, 2011) or prone to moral transgressions (Chapman and Anderson, 2013). Fear of rejection by others is a strong drive in humans (Rotge et al., 2015) and may lead to increased monitoring of other facial expressions of avoidance, including disgust. Therefore, we hypothesized that individuals with long-lasting anosmia would be more accurate in recognizing both fear and disgust. Our objective was to examine this hypothesis in humans with congenital anosmia, a rare disorder characterized by a lack of olfaction at birth. In the present study, we assessed the ability to recognize facial emotional expressions of individuals with normal olfaction, acquired anosmia and congenital anosmia due to agenesis of the olfactory bulb. Agenesis of the olfactory bulb results from abnormal embryological development of the olfactory system and may be isolated or associated with more complex malformation syndromes (e.g. Kallman syndrome). To our knowledge, the association of anosmia with the recognition of fear or disgust has never been studied apart from patients with traumatic brain injury (Neumann et al., 2012) or mental disorders (Kohler et al., 2007). Since participants in the present study were free of such conditions, only subtle differences were expected. Therefore, we used a dynamic paradigm in which a morphing technique allowed displaying emotional facial expressions increasing in intensity over time, from neutrality to a fully expressed emotion (Domes et al., 2008; Lynch et al., 2006; Robin et al., 2012). We hypothesized that individuals with congenital anosmia or long-lasting acquired anosmia, relative to healthy individuals, would correctly identify fear and disgust portrayed at lower levels of intensity (i.e., a faster recognition) and/or would be more accurate in recognizing
53
expressions of fear and disgust (i.e, a lower rate of errors).
2. Materials and methods 2.1. Participants Patients presenting with either congenital or acquired anosmia were recruited from the Head and Neck Surgery department of the European Georges Pompidou Hospital (Paris, France). Inclusion criteria were: being aged 18–74; mastery over French language. Exclusion criteria were: history of mental disorder or traumatic brain injury and current use of psychotropic medication. Eligible patients were proposed to participate after a medical consultation. Healthy subjects with no history of olfactory disease were recruited from the community to constitute a control group. For each participant, the entire experimental procedure was conducted by the same investigator (i.e., JS, EZ or YS). Each participant was given a short questionnaire covering socio-demographic and psychological data (i.e. depressive symptoms and anxiety). Participants were then screened for Axis I disorders of the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, with the Mini-International Neuropsychiatric Interview (Sheehan et al., 1998). Finally, they underwent the emotion recognition test and the olfactory tests. Type of anosmia and duration of the disease were collected from the clinical records. Participants' consent was obtained according to the Declaration of Helsinki and the study has been approved by the local ethical committee. 2.2. Measures of depressive symptoms and anxiety Since depression and anxiety may affect emotion recognition, all participants were given the 21-item Beck Depression Inventory (BDI-21) to assess depressive symptoms (Beck et al., 1988) and the Y-A form of the State-Trait Anxiety Inventory (STAI) to assess trait anxiety (Spielberger et al., 1993). The BDI-21 is a brief, reliable, and valid questionnaire of 21 items mainly related to the cognitive features of depressive mood such as pessimism, sense of failure, or guilt. For each item, subjects endorse one of the four statements that are proposed according to the perception of their current state. Each item is subsequently scored from 0 to 3 according to the endorsed sentence, yielding a total score ranging from 0 to 39 (cronbach's alpha ¼0.90 in the present study). The trait scale of the STAI includes 20 sentences describing various psychological symptoms of anxiety that are rated by the subjects from 1 (not endorsed) to 4 (fully endorsed) yielding a total score ranging from 20 to 80 (cronbach's alpha ¼0.91 in the present study). 2.3. Emotion recognition test Emotion recognition was assessed with a task that comprised 36 trials presented in a random order (Robin et al., 2012). Each trial began with a neutral face gradually morphed into one of the six primary emotions (i.e., happiness, sadness, anger, surprise, fear and disgust) using forty 2.5% incremental stages (Fig. 1). Each picture was presented for 500 ms followed immediately by the next morphed face in the sequence. Each trial therefore consisted of a 20 s continuum. We used the pictures of three men and three women, who were each expressing the six emotions of interest. Each target emotion was thus portrayed six times. Participants were told that they would see faces appearing on the computer screen, whose expressions would slowly and continuously evolve from neutral to one of the six target emotions. They were asked to watch the changes in facial expression and report the emotion expressed whenever they thought they had
54
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
Fig. 1. Stimulus example for the expression of fear.
