- Email: [email protected]
Facial emotion recognition in childhood: The effects of febrile seizures in the developing brain
Epilepsy & Behavior 29 (2013) 211–216
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Facial emotion recognition in childhood: The effects of febrile seizures in the developing brain Gaetano Cantalupo a,b,⁎,1, Stefano Meletti c,1, Alessia Miduri a, Silvia Mazzotta a, Loreto Rios-Pohl d, Francesca Benuzzi c, Francesco Pisani a, Carlo Alberto Tassinari e,f, Giuseppe Cossu a,f a
Child Neuropsychiatry Unit, Department of Neuroscience, University-Hospital of Parma, Via Gramsci 14, Parma, Italy Department of Life and Reproduction Sciences, University of Verona, P.le Scuro 10, Verona, Italy Department of Biomedical Sciences, Metabolism and Neuroscience, University of Modena and Reggio Emilia, Nuovo Ospedale Civile Sant'Agostino Estense, Via Giardini 1355, Modena, Italy d Centro Avanzado de Epilepsia, Clinica Las Condes, Lo Fontecilla 441, Las Condes, Santiago, Chile e Department of Neurological Sciences, University of Bologna, Via Foscolo 7, Bologna, Italy f Department of Neuroscience, University of Parma, Via Volturno 39/E, Parma, Italy b c
a r t i c l e
i n f o
Article history: Received 11 February 2013 Revised 4 July 2013 Accepted 7 July 2013 Available online 27 August 2013 Keywords: Emotions Facial expression Emotional intelligence Medial temporal lobe Limbic system Amygdala Febrile convulsion Epilepsy Temporal lobe epilepsy Pediatric neurology
a b s t r a c t It has been documented that anteromedial temporal lobe dysfunction can cause impairment in emotional intelligence. In particular, medial temporal lobe epilepsy (MTLE) is associated with disorders in emotion recognition from facial expressions. About one-third of patients with MTLE experienced febrile seizures (FSs) during childhood. In the present study, we investigated facial emotion recognition ability in a group of 38 school-aged children with antecedent FSs and in an age- and sex-matched control group. Children with abnormal general visuoperceptual abilities were excluded. Children with FSs showed lower recognition scores versus controls in both matching (28.64 vs 33.47; p b .0001) and labeling (21.25 vs 23.03; p = .001) facial emotions. Our findings support the hypothesis that FSs can be associated during childhood with a dysfunction within the neural network subserving the processing of facial expressions of the basic emotions. © 2013 Elsevier Inc. All rights reserved.
1. Introduction Febrile seizures (FSs) are the most common seizure disorder in childhood associated with fever but without evidence of intracranial infection or defined causes [1]. They occur in 2–5% of children younger than 5 years of age, and based on duration and clinical features, FSs are subdivided into ‘simple’ and ‘complex’ [1,2]. Epidemiological studies have substantially improved our understanding of the frequency, natural history, and recurrence of FSs, and there is now a general agreement that simple FSs are benign and rarely require chronic anticonvulsant treatment. Recent clinical and genetic studies suggest that the relationship between FSs and the development of epilepsy later in life is frequently genetic, and there are a number of genes that are either involved in susceptibility for FSs [3] or are pathogenic in epilepsies that present with FSs in their course [4,5]. However, the impact of FSs on ⁎ Corresponding author at: Child Neuropsychiatry Unit, Department of Neuroscience, University-Hospital of Parma, Via Gramsci 14, 43126, Parma, Italy. Fax: +39 0521704708. E-mail address: [email protected] (G. Cantalupo). 1 The first two authors contributed equally to this manuscript. 1525-5050/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.yebeh.2013.07.007
the developing brain has not been completely understood. Since FSs are the most common seizures in children, it is important to delineate whether early-life FSs can alter long-term neuroplasticity [6], especially the cognitive functions. In particular, it is important to evaluate whether FSs could be linked to persistent subtle brain injuries and the related cognitive dysfunctions. Regarding anatomical injury, longitudinal MRI studies have documented that prolonged and focal FSs can occasionally produce acute hippocampal injury that evolves into atrophy [7–9]. On the other hand, it has been proposed that a preexisting developmental hippocampal abnormality may predispose individuals to having a prolonged febrile seizure [9,10]. Indeed, the association between hippocampal sclerosis and FSs raises the possibility of selective cognitive sequelae involving the cognitive functions subserved by the medial temporal lobe region. In early reports from hospital-based studies, the prognosis of FSs, in terms of cognitive outcome, was fairly pessimistic (8% to 22% of cases presented mental retardation) because of the inclusion of symptomatic causes of seizures other than fever [11]. Furthermore, the hospital-based studies may be skewed toward disproportionately severe cases [12]. In contrast, long-term population-based cohort studies showed that children with FSs have similar global neurocognitive
212
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
developmental and academic performance compared with controls [13–16]. Therefore, the current viewpoint is that the vast majority of children with FSs have normal global measures of cognition and behavior. However, previous studies on the outcome of children with FSs only rarely focused on specific cognitive functions that can reveal a FS-related hippocampal injury, such as memory [17]. Neural networks underlying facial emotion recognition involve a distributed set of structures that include the visual cortices, the amygdala, the orbitofrontal cortex, and additional cerebral regions, such as the insula, the basal ganglia, and the prefrontal cortex [18]. The amygdala, which is often damaged along with the hippocampus in patients with medial temporal lobe epilepsy (MTLE) [19], has been identified as a key structure for evaluating emotional stimuli and regulating social and emotional behavior [20–22]. Accordingly, the investigation of emotional and social competences in patients with MTLE has been the focus of different studies [23–32] documenting common and widespread deficits of emotion recognition in this seizure disorder. In particular, patients with early onset of seizures, including FSs, and bilateral medial temporal lobe damage were severely impaired in emotion recognition [23,30,33]. The recognition of emotional signals, from all sensory modalities, is a critical component of human social interactions because it is through the understanding of the affective states of others that we modulate our behavioral responses [34]. Notably, facial expressions provide the greatest amount of emotional cues for recognizing emotions, both with positive and negative values [35]. The expression of basic emotions is supposed to be universal, innate, and have a specific neural substrate; on the contrary, the recognition of social emotions is learned over a lifetime. Thus, the alteration of emotional and social competences could represent another index of the potential FS-related medial temporal lobe injury. To the best of our knowledge, no study has investigated emotional–social abilities in children with antecedent FSs. In the present study, we assessed the ability of school-aged children with a history of FSs to recognize the facial expressions of basic innate emotions (happiness, sadness, fear, anger, and disgust) [36] in order to explore the possibility that FSs are associated with (or can lead to) a dysfunction within the neural network subserving the processing of facial expressions of basic emotions. Accordingly, we investigated facial emotion recognition (FER) ability in a sample of healthy school-aged children with antecedent FSs and compared their performances with an age- and sex-matched control group. 2. Material and methods 2.1. Participants We reviewed the database from the Child Neuropsychiatry Unit of the University-Hospital of Parma (Italy) to recruit children referred to our service between 2000 and 2008 for FSs, who were school-aged at the time of the enrollment. Following the International League Against Epilepsy [37], a FS was defined as a seizure occurring between 1 month and 5 years of age, associated with a febrile illness but not caused by an infection of the CNS, without previous neonatal seizures or a previous unprovoked seizure, and not meeting the criteria for other acute symptomatic seizures. Accordingly, we selected only cases with a complete diagnostic workup, including detailed eyewitness accounts of the episodes, clinical course, neurological evaluation, and an EEG recording within 48 h after the FS. Out of 163 eligible subjects aged 6–10 years, with at least one episode of FS and without subsequent epilepsy, 150 could be contacted, and 25.3% of the children's parents accepted to participate. Thus, 38 children (18 females, 20 males) were recruited as the clinical group (FS group). A group of 38 age- and sex-matched children (18 females, 20 males), attending primary school from the same district, served as the control group (CTRL). All participants' parents gave their informed consent, and the study was approved by the local ethical committee. Demographic and clinical variables were collected through a semistructured
interview of parents to explore the children's handedness, educational level, and familial and personal history. A history of neurological and/ or neuropsychiatric antecedents was an exclusion criterion for children to be enrolled in the control group. The presence of EEG abnormalities, for the FS group, was evaluated. Demographic variables of the two groups are reported in Table 1; neither the age at the time of testing (ANOVA) nor gender and handedness distributions (chi-square test) were significantly different between groups. For the FS group, the characteristics of FSs were reported, in order to be classified as ‘simple’ or ‘complex’ [2,38,39], along with the age at the first episode and the number of episodes (see Table 1). Twenty-one out of 38 (55.3%) children experienced a single episode, more than 80% (31/38) of episodes were classified as simple, and only seven patients experienced at least one episode classified as complex, being focal (4/38) and/or prolonged (4/38) and/or drug-interrupted (2/38). Half of our FS group (19/38) experienced only a single, simple, and short FS. 2.2. Neuropsychological assessment The children's IQ was evaluated by means of Raven's colored progressive matrices (CPM) [40]. The experimental protocol included three tasks built ad hoc with Ekman and Friesen's Pictures of Facial Affect [41] depicting five emotional expressions (happiness, sadness, fear, disgust, and anger) or no emotion (neutral). Ekman and Friesen's Pictures of Facial Affect represent a validated and reliable set of posed emotional expressions that generate high recognition rates across cultures; accordingly, they have been used in a wide range of studies, including developmental, cross-cultural, neuroimaging, and behavioral studies [42,43]. These black and white pictures were initially developed and used with adults, but they were later fruitfully implemented in studies with children [42]. Facial stimuli expressing surprise were not included because the intrinsic ambiguity of this emotion led some authors to consider it not among the basic emotions [43] or to consider it a ‘gradation’ of fear [44], and even healthy subjects often confuse fear for surprise [45]. The original stimuli were digitally edited in order to remove all nonface cues (i.e., clothing, neck, and hair). 2.2.1. Face Identity Matching Test Neutral stimuli from 10 subjects (five males) were used to create the Face Identity Matching Test (FIMT), aimed to control for subjects' basic visuoperceptual ability with facial stimuli. In this task, the children were required to identify the target face, portrayed in the upper half of the page, among five alternatives (one target and four confounders) depicted below (Fig. s-1A, online supplementary material). 2.2.2. Facial Expression Matching Test The second task, the Facial Expression Matching Test (FEMT), was created to assess FER ability, and it did not require any verbal expertise. This test comprised 40 emotional items in which the probe facial expressions resulted in a high rate of correct identification in the adult data [41]. All stimuli were faces of four women (C, MF, MO, and MR in the Ekman and Friesen's series) and four men (EM, GS, JJ, and WF) depicting the five basic emotions. For each item, the subjects were requested to match the probe facial expression with one of the five alternatives aligned below. The targets and confounders were expressions of the same actor, while the probe expression belonged to a different actor of the same gender (Fig. s-1B). The probe emotion and gender, as well as the position of the target among the confounders, varied across subsequent items in a pseudorandom sequence in order to minimize repetitions (see Table s-1, online supplementary material). 2.2.3. Facial Expression Labeling Test This test was already used in previous studies from our group to quantify the FER ability of adult subjects [23,30,31]. Briefly, in this task for each trial, the subject was presented with one original stimulus from Ekman and Friesen's series (including nonfacial features) and
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
213
Table 1 Demographic data and clinical features of FS.