identified it, by clicking with the mouse on one of the six corresponding boxes (located below the stimulus). They were asked to rather wait until they had recognized the expression than simply guessing. Participants were informed that they would not be provided with any information on the quality (correct or incorrect) of their response, and that the face would continue to change even after their response. They were also told that they could change their initial response at any time and as often as necessary by clicking again on one of the response buttons until the end of the trial. Participants were also instructed that, at the end of each trial (40th stage), they would have to indicate their final choice. A practice phase on each of the six emotions was first presented in order to familiarize the participant with the procedure. This practice phase involved two different actors – one man and one woman - than those used for the main task in order to avoid a familiarization bias with faces. For each emotion, three scores can be computed: (1) the final success rate (i.e., percentage of correct responses at the 40th stage), (2) the number of stages required for the accurate identification of the emotion (ranging from 1 to 40), and (3) the percentage of first responses that were incorrect. 2.4. Olfactory tests In addition to clinical records for patients and self-report for controls, olfaction was assessed with either the Biolfa test (Lecanu et al., 2002) or the Sniffin Test (Hummel et al., 1997), depending upon the period of recruitment. Only quantitative measures of olfactory thresholds will be reported here. The Biolfa test measures the olfactory threshold for three substances (i.e., 2-phenylethanol, eugenol and aldehyde C14) and yields 3 scores ranging from 1 to 9. The Sniffin test measures the olfactory threshold for either the 2-phenylethanol or the n-butanol and yields a score ranging from 1 to 16, with greater scores indicating better olfaction. Here, only the n-butanol was used. For both tests, a score of 1 was given to the participants who failed to detect the substance at the greatest concentration. 2.5. Statistical analyses Differences between the three groups regarding continuous and discrete variables were examined with one-way ANOVAs and χ² tests, respectively. To test our a priori hypotheses, we first computed 3 2 ANCOVAs with a three-level group factor (i.e., patients with congenital anosmia, acquired anosmia, healthy subjects) and a two-level emotion factor (i.e., fear and disgust versus happiness, sadness, anger and surprise), entering age, education, depression and anxiety as covariates. Thus the model included the group and emotion main effects, the group emotion interaction and the interactions of the emotion factor with each covariate. Should a significant group emotion interaction be found, post hoc pairwise comparisons across groups for each emotion type (i.e., fear and disgust versus happiness, sadness, anger and surprise) were based on estimated means computed
with one-way ANCOVAS with a three-level group factor and the same covariates. Exploratory analyses followed regarding each of the six primary emotions, separately. Since we had strong a priori hypotheses regarding the recognition of fear and disgust, specifically, significance level was set at Po 0.05. For each significant difference, mean difference and 95% confidence interval (95% CI) were computed. All the analyses were conducted with the PASW Statistics 18 software (Chicago: SPSS Inc.).