Age at testing Gender n (percentage) Handedness n (percentage) Age at first FS episode Complex FSs n (percentage) Prolonged FSs n (percentage) Number of FS episodes n (percentage)
Mean (±SD) Range Female Male Right Left Mean (±SD) Range
CTRL (n = 38)
FS (n = 38)
Statistical analysis
8.2 (±1.4) 6–10 18 (47.4%) 20 (52.6%) 33 (86.8%) 5 (13.2%)
7.7 (±1.4) 6–10 18 (47.4%) 20 (52.6%) 35 (92.1%) 3 (7.9%) 2.42 (±1.2) 1–5 7 (18.4%)
F = 2.508; p = .118 (ANOVA)
Only N15 min Including drug-interrupted #1 #2 #3 #4 #N4
χ2 = 1.000; p = 1.000 (FET) χ2 = 0.455; p = .711 (FET)
4 (10.5%) 6 (15.8%) 21 (55.3%) 9 (23.7%) 1 (2.6%) 2 (5.3%) 5 (12.9%)
SD= Standard Deviation; ANOVA= analysis of variance; FET= Fisher's Exact Test.
was required to choose one of the five verbal labels (happiness, sadness, fear, disgust, and anger) printed below the target face (and read aloud by the experimenter, if required) that best described the emotion shown (Fig. s-1C). This Facial Expression Labeling Test (FELT) consisted of 25 trials (five face stimuli for each emotion). The participants were instructed to consider carefully all five alternatives before answering. There was no time limit, and the subjects were given no feedback on their performances. Beside the children's responses, we recorded the completion time for each task. 2.3. Statistical analysis All demographical, clinical, and neuropsychological data were analyzed by means of the Statistical Package for Social Sciences (SPSS, version 18.0). The FIMT, FEMT, and FELT scores referred to the number of correct responses given by the subject. Since some previous reports of emotion recognition in school-aged children showed an improvement in performance from 6 to 10 years of age [46,47], we explored the effect of age at testing in the CTRL population. However, we failed to find a relationship between age (6, 7, 8, 9, and 10 years age) and FEMT or FELT scores (Kruskal–Wallis test: FEMT p = .606, FELT p = .106) (Fig. s-2). Accordingly, the age at testing was not considered in further analysis. To compare the recognition of emotions (FEMT and FELT scores) between groups, a nonparametric statistical method (the Mann–Whitney U test) was used because of their non-normal distribution (as revealed by the Kolmogorov–Smirnov test). Differences in time employed to accomplish each test were assessed by ANOVA. In order to compare the subscores of each of the five emotions, the level of significance was adjusted for multiple comparisons with Bonferroni correction (alpha = 0.0102 for each test). If all emotion subscores showed significant differences among groups, then the most relevant emotions were identified through the receiver operating characteristic (ROC) plots, comparing the sensitivity and specificity of each emotion subscore in distinguishing between the FS and CTRL groups. To analyze individual differences, we calculated how many subjects in the FS group were impaired compared with controls. We computed separate confidence intervals of 2 standard deviations (SDs) (z N 2, p b .01; two-tailed) from the control mean for the FEMT and FELT. The proportion of patients in each group falling below the control confidence interval by more than 2 SDs was considered to provide a subject-specific index of impaired task performance. For the FS’ group, correlations among emotion recognition scores and clinical variables were calculated by means of linear regression. Differences in distribution of clinical variables among subgroups of patients were assessed by means of the chi-square test.
3. Results The distribution of IQ scores (Table 2), assessed through Raven's CPM, differed significantly between the CTRL group and the FS group (χ2 = 0.019). In particular, 13.2% of children in the FS group had IQ scores below the 50th percentile compared to none in the CTRL group. Furthermore, none of the CTRL subjects made any mistakes in the FIMT, whereas 15.8% of the FS group did. In order to avoid biases due to deficits in basic visuoperceptual abilities on performances in the facial emotion matching and labeling tasks, 10 subjects from the FS group (either with an IQ below the 50th percentile [n = 5] or with errors in the FIMT [n = 6]) were excluded from further analysis of FER ability. The assumption that a deficit in visuoperceptual abilities might be a confounder in the assessment of emotion Table 2 Neuropsychological assessment results.
Raven's CPM Percentile
FIMT Total (max 10)
CTRL (n = 38) n (percentage)
FS (n = 38) n (percentage)
Chi-square test
N95 50–95 b50
21 (55.3%) 17 (44.72%) –
12 (31.6%) 21 (55.3%) 5 (13.1%)§
χ2 = .019*
10 b10
38 (100%) –
32 (84.2%) 6 (15.8%)§
χ2 = .011*
FER assessment CTRL (n = 38) mean (±SD)
FS’ (n = 28) mean (±SD)
Mann–Whitney test
FEMT Happiness (max 8) Sadness (max 8) Fear (max 8) Disgust (max 8) Anger (max 8) Total score (max 40)
7.97 (±0.2) 6.53 (±1.2) 7.11 (±1.0) 7.21 (±1.0) 4.63 (±1.2) 33.47 (±2.6)
7.64 (±0.7) 5.46 (±1.6) 5.75 (±1.7) 6.11 (±1.3) 3.68 (±1.6) 28.64 (±3.8)
U U U U U U
= = = = = =
412.0; p 328.5; p 264.0; p 259.5; p 333.5; p 165.0; p
= .006* = .007* b .0001* b .0001* = .009* b .0001*
FELT Happiness (max 5) Sadness (max 5) Fear (max 5) Disgust (max 5) Anger (max 5) Total score (max 25)
5.00 (±0.0) 4.74 (±0.5) 4.84 (±0.5) 4.37 (±0.7) 4.08 (±0.4) 23.03 (±1.0)
5.00 (±0.0) 4.18 (±1.0) 4.25 (±1.1) 3.61 (±1.3) 4.21 (±0.5) 21.25 (±2.4)
U U U U U U
= = = = = =
532.0; p 353.0; p 355.5; p 349.0; p 599.5; p 290.0; p
= 1.00 = .006* = .002* = .010* = .230 = .001*
IQ and Face Identity Matching score distributions among the two groups differed significantly (*), so 10 subjects from the FS group (§) were excluded from further analysis on FER.
214
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
recognition ability is supported by results of a preliminary analysis (see Table s-2, online supplementary material) indicating that the subjects excluded from subsequent analysis had significantly lower performance in the FEMT than the subjects with intact visuoperceptual abilities (p b .003; Mann–Whitney test). Also, FELT score was lower in the children with visuoperceptual deficits compared with subjects with intact visuoperceptual abilities, but this did not reach significance (p = .302). Therefore, only the FS subgroup (FS’ group; n = 28) with intact visuoperceptual abilities was further compared with the CTRL group. Neither age at the time of testing (one-way ANOVA) nor gender/handedness differed significantly (chi-square test) between the FS’ group (n = 28) and controls (n = 38). Intelligence quotient did not differ between the FS’ group (mean 90.5 ± 10.9) and the CTRL group (mean 94.1 ± 8.7) (one-way ANOVA, p = .148). No time difference in completing the two emotion-related tasks was found between controls (mean = 449 ± 128 s) and the FS’ group (mean = 427 ± 139 s). The FS’ group achieved a lower recognition score than the control group in both matching (mean ± SD: 28.64 ± 3.8 vs 33.47 ± 2.6; p b .0001) and labeling (21.25 ± 2.4 vs 23.03 ± 1.0; p = .001) facial emotions. Correlation analyses did not reveal a significant relationship between IQ scores and either FEMT (Spearman's rho = 0.281; p = .148) or FELT (Spearman's rho = 0.185; p = .347). On the contrary, we found a significant correlation between matching and labeling performance (Spearman rho = 0.625; p b .0005), supporting the validity of the two measures in assessing FER ability. At the level of each single emotion, the FS’ group showed significantly lower scores compared with controls in matching all of the five basic emotions (see Table 2 and Fig. 1A), whereas they were impaired in labeling three out of the five facial expressions: sadness (4.18 vs 4.74; p = .006), fear (4.25 vs 4.84; p = .002), and disgust (3.61 vs 4.37; p = .010) (see Table 2 and Fig. 1B). Even though the FS’ group's performance was significantly worse than that of the CTRL group across all emotions in the FEMT, the largest area under the ROC curve was obtained for disgust and fear, with an asymptotic significance of below 0.001 (Fig. s-3). Linear regression did not reveal any correlation of FEMT and FELT scores with clinical variables, such as (a) age at first FS, (b) number of FS episodes, (c) presence of prolonged and/or complex FSs, and (d) abnormal EEG within 48 h since FSs.
Fig. 2. Distributions of FEMT and FELT scores. Frequency distributions of subjects as a function of facial emotion recognition scores (scores on the y-axis) for FEMT (A) and FELT (B). In both histograms, controls are represented on the left and the FS’ group on the right. The horizontal dashed lines indicate the cutoff of two standard deviations from the controls' mean.
Fig. 1. Facial emotion recognition ability across the five basic emotions. Performance of the FS’ group and controls in matching (A) and labeling (B) each of the five basic emotions, expressed as the mean amount of correct responses (±standard error). Ordinate axis is set in both absolute number and percentage. Asterisks correspond to significant comparisons (Mann–Whitney test; p ≤ .01).
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
Finally, we explored differences in clinical features at a single subject level. Twelve patients (31.6%) were impaired in both FEMT and FELT, while four (10.5%) were impaired only in matching and another four exclusively in labeling (Fig. 2). These patients did not differ in any clinical variables from the remaining subjects of the FS’ group (age: p = .344; gender: p = .549; handedness: p = .353; age at first FS: p = .246; number of FS episodes: p = .760; presence of complex FSs: p = .847; presence of prolonged FSs: p = .393; presence of EEG abnormalities: p = .234). 4. Discussion The main finding of our study is that school-aged healthy children with antecedent FSs show lower scores in recognition of basic emotions compared with age-matched children without FSs. This result raises the question of whether FSs can alter long-term plasticity in extrahippocampal limbic regions, such as the amygdala, insular cortex, and other cortical–subcortical structures involved in emotion recognition. The demonstration of subtle emotion recognition deficits in children with antecedent FSs can be relevant both to the field of clinical epileptology and to emotion research. Thus, before discussing the significance of our results, some important methodological issues and limitations of our study should be taken into account. A first point concerns patient selection. Indeed, severe cases might have been overrepresented in our FS group because of the selection bias of a hospital-based series. Furthermore, the low participation rate of eligible subjects can also generate a selection bias. However, both these biases would affect the epidemiological value of the study (that is beyond the aim of our study). On the contrary, we consider that the intrinsic validity of the study, demonstrating FER deficits even in subjects with a normal IQ and face identity recognition, is scarcely affected by these biases. This is even more plausible considering that the distribution of clinical features in our sample was not exceptional. In fact, although our work was not designed to obtain epidemiological data, the clinical features of our FS group did not differ from the ones provided by populationbased cohort studies [48,49]. A second point concerns the ‘specificity’ of the observed impairments in FER. Because of the limited time made available by the children's parents, we could not perform a comprehensive examination of neuropsychological functions. Thus, we were forced to limit our control tasks to a general intelligence assessment and a visuoperceptual task employing the same type of stimuli as the FER tasks. Indeed, 13.2% of the children in the FS group had an IQ below the 50th percentile, and 16% showed some mistakes in matching facial identity. Aiming to overcome this possible and relevant bias, we then restricted FER analysis only to children with a normal IQ and intact face recognition abilities; yet, these children still showed a significant impairment in FER. Furthermore, correlation analysis of CPM and FER scores revealed that a general intelligence measure could not predict performance in evaluating facial expressions. Hence, we are quite confident that lower scores in both the FEMT and FELT are related to specific difficulties in emotion recognition rather than to other more general cognitive impairments. A third point concerns the fact that none of our patients underwent a brain MRI in order to assess medial temporal lobe structures and possible FS-induced amygdala–hippocampal damage. However, there was no clinical indication to perform MRI in this group of children [39] and, because of ethical issues, a brain MRI study has not been proposed. Nevertheless, our data demonstrate that school-aged children with antecedent FSs have difficulties in matching and labeling emotional expressions compared with an age-matched control group from the same sociocultural background. These results support the hypothesis that FSs can be associated with specific cognitive function deficits. To the best of our knowledge, only one study has addressed specific medial temporal lobe functions, including learning, memory retrieval, and consolidation, in school-aged children with a history of FSs [17]. These authors
215
observed that the occurrence of FS before 1 year of age was associated with deficits in memory functions. Furthermore, MRI studies detected hippocampal abnormalities not only in children after complex FSs [9] but also in adults with a history of simple FSs without subsequent epilepsy [50]. Morphological brain changes have also been demonstrated in animal models including FSinduced neuroanatomical changes (such as mossy fiber sprouting, loss of inhibitory interneurons, and persistent granule cell ectopia) are associated with impairment in functions subserved by medial temporal lobe structures, i.e., learning and memory [51,52]. Although we do not provide MRI evidence of medial temporal lobe (or of amygdala) damage in the FS group, it is possible to consider that the neural network involved in emotion processing is somewhat dysfunctional in these children. It is important to note that the difference observed between the FS and CTRL groups, while significant, was relatively small. However, FER deficits were quite common among children in the FS group. Indeed, 31.6% of children were impaired in both matching and labeling facial expressions. As far as the recognition of specific emotions is concerned, we observed that in the FS group, disgust and fear were the two emotions for which errors were present in both tasks (labeling and matching). However, the emotion-matching task appeared particularly difficult for children in the FS group, with deficits observed in matching every emotion category, including happiness. Recently, we reported FER abilities in a large cohort of 140 adult patients with chronic MTLE evaluated across a 5-year period [30]. This study confirmed that widespread deficits in the recognition of negative emotions are common in this seizure disorder. In particular, patients with early onset of seizures, including FSs, and bilateral damage are severely impaired in emotion recognition. Indeed, emotion recognition deficits seem to be clinically relevant and severe in patients suffering from early-life amygdala damage [53,54] and in adult patients with temporal lobe epilepsy with seizure onset during childhood, including FSs [23,24,55–57]. Overall, these findings raise the important question of whether a critical period of life exists for establishing the neural network underlying FER abilities [23]. Notably, Golouboff and coworkers [33] assessed this question in young patients with MTLE, documenting that FER deficits are already present during childhood and adolescence in patients with early-onset MTLE. These authors underlined that early seizure onset (including FSs) was associated with poor recognition of facial expression, especially for fear. In this context, the results of our study suggest that FSs, even in the absence of overt epilepsy, are sufficient to cause subtle emotion recognition deficits, thus strengthening the view that FSs are a marker of medial temporal lobe dysfunction. Clearly, we cannot demonstrate the putative causative role of FSs on the long-term plasticity of the amygdala and interconnected areas subserving emotion processing. Indeed, it could be possible that FER impairments and FSs are the markers of a preexisting damage within the temporal lobe. In conclusion, we demonstrated that emotion recognition abilities could be defective in school-aged children with a history of FSs, even in children with an isolated simple FS. This deficit was quite common in our patient population, even if it was not clinically relevant in daily life. Because of the limited sample size, our results need to be confirmed in longitudinal studies and/or larger populations in order to establish their specificity and sensitivity. In particular, the use of a more comprehensive neuropsychological test battery would allow us to find out if other functions are altered in children with FSs, in addition to FER deficit. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.yebeh.2013.07.007. Acknowledgments Dr. Meletti received a research grant from the Emilia Romagna Region (Grant Innovative Research Area 1A) and from CaRisMo Foundation.