3. Results 3.1. Participants A total of 90 participants were recruited, including 17 patients with congenital anosmia, 34 patients with acquired anosmia and 39 healthy subjects. For each patient with congenital anosmia, the agenesis of the olfactory bulb was ascertained through magnetic resonance imaging (Fig. 2). The olfactory bulb agenesis was isolated for 13 patients and associated with a Kallman syndrome for 4 patients. Acquired anosmia resulted from post-viral anosmia (N ¼15), nasal polyposis (N ¼ 9), head trauma without brain injury (N ¼3) and undetermined causes (i.e., idiopathic anosmia) (N ¼7). The characteristics of the three groups are displayed in Table 1. The three groups differed regarding age, education and depression. There was also a trend regarding differences in anxiety (P ¼0.062). Therefore, subsequent analyses were adjusted for age, education, depression and anxiety by entering these variables as covariates. Because of limited availability of the olfactory tests, only a subsample of the healthy subjects (N ¼15) were tested for olfaction. None of the healthy subjects had olfactory complaint and those who were tested had normal olfaction. 3.2. Emotion recognition There was no significant group main effect or group emotion interaction regarding the rate of final correct recognition or the stage for correct recognition (all P Z0.248). However, there was a significant group main effect regarding the error rate at first response (F¼3.235; df ¼2, 83; P ¼0.044; η² ¼0.072), which was driven by a significant group emotion interaction (F¼ 5.088; df¼ 2, 83; P¼ 0.008; η² ¼0.109). The other interactions were not significant (all PZ 0.171). In line with our hypotheses, post hoc pairwise comparisons across groups based on estimated means found that patients with congenital anosmia displayed lower error rates at first response than healthy subjects for fear and disgust (Mean difference [95% CI] ¼ 21.8% [ 35.9%, 7.7%]; t¼3.077; df¼83; P ¼0.003) with no other significant between-group difference (all P Z0.067). There was no significant between-group difference regarding the other four primary emotions (all PZ 0.290). Likewise, exploratory analyses regarding each primary emotion separately found that patients with congenital anosmia displayed lower error rates at first response than healthy subjects for both fear (Mean difference [95% CI] ¼ 28.3% [ 46.3%,
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
55
Fig. 2. MRI coronal view of the olfactory tract of a healthy subject (A) and of a patient with congenital anosmia (B).
10.2%]; t ¼3.111; df¼83; P¼ 0.003) and disgust (Mean difference [95% CI] ¼ 15.8% [ 31.5%, 0.2%]; t ¼2.013; df¼ 83; P ¼0.048) (Table 2, Fig. 3). In contrast, there was no significant betweengroup difference regarding the other four primary emotions (Table 2). There was no interaction between group and emotion type when considering fear versus disgust only (P ¼0.376), suggesting that between-group differences were similar for these two emotions. There was no indication of a greater tendency for patients with congenital anosmia, compared to healthy subjects, to falsely recognize fear (P ¼0.681) or disgust (P ¼0.344) when presented with other emotions. Contrary to patients with congenital anosmia, those with acquired anosmia did not display lower error rate at first response than healthy subjects for fear (P ¼0.259) or disgust (P ¼0.971)
(Table 2, Fig. 3). However, post hoc analyses among patients with acquired anosmia found a negative correlation between the duration of anosmia and the error rate at first response for both fear (Spearman’s ρ ¼ 0.531, P ¼0.001) and disgust (Spearman's ρ ¼ 0.719, P o0.001) (Fig. 4). In other words, the longer the duration of anosmia was, the lower the rate of errors for fear and disgust. This issue was further explored using between-group comparisons (i.e., long-lasting versus recent acquired anosmia) based on a median split of the duration of anosmia (i.e. 3 years). Comparing the two groups of patients with acquired anosmia with healthy controls, there was a significant group main effect regarding the error rate at first response (F¼ 8.409; df¼2, 66; P¼ 0.001; η² ¼0.203), which was driven by a significant group emotion interaction (F ¼4.032; df¼ 2, 66; P¼ 0.022; η²¼ 0.109).
Table 1 Characteristics of the participants. Healthy subjects (N ¼39)
Patients with congenital anosmia (N¼ 17)
Patients with acquired anosmia (N¼ 34)
Χ²
P
Discrete variables Male gender
N 22
% 56.4
N 10
% 58.8
N 18
% 52.9
0.179
0.914
Continuous variables Age Education (years) STAI-Trait BDI-21 Anosmia duration (years) Biolfa Testa 2-Phenylethanol (1–9) Eugenol (1–-9) Aldehyde C14 (1–9) Sniffin Testb n-Butanol (1–16)
Mean 36.7 16.8 35.7 5.2
SD 13.2 3.4 9.3 6.0
Mean 36.5 14.5 42.8 10.3
SD 14.8 3.0 9.4 8.6
Mean 57.2 15.1 38.6 8.8 5.5
SD 11.8 4.0 11.6 8.7 6.2
F 26.562 3.267 2.880 3.354
P o 0.001 0.043 0.062 0.040
7.3 7.7 7.5
0.9 1.3 1.5
1.6 1.4 1.4
1.5 1.5 1.3
1.5 1.6 1.5
1.3 1.8 1.5
111.54 82.485 93.876
o 0.001 o 0.001 o 0.001
1
1.4
0.8
a b
The Biolfa Test was used for 15 healthy subjects, 16 patients with congenital anosmia and 22 patients with acquired anosmia. The Sniffin Test was used for one patient with congenital anosmia and 12 patients with acquired anosmia.