216
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
References [1] Baram TZ, Shinnar S. Febrile seizures. Academic Press; 2002. [2] Nelson KB, Ellenberg JH. Predictors of epilepsy in children who have experienced febrile seizures. N Engl J Med 1976;295(19):1029–33. [3] Kira R, Ishizaki Y, Torisu H, Sanefuji M, Takemoto M, Sakamoto K, et al. Genetic susceptibility to febrile seizures: case–control association studies. Brain Dev 2010;32(1):57–63. [4] Baulac S, Gourfinkel-An I, Nabbout R, Huberfeld G, Serratosa J, Leguern E, et al. Fever, genes, and epilepsy. Lancet Neurol 2004;3(7):421–30. [5] Marini C, Mantegazza M. Na+ channelopathies and epilepsy: recent advances and new perspectives. Expert Rev Clin Pharmacol 2010;3(3):371–84. [6] Lado FA, Laureta EC, Moshé SL. Seizure-induced hippocampal damage in the mature and immature brain. Epileptic Disord 2002;4(2):83–97. [7] VanLandingham KE, Heinz ER, Cavazos JE, Lewis DV. Magnetic resonance imaging evidence of hippocampal injury after prolonged focal febrile convulsions. Ann Neurol 1998;43(4):413–26. [8] Scott RC, Gadian DG, King MD, Chong WK, Cox TC, Neville BGR, et al. Magnetic resonance imaging findings within 5 days of status epilepticus in childhood. Brain 2002;125(Pt 9):1951–9. [9] Scott RC, King MD, Gadian DG, Neville BGR, Connelly A. Hippocampal abnormalities after prolonged febrile convulsion: a longitudinal MRI study. Brain 2003;126(Pt 11): 2551–7. [10] Scott RC, King MD, Gadian DG, Neville BGR, Connelly A. Prolonged febrile seizures are associated with hippocampal vasogenic edema and developmental changes. Epilepsia 2006;47(9):1493–8. [11] Wallace SJ. They don't do very well. Pediatrics 1980;65(3):678–9. [12] Tsai JJ, Huang MC, Lung FW, Huang CC, Change YC. Differences in factors influencing the familial aggregation of febrile convulsion in population and hospital patients. Acta Neurol Scand 1996;94(5):314–9. [13] Ellenberg JH, Nelson KB. Febrile seizures and later intellectual performance. Arch Neurol 1978;35(1):17–21. [14] Verity CM, Butler NR, Golding J. Febrile convulsions in a national cohort followed up from birth. II—medical history and intellectual ability at 5 years of age. Br Med J 1985;290(6478):1311–5. [15] Verity CM, Greenwood R, Golding J. Long-term intellectual and behavioral outcomes of children with febrile convulsions. N Engl J Med 1998;338(24):1723–8. [16] Chang YC, Guo NW, Huang CC, Wang ST, Tsai JJ. Neurocognitive attention and behavior outcome of school-age children with a history of febrile convulsions: a population study. Epilepsia 2000;41(4):412–20. [17] Chang YC, Guo NW, Wang ST, Huang CC, Tsai JJ. Working memory of school-aged children with a history of febrile convulsions: a population study. Neurology 2001;57(1):37–42. [18] Adolphs R. Neural systems for recognizing emotion. Curr Opin Neurobiol 2002;12(2): 169–77. [19] Pitkänen A, Tuunanen J, Kälviäinen R, Partanen K, Salmenperä T. Amygdala damage in experimental and human temporal lobe epilepsy. Epilepsy Res 1998;32(1–2): 233–53. [20] Aggleton JP. The contribution of the amygdala to normal and abnormal emotional states. Trends Neurosci 1993;16(8):328–33. [21] LeDoux JE. Emotion circuits in the brain. Annu Rev Neurosci 2000;23:155–84. [22] LeDoux J. The amygdala. Curr Biol 2007;17(20):R868–74. [23] Meletti S, Benuzzi F, Rubboli G, Cantalupo G, Stanzani-Maserati M, Nichelli P, et al. Impaired facial emotion recognition in early-onset right mesial temporal lobe epilepsy. Neurology 2003;60(3):426–31. [24] Benuzzi F, Meletti S, Zamboni G, Calandra-Buonaura G, Serafini M, Lui F, et al. Impaired fear processing in right mesial temporal sclerosis: a fMRI study. Brain Res Bull 2004;63(4):269–81. [25] Reynders HJ, Broks P, Dickson JM, Lee CE, Turpin G. Investigation of social and emotion information processing in temporal lobe epilepsy with ictal fear. Epilepsy Behav 2005;7(3):419–29. [26] Yamada M, Murai T, Sato W, Namiki C, Miyamoto T, Ohigashi Y. Emotion recognition from facial expressions in a temporal lobe epileptic patient with ictal fear. Neuropsychologia 2005;43(3):434–41. [27] Schacher M, Winkler R, Grunwald T, Kraemer G, Kurthen M, Reed V, et al. Mesial temporal lobe epilepsy impairs advanced social cognition. Epilepsia 2006;47(12): 2141–6. [28] Shaw P, Lawrence E, Bramham J, Brierley B, Radbourne C, David AS. A prospective study of the effects of anterior temporal lobectomy on emotion recognition and theory of mind. Neuropsychologia 2007;45(12):2783–90.
[29] Walpole P, Isaac CL, Reynders HJ. A comparison of emotional and cognitive intelligence in people with and without temporal lobe epilepsy. Epilepsia 2008;49(8):1470–4. [30] Meletti S, Benuzzi F, Cantalupo G, Rubboli G, Tassinari CA, Nichelli P. Facial emotion recognition impairment in chronic temporal lobe epilepsy. Epilepsia 2009;50(6): 1547–59. [31] Bonora A, Benuzzi F, Monti G, Mirandola L, Pugnaghi M, Nichelli P, et al. Recognition of emotions from faces and voices in medial temporal lobe epilepsy. Epilepsy Behav 2011;20(4):648–54. [32] Broicher SD, Kuchukhidze G, Grunwald T, Krämer G, Kurthen M, Jokeit H. “Tell me how do I feel”—emotion recognition and theory of mind in symptomatic mesial temporal lobe epilepsy. Neuropsychologia 2012;50(1):118–28. [33] Golouboff N, Fiori N, Delalande O, Fohlen M, Dellatolas G, Jambaqué I. Impaired facial expression recognition in children with temporal lobe epilepsy: impact of early seizure onset on fear recognition. Neuropsychologia 2008;46(5):1415–28. [34] Adolphs R. The social brain: neural basis of social knowledge. Annu Rev Psychol 2009;60:693–716. [35] Ekman P, Freisen WV, Ancoli S. Facial signs of emotional experience. J Pers Soc Psychol 1980;39(6):1125–34. [36] Ekman P. Expression and the nature of emotion. In: Scherer K, Ekman P, editors. Approaches to emotion. Hillsdale, New Jersey: Lawrence Erlbaum; 1984. p. 319–44. [37] Commission on Epidemiology and Prognosis, International League Against Epilepsy. Guidelines for epidemiologic studies on epilepsy. Epilepsia 1993;34(4):592–6. [38] Berg AT, Shinnar S. Complex febrile seizures. Epilepsia 1996;37(2):126–33. [39] Capovilla G, Mastrangelo M, Romeo A, Vigevano F. Recommendations for the management of “febrile seizures”: Ad Hoc Task Force of LICE Guidelines Commission. Epilepsia 2009;50(Suppl. 1):2–6. [40] Raven JC. CPM: Coloured Progressive Matrices. Firenze: O.S. Organizzazioni Speciali; 1996. [41] Ekman P, Friesen WV. Pictures of Facial Affect. Palo Alto, CA: Consulting Psychologists Press; 1976. [42] O'Sullivan M, Ekman P. Facial expression recognition and emotional intelligence. In: Leeland KB, editor. Face recognition: new research. New York, NY: Nova Science Publishers; 2008. p. 25–40. [43] Ekman P. Facial expressions of emotion: an old controversy and new findings. Philos Trans R Soc Lond B Biol Sci 1992;335(1273):63–9. [44] Darwin C. The expression of the emotions in man and animals. New York: Philosophical Library; 1955. [45] Rapcsak SZ, Galper SR, Comer JF, Reminger SL, Nielsen L, Kaszniak AW, et al. Fear recognition deficits after focal brain damage: a cautionary note. Neurology 2000;54(3): 575–81. [46] Vicari S, Reilly JS, Pasqualetti P, Vizzotto A, Caltagirone C. Recognition of facial expressions of emotions in school-age children: the intersection of perceptual and semantic categories. Acta Paediatr 2000;89(7):836–45. [47] Durand K, Gallay M, Seigneuric A, Robichon F, Baudouin J-Y. The development of facial emotion recognition: the role of configural information. J Exp Child Psychol 2007;97(1):14–27. [48] Verity CM, Golding J. Risk of epilepsy after febrile convulsions: a national cohort study. BMJ 1991;303(6814):1373–6. [49] Berg AT, Shinnar S, Darefsky AS, Holford TR, Shapiro ED, Salomon ME, et al. Predictors of recurrent febrile seizures. A prospective cohort study. Arch Pediatr Adolesc Med 1997;151(4):371–8. [50] Auer T, Barsi P, Bone B, Angyalosi A, Aradi M, Szalay C, et al. History of simple febrile seizures is associated with hippocampal abnormalities in adults. Epilepsia 2008;49(9): 1562–9. [51] Dubé CM, Zhou J-L, Hamamura M, Zhao Q, Ring A, Abrahams J, et al. Cognitive dysfunction after experimental febrile seizures. Exp Neurol 2009;215(1):167–77. [52] Koyama R, Tao K, Sasaki T, Ichikawa J, Miyamoto D, Muramatsu R, et al. GABAergic excitation after febrile seizures induces ectopic granule cells and adult epilepsy. Nat Med 2012. http://dx.doi.org/10.1038/nm.2850. [53] Calder AJ. Facial emotion recognition after bilateral amygdala damage: differentially severe impairment of fear. Cogn Neuropsychol 1996;13(5):699–745. [54] Hamann SB, Stefanacci L, Squire LR, Adolphs R, Tranel D, Damasio H, et al. Recognizing facial emotion. Nature 1996;379(6565):497. [55] Anderson AK, Spencer DD, Fulbright RK, Phelps EA. Contribution of the anteromedial temporal lobes to the evaluation of facial emotion. Neuropsychology 2000;14(4): 526–36. [56] Adolphs R, Tranel D, Damasio H. Emotion recognition from faces and prosody following temporal lobectomy. Neuropsychology 2001;15(3):396–404. [57] McClelland III S, Garcia RE, Peraza DM, Shih TT, Hirsch LJ, Hirsch J, et al. Facial emotion recognition after curative nondominant temporal lobectomy in patients with mesial temporal sclerosis. Epilepsia 2006;47(8):1337–42.