56
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
The other interactions were not significant (all P Z0.200). Patients with long-lasting acquired anosmia displayed lower error rates for fear and disgust than both healthy subjects (Mean difference [95% CI] ¼ 22.5% [ 37.9%, 7.1%]; t¼ 2.913; df ¼66; P¼ 0.005) and patients with more recent acquired anosmia (Mean difference [95% CI]¼ 30.4% [ 46.1%, 14.6%]; t¼ 3.845; df ¼66; P o0.001). Table 2 Emotion recognition performances. Healthy subject Patients with con(N ¼ 39) genital anosmia (N ¼ 17)
Patients with acquired anosmia (N¼ 34)
Mean
SD
Mean
SD
Mean
SD
Correct final recognition rate (%) Happiness Sadness Anger Surprise Fear Disgust
100.0 83.8 80.3 87.6 85.5 76.1
0.0 17.7 17.9 15.2 21.0 16.1
100.0 86.3 77.5 81.4 85.3 75.5
0.0 16.9 22.0 23.5 14.3 14.6
98.0 86.7 78.9 77.4 73.5 78.9
9.0 15.8 16.0 21.3 20.5 18.0
Stage of correct recognition (1–40) Happiness Sadness Anger Surprise Fear Disgust
19.2 27.1 28.6 24.4 29.3 26.6
3.7 4.1 3.5 4.6 3.6 4.9
21.4 29.0 30.0 25.3 30.9 27.3
5.7 3.8 4.1 4.5 3.6 4.3
21.9 28.6 29.6 27.2 31.1 29.0
5.6 4.3 4.3 4.6 3.6 4.4
Error rate at first response (%) Happiness Sadness Anger Surprise Fear Disgust
2.6 11.9 12.1 9.8 33.5a 21.4b
6.2 23.0 18.6 14.7 29.4 23.7
2.0 7.8 11.4 6.6 10.4a 11.7b
5.5 18.6 21.8 13.9 18.0 24.2
7.2 13.6 19.5 12.2 34.1 24.5
14.1 23.8 26.9 23.4 35.3 29.0
a P ¼0.003 for the comparison of patients with congenital anosmia versus healthy controls, adjusting for age, education, depression and anxiety in general linear model. b P¼0.048 for the comparison of patients with congenital anosmia versus healthy controls adjusting for age, education, depression and anxiety in general linear model.
4. Discussion The aim of the present study was to examine whether individuals with congenital or long-lasting acquired anosmia may display enhanced ability to recognize facial expressions of fear and disgust. We found mixed results in favor of this hypothesis. Indeed, patients with congenital anosmia presented no advantage for detecting fear and disgust at lower levels of emotional intensity than healthy subjects while they displayed a decreased rate of errors for these two emotions. Furthermore, these results seem to be specific to fear and disgust, as suggested by a significant interaction between group and emotion type (i.e. fear and disgust versus the other four primary emotions). Finally, as regards patients with acquired anosmia, we found a negative correlation between duration of anosmia and the error rates for the recognition of fear and disgust. In other words, the longer the illness duration was, the greater the accuracy for recognizing fear and disgust. Overall, these results suggest that patients with congenital anosmia and long-lasting acquired anosmia may compensate their inability to detect environmental hazards through olfaction by increasing their ability to detect fear or disgust as expressed by others. One may argue that these results could be explained by either a more careful appraisal of the stimuli by patients with congenital anosmia, or by a greater tendency to consider facial expressions as expressing fear or disgust, regardless of the actual emotion that is portrayed. Nevertheless, patients with congenital anosmia did not display longer delays for correctly identifying fear and disgust than healthy subjects, providing no support for the first hypothesis. In addition, patients with congenital anosmia were not more likely to be biased towards misrecognition of fear or disgust than healthy subjects when presented with other emotions, ruling out the second hypothesis as well. Of particular note, the association between congenital anosmia and the recognition of fear appeared to be somewhat stronger than with disgust, though there was no significant interaction between group and emotion type when considering only fear and disgust. Apart from phenomenological aspects, several features distinguish fear from disgust on a physiological level, including facial mimicry and their neural bases. For instance, facial expressions of fear activates the amygdala, but not the insula, whereas facial expressions of disgust activates the insula, but not the amygdala (Fusar-Poli et al., 2009). More specific to the present
Fig. 3. Mean error rate at first response and upper limit of the 95% confidence interval across groups and emotions. (a) P¼ 0.003 for the comparison of patients with congenital anosmia versus healthy controls, adjusting for age, education, depression and anxiety in general linear model. (b) P¼ 0.048 for the comparison of patients with congenital anosmia versus healthy controls adjusting for age, education, depression and anxiety in general linear model.
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
57
Fig. 4. Error rate at first response for fear and disgust as a function of duration of anosmia among patients with acquired anosmia. Fit lines were drawn with locally weighted scatterplot smoothing based on an Epanechnikov kernel.
study is the fact that the recognition of facial expression of disgust occurs later in development than the recognition of fear. Although 5-year-old children are able to produce facial and verbal expressions of disgust and to infer disgust in others from a situation or a behavior, the ability to infer disgust in others from facial mimicry only gradually improves until the late teens (Widen and Russell, 2013). In contrast, 5-year old children are able to readily recognize fear in others, even from partial facial expressions (Gagnon et al., 2014). In other words, the association between odors and disgust might be learned several years before the association between facial expressions of disgust in others and this emotion. Should the first association facilitate the learning of the second one (e.g., by perceiving disgusting odors and facial expressions of disgust in others, simultaneously), the absence of olfaction at birth may impede the development of disgust recognition among patients with congenital anosmia. This might account for the somewhat smaller effect observed with disgust. Strength of the present study includes its novelty, the use of a sensitive protocol of emotion recognition and the recruitment of patients with congenital anosmia, a rare disorder characterized by the absence of olfaction at birth. These patients offer the unique opportunity to examine the association between a complete lack of olfaction and the ability to recognize facial expressions, particularly of fear and disgust. In addition, the participants were carefully screened for potential confounding effects, such as a history of mental disorders, with a standardized interview and assessed for subclinical levels of depression and anxiety with validated measures. Previous studies that addressed the association of anosmia with the recognition of facial emotional expressions included patients with traumatic brain injury (Neumann et al., 2012) or mental disorders (Kohler et al., 2007). Given that olfaction and emotion share some neural substrates (Soudry et al., 2011), any brain disorder may confound their association. Abnormal olfaction has indeed been proposed as a marker of psychopathology (Atanasova et al., 2008; Naudin et al., 2012) and traumatic brain injury may be associated with impaired recognition of facial expressions of emotion (Biszak and Babbage, 2014). Finally, psychotropic medications also influence emotion recognition abilities (Harmer, 2013). It is thus noteworthy that the present results were observed in participants free of brain disorders and psychotropic medications, and adjusted for subclinical levels of depression and anxiety. Some limitations should be noted. First, one may argue that differences regarding fear and disgust resulted from chance because of multiple comparisons. However, these results were a priori expected as exposed in the introduction. In addition, there was no significant difference as regards the other four primary
emotions (i.e. happiness, sadness, anger and surprise), whatever the measure. Furthermore, the interaction between group and emotion type (i.e., fear and disgust versus happiness, sadness, anger and surprise) was significant, suggesting that the results may indeed be specific to fear and disgust. However, we cannot rule out that potential ceiling effects might account for null findings regarding the other four primary emotions. Further studies are warranted to confirm the specificity of our results regarding fear and disgust. Increased attention to the facial expression of others to detect fear and disgust may indeed have resulted in better emotion recognition regardless of emotion type. Second, our experimental design did not allow testing the implicit recognition of emotions. Future studies might use subliminal stimuli and physiological measures to address this issue. Finally, dynamic facial recognition paradigms might be viewed as more ecological than static paradigms, but duration of 20 s is very far from the duration of expression of emotion by humans and constitutes a limit of morphing-based studies. In conclusion, we used a sensitive measure of emotion recognition in a population of patients with congenital and acquired anosmia to test the role of olfaction in the recognition of facial expressions of fear and disgust. Our results suggest that individuals with long-standing anosmia, either congenital or acquired, may develop enhanced ability to recognize these two emotions, possibly to compensate their inabilities to detect olfactory signals of both infectious and non-infectious environmental hazards.