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Facial emotion recognition in childhood: The effects of febrile seizures in the developing brain Gaetano Cantalupo a,b,⁎,1, Stefano Meletti c,1, Alessia Miduri a, Silvia Mazzotta a, Loreto Rios-Pohl d, Francesca Benuzzi c, Francesco Pisani a, Carlo Alberto Tassinari e,f, Giuseppe Cossu a,f a
Child Neuropsychiatry Unit, Department of Neuroscience, University-Hospital of Parma, Via Gramsci 14, Parma, Italy Department of Life and Reproduction Sciences, University of Verona, P.le Scuro 10, Verona, Italy Department of Biomedical Sciences, Metabolism and Neuroscience, University of Modena and Reggio Emilia, Nuovo Ospedale Civile Sant'Agostino Estense, Via Giardini 1355, Modena, Italy d Centro Avanzado de Epilepsia, Clinica Las Condes, Lo Fontecilla 441, Las Condes, Santiago, Chile e Department of Neurological Sciences, University of Bologna, Via Foscolo 7, Bologna, Italy f Department of Neuroscience, University of Parma, Via Volturno 39/E, Parma, Italy b c
a r t i c l e
i n f o
Article history: Received 11 February 2013 Revised 4 July 2013 Accepted 7 July 2013 Available online 27 August 2013 Keywords: Emotions Facial expression Emotional intelligence Medial temporal lobe Limbic system Amygdala Febrile convulsion Epilepsy Temporal lobe epilepsy Pediatric neurology
a b s t r a c t It has been documented that anteromedial temporal lobe dysfunction can cause impairment in emotional intelligence. In particular, medial temporal lobe epilepsy (MTLE) is associated with disorders in emotion recognition from facial expressions. About one-third of patients with MTLE experienced febrile seizures (FSs) during childhood. In the present study, we investigated facial emotion recognition ability in a group of 38 school-aged children with antecedent FSs and in an age- and sex-matched control group. Children with abnormal general visuoperceptual abilities were excluded. Children with FSs showed lower recognition scores versus controls in both matching (28.64 vs 33.47; p b .0001) and labeling (21.25 vs 23.03; p = .001) facial emotions. Our findings support the hypothesis that FSs can be associated during childhood with a dysfunction within the neural network subserving the processing of facial expressions of the basic emotions. © 2013 Elsevier Inc. All rights reserved.
1. Introduction Febrile seizures (FSs) are the most common seizure disorder in childhood associated with fever but without evidence of intracranial infection or defined causes [1]. They occur in 2–5% of children younger than 5 years of age, and based on duration and clinical features, FSs are subdivided into ‘simple’ and ‘complex’ [1,2]. Epidemiological studies have substantially improved our understanding of the frequency, natural history, and recurrence of FSs, and there is now a general agreement that simple FSs are benign and rarely require chronic anticonvulsant treatment. Recent clinical and genetic studies suggest that the relationship between FSs and the development of epilepsy later in life is frequently genetic, and there are a number of genes that are either involved in susceptibility for FSs [3] or are pathogenic in epilepsies that present with FSs in their course [4,5]. However, the impact of FSs on ⁎ Corresponding author at: Child Neuropsychiatry Unit, Department of Neuroscience, University-Hospital of Parma, Via Gramsci 14, 43126, Parma, Italy. Fax: +39 0521704708. E-mail address: [email protected] (G. Cantalupo). 1 The first two authors contributed equally to this manuscript. 1525-5050/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.yebeh.2013.07.007
the developing brain has not been completely understood. Since FSs are the most common seizures in children, it is important to delineate whether early-life FSs can alter long-term neuroplasticity [6], especially the cognitive functions. In particular, it is important to evaluate whether FSs could be linked to persistent subtle brain injuries and the related cognitive dysfunctions. Regarding anatomical injury, longitudinal MRI studies have documented that prolonged and focal FSs can occasionally produce acute hippocampal injury that evolves into atrophy [7–9]. On the other hand, it has been proposed that a preexisting developmental hippocampal abnormality may predispose individuals to having a prolonged febrile seizure [9,10]. Indeed, the association between hippocampal sclerosis and FSs raises the possibility of selective cognitive sequelae involving the cognitive functions subserved by the medial temporal lobe region. In early reports from hospital-based studies, the prognosis of FSs, in terms of cognitive outcome, was fairly pessimistic (8% to 22% of cases presented mental retardation) because of the inclusion of symptomatic causes of seizures other than fever [11]. Furthermore, the hospital-based studies may be skewed toward disproportionately severe cases [12]. In contrast, long-term population-based cohort studies showed that children with FSs have similar global neurocognitive
212
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
developmental and academic performance compared with controls [13–16]. Therefore, the current viewpoint is that the vast majority of children with FSs have normal global measures of cognition and behavior. However, previous studies on the outcome of children with FSs only rarely focused on specific cognitive functions that can reveal a FS-related hippocampal injury, such as memory [17]. Neural networks underlying facial emotion recognition involve a distributed set of structures that include the visual cortices, the amygdala, the orbitofrontal cortex, and additional cerebral regions, such as the insula, the basal ganglia, and the prefrontal cortex [18]. The amygdala, which is often damaged along with the hippocampus in patients with medial temporal lobe epilepsy (MTLE) [19], has been identified as a key structure for evaluating emotional stimuli and regulating social and emotional behavior [20–22]. Accordingly, the investigation of emotional and social competences in patients with MTLE has been the focus of different studies [23–32] documenting common and widespread deficits of emotion recognition in this seizure disorder. In particular, patients with early onset of seizures, including FSs, and bilateral medial temporal lobe damage were severely impaired in emotion recognition [23,30,33]. The recognition of emotional signals, from all sensory modalities, is a critical component of human social interactions because it is through the understanding of the affective states of others that we modulate our behavioral responses [34]. Notably, facial expressions provide the greatest amount of emotional cues for recognizing emotions, both with positive and negative values [35]. The expression of basic emotions is supposed to be universal, innate, and have a specific neural substrate; on the contrary, the recognition of social emotions is learned over a lifetime. Thus, the alteration of emotional and social competences could represent another index of the potential FS-related medial temporal lobe injury. To the best of our knowledge, no study has investigated emotional–social abilities in children with antecedent FSs. In the present study, we assessed the ability of school-aged children with a history of FSs to recognize the facial expressions of basic innate emotions (happiness, sadness, fear, anger, and disgust) [36] in order to explore the possibility that FSs are associated with (or can lead to) a dysfunction within the neural network subserving the processing of facial expressions of basic emotions. Accordingly, we investigated facial emotion recognition (FER) ability in a sample of healthy school-aged children with antecedent FSs and compared their performances with an age- and sex-matched control group. 2. Material and methods 2.1. Participants We reviewed the database from the Child Neuropsychiatry Unit of the University-Hospital of Parma (Italy) to recruit children referred to our service between 2000 and 2008 for FSs, who were school-aged at the time of the enrollment. Following the International League Against Epilepsy [37], a FS was defined as a seizure occurring between 1 month and 5 years of age, associated with a febrile illness but not caused by an infection of the CNS, without previous neonatal seizures or a previous unprovoked seizure, and not meeting the criteria for other acute symptomatic seizures. Accordingly, we selected only cases with a complete diagnostic workup, including detailed eyewitness accounts of the episodes, clinical course, neurological evaluation, and an EEG recording within 48 h after the FS. Out of 163 eligible subjects aged 6–10 years, with at least one episode of FS and without subsequent epilepsy, 150 could be contacted, and 25.3% of the children's parents accepted to participate. Thus, 38 children (18 females, 20 males) were recruited as the clinical group (FS group). A group of 38 age- and sex-matched children (18 females, 20 males), attending primary school from the same district, served as the control group (CTRL). All participants' parents gave their informed consent, and the study was approved by the local ethical committee. Demographic and clinical variables were collected through a semistructured
interview of parents to explore the children's handedness, educational level, and familial and personal history. A history of neurological and/ or neuropsychiatric antecedents was an exclusion criterion for children to be enrolled in the control group. The presence of EEG abnormalities, for the FS group, was evaluated. Demographic variables of the two groups are reported in Table 1; neither the age at the time of testing (ANOVA) nor gender and handedness distributions (chi-square test) were significantly different between groups. For the FS group, the characteristics of FSs were reported, in order to be classified as ‘simple’ or ‘complex’ [2,38,39], along with the age at the first episode and the number of episodes (see Table 1). Twenty-one out of 38 (55.3%) children experienced a single episode, more than 80% (31/38) of episodes were classified as simple, and only seven patients experienced at least one episode classified as complex, being focal (4/38) and/or prolonged (4/38) and/or drug-interrupted (2/38). Half of our FS group (19/38) experienced only a single, simple, and short FS. 2.2. Neuropsychological assessment The children's IQ was evaluated by means of Raven's colored progressive matrices (CPM) [40]. The experimental protocol included three tasks built ad hoc with Ekman and Friesen's Pictures of Facial Affect [41] depicting five emotional expressions (happiness, sadness, fear, disgust, and anger) or no emotion (neutral). Ekman and Friesen's Pictures of Facial Affect represent a validated and reliable set of posed emotional expressions that generate high recognition rates across cultures; accordingly, they have been used in a wide range of studies, including developmental, cross-cultural, neuroimaging, and behavioral studies [42,43]. These black and white pictures were initially developed and used with adults, but they were later fruitfully implemented in studies with children [42]. Facial stimuli expressing surprise were not included because the intrinsic ambiguity of this emotion led some authors to consider it not among the basic emotions [43] or to consider it a ‘gradation’ of fear [44], and even healthy subjects often confuse fear for surprise [45]. The original stimuli were digitally edited in order to remove all nonface cues (i.e., clothing, neck, and hair). 2.2.1. Face Identity Matching Test Neutral stimuli from 10 subjects (five males) were used to create the Face Identity Matching Test (FIMT), aimed to control for subjects' basic visuoperceptual ability with facial stimuli. In this task, the children were required to identify the target face, portrayed in the upper half of the page, among five alternatives (one target and four confounders) depicted below (Fig. s-1A, online supplementary material). 2.2.2. Facial Expression Matching Test The second task, the Facial Expression Matching Test (FEMT), was created to assess FER ability, and it did not require any verbal expertise. This test comprised 40 emotional items in which the probe facial expressions resulted in a high rate of correct identification in the adult data [41]. All stimuli were faces of four women (C, MF, MO, and MR in the Ekman and Friesen's series) and four men (EM, GS, JJ, and WF) depicting the five basic emotions. For each item, the subjects were requested to match the probe facial expression with one of the five alternatives aligned below. The targets and confounders were expressions of the same actor, while the probe expression belonged to a different actor of the same gender (Fig. s-1B). The probe emotion and gender, as well as the position of the target among the confounders, varied across subsequent items in a pseudorandom sequence in order to minimize repetitions (see Table s-1, online supplementary material). 2.2.3. Facial Expression Labeling Test This test was already used in previous studies from our group to quantify the FER ability of adult subjects [23,30,31]. Briefly, in this task for each trial, the subject was presented with one original stimulus from Ekman and Friesen's series (including nonfacial features) and
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
213
Table 1 Demographic data and clinical features of FS.