Conflict of interests None of the authors have any conflict of interests to report.
Role of funding sources None.
Acknowledgments The authors thank Patrick Faulcon for his help in recruiting the participants and Laurence Laccourreye for having reviewed the manuscript.
58
C. Lemogne et al. / Neuropsychologia 72 (2015) 52–58
References Atanasova, B., Graux, J., El Hage, W., Hommet, C., Camus, V., Belzung, C., 2008. Olfaction: a potential cognitive marker of psychiatric disorders. Neurosci. Biobehav. Rev. 32, 1315–1325. Beck, A., Steer, R., Garbin, M., 1988. Psychometric properties of the Beck depression inventory-25 years of evaluation. Clin. Psychol. Rev. 8, 77–100. Biszak, A.M., Babbage, D.R., 2014. Facial affect recognition difficulties in traumatic brain injury rehabilitation services. Brain Injury 28, 97–104. Bonfils, P., Faulcon, P., Tavernier, L., Bonfils, N.A., Malinvaud, D., 2009. Home Accidents Associated With Anosmia. vol. 37. Presse Médicale, Paris, France, pp. 742–745. Bulsing, P.J., Smeets, M.A.M., Van den Hout, M.A., 2009. The implicit association between odors and illness. Chem. Senses 34, 111–119. Chapman, H.A., Anderson, A.K., 2013. Things rank and gross in nature: a review and synthesis of moral disgust. Psychol. Bull. 139, 300–327. Cosmides, L., Tooby, J., 2013. Evolutionary psychology: new perspectives on cognition and motivation. Annu. Rev. Psychol. 64, 201–229. Curtis, V., 2011. Why disgust matters. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 366, 3478–3490. Darwin, C., 1965. The Expression of the Emotions in Man and Animals. University of Chicago Press. Davey, G.C.L., 2011. Disgust: the disease-avoidance emotion and its dysfunctions. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 366, 3453–3465. De Gelder, B., Snyder, J., Greve, D., Gerard, G., Hadjikhani, N., 2004. Fear fosters flight: a mechanism for fear contagion when perceiving emotion expressed by a whole body. Proc. Natl. Acad. Sci. USA 101, 16701–16706. Domes, G., Czieschnek, D., Weidler, F., Berger, C., Fast, K., Herpertz, S.C., 2008. Recognition of facial affect in Borderline Personality Disorder. J. Pers. Disord 22, 135–147. Eisenberger, N.I., 2012. The neural bases of social pain: evidence for shared representations with physical pain. Psychosom. Med. 74, 126–135. Ekman, P., Sorenson, E.R., Friesen, W.V., 1969. Pan-cultural elements in facial displays of emotion. Science 164, 86–88. Fusar-Poli, P., Placentino, A., Carletti, F., Landi, P., Allen, P., Surguladze, S., Benedetti, F., Abbamonte, M., Gasparotti, R., Barale, F., Perez, J., McGuire, P., Politi, P., 2009. Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional magnetic resonance imaging studies. J. Psychiatry Neurosci. 34, 418–432. Gagnon, M., Gosselin, P., Maassarani, R., 2014. Children's ability to recognize emotions from partial and complete facial expressions. J. Genet. Psychol. 175, 416–430. Harmer, C.J., 2013. Emotional processing and antidepressant action. Curr. Top. Behav. Neurosci. 14, 209–222. Hummel, T., Sekinger, B., Wolf, S.R., Pauli, E., Kobal, G., 1997. “Sniffin” sticks': olfactory performance assessed by the combined testing of odor identification, odor discrimination and olfactory threshold. Chem. Senses 22, 39–52. Kobayakawa, K., Kobayakawa, R., Matsumoto, H., Oka, Y., Imai, T., Ikawa, M., Okabe, M., Ikeda, T., Itohara, S., Kikusui, T., Mori, K., Sakano, H., 2007. Innate versus learned odour processing in the mouse olfactory bulb. Nature 450, 503–508. Kohler, C.G., Barrett, F.S., Gur, R.C., Turetsky, B.I., Moberg, P.J., 2007. Association between facial emotion recognition and odor identification in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 19, 128–131.
Lecanu, J.B., Faulcon, P., Werner, A., Bonfils, P., 2002. [Normative data of the Biolfa (R) olfactory test]. Ann. Oto-Laryngol. Chir. Cervico Faciale Bull. Soc. Oto-Laryngol. Hôp. Paris 119, 164–169. Lynch, T.R., Rosenthal, M.Z., Kosson, D.S., Cheavens, J.S., Lejuez, C.W., Blair, R.J.R., 2006. Heightened sensitivity to facial expressions of emotion in borderline personality disorder. Emot. Wash. DC 6, 647–655. Matsumoto, D., Willingham, B., 2009. Spontaneous facial expressions of emotion of congenitally and noncongenitally blind individuals. J. Pers. Soc. Psychol. 96, 1–10. Mesquita, B., Frijda, N.H., 1992. Cultural variations in emotions: a review. Psychol. Bull 112, 179–204. Miwa, T., Furukawa, M., Tsukatani, T., Costanzo, R.M., DiNardo, L.J., Reiter, E.R., 2001. Impact of olfactory impairment on quality of life and disability. Arch. Otolaryngol. Head Neck Surg. 127, 497–503. Naudin, M., El-Hage, W., Gomes, M., Gaillard, P., Belzung, C., Atanasova, B., 2012. State and trait olfactory markers of major depression. PloS One 7, e46938. Neumann, D., Zupan, B., Babbage, D.R., Radnovich, A.J., Tomita, M., Hammond, F., Willer, B., 2012. Affect recognition, empathy, and dysosmia after traumatic brain injury. Arch. Phys. Med. Rehab. 93, 1414–1420. Robin, M., Pham-Scottez, A., Curt, F., Dugre-Le Bigre, C., Speranza, M., Sapinho, D., Corcos, M., Berthoz, S., Kedia, G., 2012. Decreased sensitivity to facial emotions in adolescents with Borderline Personality Disorder. Psychiatry Res. 200, 417–421. Rotge, J.-Y., Lemogne, C., Hinfray, S., Huguet, P., Grynszpan, O., Tartour, E., George, N., Fossati, P., 2015. A meta-analysis of the anterior cingulate contribution to social pain. Soc. Cogn. Affect. Neurosci. 10, 19–27. Santos, D.V., Reiter, E.R., DiNardo, L.J., Costanzo, R.M., 2004. Hazardous events associated with impaired olfactory function. Arch. Otolaryngol. Head Neck Surg. 130, 317–319. Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., Hergueta, T., Baker, R., Dunbar, G.C., 1998. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J. Clin. Psychiatry 59 (Suppl. 20), S22–S33 ;quiz 34–57. Soudry, Y., Lemogne, C., Malinvaud, D., Consoli, S.-M., Bonfils, P., 2011. Olfactory system and emotion: common substrates. Eur. Ann. Otorhinolaryngol. Head Neck Dis. 128, 18–23. Spielberger, C.D., Bruchon-Schweitzer, M., Paulhan, I., 1993. Inventaire d’anxiété état-trait forme Y (STAI-Y). du Centre de psychologie appliquée, Paris, France. Stevenson, R.J., 2010. An initial evaluation of the functions of human olfaction. Chem. Senses 35, 3–20. Stevenson, R.J., Oaten, M.J., Case, T.I., Repacholi, B.M., Wagland, P., 2010. Children's response to adult disgust elicitors: development and acquisition. Dev. Psychol. 46, 165–177. Temmel, A.F.P., Quint, C., Schickinger-Fischer, B., Klimek, L., Stoller, E., Hummel, T., 2002. Characteristics of olfactory disorders in relation to major causes of olfactory loss. Arch. Otolaryngol. Head Neck Surg. 128, 635–641. Whalen, P.J., Kagan, J., Cook, R.G., Davis, F.C., Kim, H., Polis, S., McLaren, D.G., Somerville, L.H., McLean, A.A., Maxwell, J.S., Johnstone, T., 2004. Human amygdala responsivity to masked fearful eye whites. Science 306, 2061. Widen, S.C., Russell, J.A., 2013. Children's recognition of disgust in others. Psychol. Bull 139, 271–299.