Age at testing Gender n (percentage) Handedness n (percentage) Age at first FS episode Complex FSs n (percentage) Prolonged FSs n (percentage) Number of FS episodes n (percentage)
Mean (±SD) Range Female Male Right Left Mean (±SD) Range
CTRL (n = 38)
FS (n = 38)
Statistical analysis
8.2 (±1.4) 6–10 18 (47.4%) 20 (52.6%) 33 (86.8%) 5 (13.2%)
7.7 (±1.4) 6–10 18 (47.4%) 20 (52.6%) 35 (92.1%) 3 (7.9%) 2.42 (±1.2) 1–5 7 (18.4%)
F = 2.508; p = .118 (ANOVA)
Only N15 min Including drug-interrupted #1 #2 #3 #4 #N4
χ2 = 1.000; p = 1.000 (FET) χ2 = 0.455; p = .711 (FET)
4 (10.5%) 6 (15.8%) 21 (55.3%) 9 (23.7%) 1 (2.6%) 2 (5.3%) 5 (12.9%)
SD= Standard Deviation; ANOVA= analysis of variance; FET= Fisher's Exact Test.
was required to choose one of the five verbal labels (happiness, sadness, fear, disgust, and anger) printed below the target face (and read aloud by the experimenter, if required) that best described the emotion shown (Fig. s-1C). This Facial Expression Labeling Test (FELT) consisted of 25 trials (five face stimuli for each emotion). The participants were instructed to consider carefully all five alternatives before answering. There was no time limit, and the subjects were given no feedback on their performances. Beside the children's responses, we recorded the completion time for each task. 2.3. Statistical analysis All demographical, clinical, and neuropsychological data were analyzed by means of the Statistical Package for Social Sciences (SPSS, version 18.0). The FIMT, FEMT, and FELT scores referred to the number of correct responses given by the subject. Since some previous reports of emotion recognition in school-aged children showed an improvement in performance from 6 to 10 years of age [46,47], we explored the effect of age at testing in the CTRL population. However, we failed to find a relationship between age (6, 7, 8, 9, and 10 years age) and FEMT or FELT scores (Kruskal–Wallis test: FEMT p = .606, FELT p = .106) (Fig. s-2). Accordingly, the age at testing was not considered in further analysis. To compare the recognition of emotions (FEMT and FELT scores) between groups, a nonparametric statistical method (the Mann–Whitney U test) was used because of their non-normal distribution (as revealed by the Kolmogorov–Smirnov test). Differences in time employed to accomplish each test were assessed by ANOVA. In order to compare the subscores of each of the five emotions, the level of significance was adjusted for multiple comparisons with Bonferroni correction (alpha = 0.0102 for each test). If all emotion subscores showed significant differences among groups, then the most relevant emotions were identified through the receiver operating characteristic (ROC) plots, comparing the sensitivity and specificity of each emotion subscore in distinguishing between the FS and CTRL groups. To analyze individual differences, we calculated how many subjects in the FS group were impaired compared with controls. We computed separate confidence intervals of 2 standard deviations (SDs) (z N 2, p b .01; two-tailed) from the control mean for the FEMT and FELT. The proportion of patients in each group falling below the control confidence interval by more than 2 SDs was considered to provide a subject-specific index of impaired task performance. For the FS’ group, correlations among emotion recognition scores and clinical variables were calculated by means of linear regression. Differences in distribution of clinical variables among subgroups of patients were assessed by means of the chi-square test.
3. Results The distribution of IQ scores (Table 2), assessed through Raven's CPM, differed significantly between the CTRL group and the FS group (χ2 = 0.019). In particular, 13.2% of children in the FS group had IQ scores below the 50th percentile compared to none in the CTRL group. Furthermore, none of the CTRL subjects made any mistakes in the FIMT, whereas 15.8% of the FS group did. In order to avoid biases due to deficits in basic visuoperceptual abilities on performances in the facial emotion matching and labeling tasks, 10 subjects from the FS group (either with an IQ below the 50th percentile [n = 5] or with errors in the FIMT [n = 6]) were excluded from further analysis of FER ability. The assumption that a deficit in visuoperceptual abilities might be a confounder in the assessment of emotion Table 2 Neuropsychological assessment results.
Raven's CPM Percentile
FIMT Total (max 10)
CTRL (n = 38) n (percentage)
FS (n = 38) n (percentage)
Chi-square test
N95 50–95 b50
21 (55.3%) 17 (44.72%) –
12 (31.6%) 21 (55.3%) 5 (13.1%)§
χ2 = .019*
10 b10
38 (100%) –
32 (84.2%) 6 (15.8%)§
χ2 = .011*
FER assessment CTRL (n = 38) mean (±SD)
FS’ (n = 28) mean (±SD)
Mann–Whitney test
FEMT Happiness (max 8) Sadness (max 8) Fear (max 8) Disgust (max 8) Anger (max 8) Total score (max 40)
7.97 (±0.2) 6.53 (±1.2) 7.11 (±1.0) 7.21 (±1.0) 4.63 (±1.2) 33.47 (±2.6)
7.64 (±0.7) 5.46 (±1.6) 5.75 (±1.7) 6.11 (±1.3) 3.68 (±1.6) 28.64 (±3.8)
U U U U U U
= = = = = =
412.0; p 328.5; p 264.0; p 259.5; p 333.5; p 165.0; p
= .006* = .007* b .0001* b .0001* = .009* b .0001*
FELT Happiness (max 5) Sadness (max 5) Fear (max 5) Disgust (max 5) Anger (max 5) Total score (max 25)
5.00 (±0.0) 4.74 (±0.5) 4.84 (±0.5) 4.37 (±0.7) 4.08 (±0.4) 23.03 (±1.0)
5.00 (±0.0) 4.18 (±1.0) 4.25 (±1.1) 3.61 (±1.3) 4.21 (±0.5) 21.25 (±2.4)
U U U U U U
= = = = = =
532.0; p 353.0; p 355.5; p 349.0; p 599.5; p 290.0; p
= 1.00 = .006* = .002* = .010* = .230 = .001*
IQ and Face Identity Matching score distributions among the two groups differed significantly (*), so 10 subjects from the FS group (§) were excluded from further analysis on FER.
214
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
recognition ability is supported by results of a preliminary analysis (see Table s-2, online supplementary material) indicating that the subjects excluded from subsequent analysis had significantly lower performance in the FEMT than the subjects with intact visuoperceptual abilities (p b .003; Mann–Whitney test). Also, FELT score was lower in the children with visuoperceptual deficits compared with subjects with intact visuoperceptual abilities, but this did not reach significance (p = .302). Therefore, only the FS subgroup (FS’ group; n = 28) with intact visuoperceptual abilities was further compared with the CTRL group. Neither age at the time of testing (one-way ANOVA) nor gender/handedness differed significantly (chi-square test) between the FS’ group (n = 28) and controls (n = 38). Intelligence quotient did not differ between the FS’ group (mean 90.5 ± 10.9) and the CTRL group (mean 94.1 ± 8.7) (one-way ANOVA, p = .148). No time difference in completing the two emotion-related tasks was found between controls (mean = 449 ± 128 s) and the FS’ group (mean = 427 ± 139 s). The FS’ group achieved a lower recognition score than the control group in both matching (mean ± SD: 28.64 ± 3.8 vs 33.47 ± 2.6; p b .0001) and labeling (21.25 ± 2.4 vs 23.03 ± 1.0; p = .001) facial emotions. Correlation analyses did not reveal a significant relationship between IQ scores and either FEMT (Spearman's rho = 0.281; p = .148) or FELT (Spearman's rho = 0.185; p = .347). On the contrary, we found a significant correlation between matching and labeling performance (Spearman rho = 0.625; p b .0005), supporting the validity of the two measures in assessing FER ability. At the level of each single emotion, the FS’ group showed significantly lower scores compared with controls in matching all of the five basic emotions (see Table 2 and Fig. 1A), whereas they were impaired in labeling three out of the five facial expressions: sadness (4.18 vs 4.74; p = .006), fear (4.25 vs 4.84; p = .002), and disgust (3.61 vs 4.37; p = .010) (see Table 2 and Fig. 1B). Even though the FS’ group's performance was significantly worse than that of the CTRL group across all emotions in the FEMT, the largest area under the ROC curve was obtained for disgust and fear, with an asymptotic significance of below 0.001 (Fig. s-3). Linear regression did not reveal any correlation of FEMT and FELT scores with clinical variables, such as (a) age at first FS, (b) number of FS episodes, (c) presence of prolonged and/or complex FSs, and (d) abnormal EEG within 48 h since FSs.
Fig. 2. Distributions of FEMT and FELT scores. Frequency distributions of subjects as a function of facial emotion recognition scores (scores on the y-axis) for FEMT (A) and FELT (B). In both histograms, controls are represented on the left and the FS’ group on the right. The horizontal dashed lines indicate the cutoff of two standard deviations from the controls' mean.
Fig. 1. Facial emotion recognition ability across the five basic emotions. Performance of the FS’ group and controls in matching (A) and labeling (B) each of the five basic emotions, expressed as the mean amount of correct responses (±standard error). Ordinate axis is set in both absolute number and percentage. Asterisks correspond to significant comparisons (Mann–Whitney test; p ≤ .01).
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
Finally, we explored differences in clinical features at a single subject level. Twelve patients (31.6%) were impaired in both FEMT and FELT, while four (10.5%) were impaired only in matching and another four exclusively in labeling (Fig. 2). These patients did not differ in any clinical variables from the remaining subjects of the FS’ group (age: p = .344; gender: p = .549; handedness: p = .353; age at first FS: p = .246; number of FS episodes: p = .760; presence of complex FSs: p = .847; presence of prolonged FSs: p = .393; presence of EEG abnormalities: p = .234). 4. Discussion The main finding of our study is that school-aged healthy children with antecedent FSs show lower scores in recognition of basic emotions compared with age-matched children without FSs. This result raises the question of whether FSs can alter long-term plasticity in extrahippocampal limbic regions, such as the amygdala, insular cortex, and other cortical–subcortical structures involved in emotion recognition. The demonstration of subtle emotion recognition deficits in children with antecedent FSs can be relevant both to the field of clinical epileptology and to emotion research. Thus, before discussing the significance of our results, some important methodological issues and limitations of our study should be taken into account. A first point concerns patient selection. Indeed, severe cases might have been overrepresented in our FS group because of the selection bias of a hospital-based series. Furthermore, the low participation rate of eligible subjects can also generate a selection bias. However, both these biases would affect the epidemiological value of the study (that is beyond the aim of our study). On the contrary, we consider that the intrinsic validity of the study, demonstrating FER deficits even in subjects with a normal IQ and face identity recognition, is scarcely affected by these biases. This is even more plausible considering that the distribution of clinical features in our sample was not exceptional. In fact, although our work was not designed to obtain epidemiological data, the clinical features of our FS group did not differ from the ones provided by populationbased cohort studies [48,49]. A second point concerns the ‘specificity’ of the observed impairments in FER. Because of the limited time made available by the children's parents, we could not perform a comprehensive examination of neuropsychological functions. Thus, we were forced to limit our control tasks to a general intelligence assessment and a visuoperceptual task employing the same type of stimuli as the FER tasks. Indeed, 13.2% of the children in the FS group had an IQ below the 50th percentile, and 16% showed some mistakes in matching facial identity. Aiming to overcome this possible and relevant bias, we then restricted FER analysis only to children with a normal IQ and intact face recognition abilities; yet, these children still showed a significant impairment in FER. Furthermore, correlation analysis of CPM and FER scores revealed that a general intelligence measure could not predict performance in evaluating facial expressions. Hence, we are quite confident that lower scores in both the FEMT and FELT are related to specific difficulties in emotion recognition rather than to other more general cognitive impairments. A third point concerns the fact that none of our patients underwent a brain MRI in order to assess medial temporal lobe structures and possible FS-induced amygdala–hippocampal damage. However, there was no clinical indication to perform MRI in this group of children [39] and, because of ethical issues, a brain MRI study has not been proposed. Nevertheless, our data demonstrate that school-aged children with antecedent FSs have difficulties in matching and labeling emotional expressions compared with an age-matched control group from the same sociocultural background. These results support the hypothesis that FSs can be associated with specific cognitive function deficits. To the best of our knowledge, only one study has addressed specific medial temporal lobe functions, including learning, memory retrieval, and consolidation, in school-aged children with a history of FSs [17]. These authors
215
observed that the occurrence of FS before 1 year of age was associated with deficits in memory functions. Furthermore, MRI studies detected hippocampal abnormalities not only in children after complex FSs [9] but also in adults with a history of simple FSs without subsequent epilepsy [50]. Morphological brain changes have also been demonstrated in animal models including FSinduced neuroanatomical changes (such as mossy fiber sprouting, loss of inhibitory interneurons, and persistent granule cell ectopia) are associated with impairment in functions subserved by medial temporal lobe structures, i.e., learning and memory [51,52]. Although we do not provide MRI evidence of medial temporal lobe (or of amygdala) damage in the FS group, it is possible to consider that the neural network involved in emotion processing is somewhat dysfunctional in these children. It is important to note that the difference observed between the FS and CTRL groups, while significant, was relatively small. However, FER deficits were quite common among children in the FS group. Indeed, 31.6% of children were impaired in both matching and labeling facial expressions. As far as the recognition of specific emotions is concerned, we observed that in the FS group, disgust and fear were the two emotions for which errors were present in both tasks (labeling and matching). However, the emotion-matching task appeared particularly difficult for children in the FS group, with deficits observed in matching every emotion category, including happiness. Recently, we reported FER abilities in a large cohort of 140 adult patients with chronic MTLE evaluated across a 5-year period [30]. This study confirmed that widespread deficits in the recognition of negative emotions are common in this seizure disorder. In particular, patients with early onset of seizures, including FSs, and bilateral damage are severely impaired in emotion recognition. Indeed, emotion recognition deficits seem to be clinically relevant and severe in patients suffering from early-life amygdala damage [53,54] and in adult patients with temporal lobe epilepsy with seizure onset during childhood, including FSs [23,24,55–57]. Overall, these findings raise the important question of whether a critical period of life exists for establishing the neural network underlying FER abilities [23]. Notably, Golouboff and coworkers [33] assessed this question in young patients with MTLE, documenting that FER deficits are already present during childhood and adolescence in patients with early-onset MTLE. These authors underlined that early seizure onset (including FSs) was associated with poor recognition of facial expression, especially for fear. In this context, the results of our study suggest that FSs, even in the absence of overt epilepsy, are sufficient to cause subtle emotion recognition deficits, thus strengthening the view that FSs are a marker of medial temporal lobe dysfunction. Clearly, we cannot demonstrate the putative causative role of FSs on the long-term plasticity of the amygdala and interconnected areas subserving emotion processing. Indeed, it could be possible that FER impairments and FSs are the markers of a preexisting damage within the temporal lobe. In conclusion, we demonstrated that emotion recognition abilities could be defective in school-aged children with a history of FSs, even in children with an isolated simple FS. This deficit was quite common in our patient population, even if it was not clinically relevant in daily life. Because of the limited sample size, our results need to be confirmed in longitudinal studies and/or larger populations in order to establish their specificity and sensitivity. In particular, the use of a more comprehensive neuropsychological test battery would allow us to find out if other functions are altered in children with FSs, in addition to FER deficit. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.yebeh.2013.07.007. Acknowledgments Dr. Meletti received a research grant from the Emilia Romagna Region (Grant Innovative Research Area 1A) and from CaRisMo Foundation.
216
G. Cantalupo et al. / Epilepsy & Behavior 29 (2013) 211–216
References [1] Baram TZ, Shinnar S. Febrile seizures. Academic Press; 2002. [2] Nelson KB, Ellenberg JH. Predictors of epilepsy in children who have experienced febrile seizures. N Engl J Med 1976;295(19):1029–33. [3] Kira R, Ishizaki Y, Torisu H, Sanefuji M, Takemoto M, Sakamoto K, et al. Genetic susceptibility to febrile seizures: case–control association studies. Brain Dev 2010;32(1):57–63. [4] Baulac S, Gourfinkel-An I, Nabbout R, Huberfeld G, Serratosa J, Leguern E, et al. Fever, genes, and epilepsy. Lancet Neurol 2004;3(7):421–30. [5] Marini C, Mantegazza M. Na+ channelopathies and epilepsy: recent advances and new perspectives. Expert Rev Clin Pharmacol 2010;3(3):371–84. [6] Lado FA, Laureta EC, Moshé SL. Seizure-induced hippocampal damage in the mature and immature brain. Epileptic Disord 2002;4(2):83–97. [7] VanLandingham KE, Heinz ER, Cavazos JE, Lewis DV. Magnetic resonance imaging evidence of hippocampal injury after prolonged focal febrile convulsions. Ann Neurol 1998;43(4):413–26. [8] Scott RC, Gadian DG, King MD, Chong WK, Cox TC, Neville BGR, et al. Magnetic resonance imaging findings within 5 days of status epilepticus in childhood. Brain 2002;125(Pt 9):1951–9. [9] Scott RC, King MD, Gadian DG, Neville BGR, Connelly A. Hippocampal abnormalities after prolonged febrile convulsion: a longitudinal MRI study. Brain 2003;126(Pt 11): 2551–7. [10] Scott RC, King MD, Gadian DG, Neville BGR, Connelly A. Prolonged febrile seizures are associated with hippocampal vasogenic edema and developmental changes. Epilepsia 2006;47(9):1493–8. [11] Wallace SJ. They don't do very well. Pediatrics 1980;65(3):678–9. [12] Tsai JJ, Huang MC, Lung FW, Huang CC, Change YC. Differences in factors influencing the familial aggregation of febrile convulsion in population and hospital patients. Acta Neurol Scand 1996;94(5):314–9. [13] Ellenberg JH, Nelson KB. Febrile seizures and later intellectual performance. Arch Neurol 1978;35(1):17–21. [14] Verity CM, Butler NR, Golding J. Febrile convulsions in a national cohort followed up from birth. II—medical history and intellectual ability at 5 years of age. Br Med J 1985;290(6478):1311–5. [15] Verity CM, Greenwood R, Golding J. Long-term intellectual and behavioral outcomes of children with febrile convulsions. N Engl J Med 1998;338(24):1723–8. [16] Chang YC, Guo NW, Huang CC, Wang ST, Tsai JJ. Neurocognitive attention and behavior outcome of school-age children with a history of febrile convulsions: a population study. Epilepsia 2000;41(4):412–20. [17] Chang YC, Guo NW, Wang ST, Huang CC, Tsai JJ. Working memory of school-aged children with a history of febrile convulsions: a population study. Neurology 2001;57(1):37–42. [18] Adolphs R. Neural systems for recognizing emotion. Curr Opin Neurobiol 2002;12(2): 169–77. [19] Pitkänen A, Tuunanen J, Kälviäinen R, Partanen K, Salmenperä T. Amygdala damage in experimental and human temporal lobe epilepsy. Epilepsy Res 1998;32(1–2): 233–53. [20] Aggleton JP. The contribution of the amygdala to normal and abnormal emotional states. Trends Neurosci 1993;16(8):328–33. [21] LeDoux JE. Emotion circuits in the brain. Annu Rev Neurosci 2000;23:155–84. [22] LeDoux J. The amygdala. Curr Biol 2007;17(20):R868–74. [23] Meletti S, Benuzzi F, Rubboli G, Cantalupo G, Stanzani-Maserati M, Nichelli P, et al. Impaired facial emotion recognition in early-onset right mesial temporal lobe epilepsy. Neurology 2003;60(3):426–31. [24] Benuzzi F, Meletti S, Zamboni G, Calandra-Buonaura G, Serafini M, Lui F, et al. Impaired fear processing in right mesial temporal sclerosis: a fMRI study. Brain Res Bull 2004;63(4):269–81. [25] Reynders HJ, Broks P, Dickson JM, Lee CE, Turpin G. Investigation of social and emotion information processing in temporal lobe epilepsy with ictal fear. Epilepsy Behav 2005;7(3):419–29. [26] Yamada M, Murai T, Sato W, Namiki C, Miyamoto T, Ohigashi Y. Emotion recognition from facial expressions in a temporal lobe epileptic patient with ictal fear. Neuropsychologia 2005;43(3):434–41. [27] Schacher M, Winkler R, Grunwald T, Kraemer G, Kurthen M, Reed V, et al. Mesial temporal lobe epilepsy impairs advanced social cognition. Epilepsia 2006;47(12): 2141–6. [28] Shaw P, Lawrence E, Bramham J, Brierley B, Radbourne C, David AS. A prospective study of the effects of anterior temporal lobectomy on emotion recognition and theory of mind. Neuropsychologia 2007;45(12):2783–90.
[29] Walpole P, Isaac CL, Reynders HJ. A comparison of emotional and cognitive intelligence in people with and without temporal lobe epilepsy. Epilepsia 2008;49(8):1470–4. [30] Meletti S, Benuzzi F, Cantalupo G, Rubboli G, Tassinari CA, Nichelli P. Facial emotion recognition impairment in chronic temporal lobe epilepsy. Epilepsia 2009;50(6): 1547–59. [31] Bonora A, Benuzzi F, Monti G, Mirandola L, Pugnaghi M, Nichelli P, et al. Recognition of emotions from faces and voices in medial temporal lobe epilepsy. Epilepsy Behav 2011;20(4):648–54. [32] Broicher SD, Kuchukhidze G, Grunwald T, Krämer G, Kurthen M, Jokeit H. “Tell me how do I feel”—emotion recognition and theory of mind in symptomatic mesial temporal lobe epilepsy. Neuropsychologia 2012;50(1):118–28. [33] Golouboff N, Fiori N, Delalande O, Fohlen M, Dellatolas G, Jambaqué I. Impaired facial expression recognition in children with temporal lobe epilepsy: impact of early seizure onset on fear recognition. Neuropsychologia 2008;46(5):1415–28. [34] Adolphs R. The social brain: neural basis of social knowledge. Annu Rev Psychol 2009;60:693–716. [35] Ekman P, Freisen WV, Ancoli S. Facial signs of emotional experience. J Pers Soc Psychol 1980;39(6):1125–34. [36] Ekman P. Expression and the nature of emotion. In: Scherer K, Ekman P, editors. Approaches to emotion. Hillsdale, New Jersey: Lawrence Erlbaum; 1984. p. 319–44. [37] Commission on Epidemiology and Prognosis, International League Against Epilepsy. Guidelines for epidemiologic studies on epilepsy. Epilepsia 1993;34(4):592–6. [38] Berg AT, Shinnar S. Complex febrile seizures. Epilepsia 1996;37(2):126–33. [39] Capovilla G, Mastrangelo M, Romeo A, Vigevano F. Recommendations for the management of “febrile seizures”: Ad Hoc Task Force of LICE Guidelines Commission. Epilepsia 2009;50(Suppl. 1):2–6. [40] Raven JC. CPM: Coloured Progressive Matrices. Firenze: O.S. Organizzazioni Speciali; 1996. [41] Ekman P, Friesen WV. Pictures of Facial Affect. Palo Alto, CA: Consulting Psychologists Press; 1976. [42] O'Sullivan M, Ekman P. Facial expression recognition and emotional intelligence. In: Leeland KB, editor. Face recognition: new research. New York, NY: Nova Science Publishers; 2008. p. 25–40. [43] Ekman P. Facial expressions of emotion: an old controversy and new findings. Philos Trans R Soc Lond B Biol Sci 1992;335(1273):63–9. [44] Darwin C. The expression of the emotions in man and animals. New York: Philosophical Library; 1955. [45] Rapcsak SZ, Galper SR, Comer JF, Reminger SL, Nielsen L, Kaszniak AW, et al. Fear recognition deficits after focal brain damage: a cautionary note. Neurology 2000;54(3): 575–81. [46] Vicari S, Reilly JS, Pasqualetti P, Vizzotto A, Caltagirone C. Recognition of facial expressions of emotions in school-age children: the intersection of perceptual and semantic categories. Acta Paediatr 2000;89(7):836–45. [47] Durand K, Gallay M, Seigneuric A, Robichon F, Baudouin J-Y. The development of facial emotion recognition: the role of configural information. J Exp Child Psychol 2007;97(1):14–27. [48] Verity CM, Golding J. Risk of epilepsy after febrile convulsions: a national cohort study. BMJ 1991;303(6814):1373–6. [49] Berg AT, Shinnar S, Darefsky AS, Holford TR, Shapiro ED, Salomon ME, et al. Predictors of recurrent febrile seizures. A prospective cohort study. Arch Pediatr Adolesc Med 1997;151(4):371–8. [50] Auer T, Barsi P, Bone B, Angyalosi A, Aradi M, Szalay C, et al. History of simple febrile seizures is associated with hippocampal abnormalities in adults. Epilepsia 2008;49(9): 1562–9. [51] Dubé CM, Zhou J-L, Hamamura M, Zhao Q, Ring A, Abrahams J, et al. Cognitive dysfunction after experimental febrile seizures. Exp Neurol 2009;215(1):167–77. [52] Koyama R, Tao K, Sasaki T, Ichikawa J, Miyamoto D, Muramatsu R, et al. GABAergic excitation after febrile seizures induces ectopic granule cells and adult epilepsy. Nat Med 2012. http://dx.doi.org/10.1038/nm.2850. [53] Calder AJ. Facial emotion recognition after bilateral amygdala damage: differentially severe impairment of fear. Cogn Neuropsychol 1996;13(5):699–745. [54] Hamann SB, Stefanacci L, Squire LR, Adolphs R, Tranel D, Damasio H, et al. Recognizing facial emotion. Nature 1996;379(6565):497. [55] Anderson AK, Spencer DD, Fulbright RK, Phelps EA. Contribution of the anteromedial temporal lobes to the evaluation of facial emotion. Neuropsychology 2000;14(4): 526–36. [56] Adolphs R, Tranel D, Damasio H. Emotion recognition from faces and prosody following temporal lobectomy. Neuropsychology 2001;15(3):396–404. [57] McClelland III S, Garcia RE, Peraza DM, Shih TT, Hirsch LJ, Hirsch J, et al. Facial emotion recognition after curative nondominant temporal lobectomy in patients with mesial temporal sclerosis. Epilepsia 2006;47(8):1337–42.