Facial emotion recognition in adolescent depression: The role of childhood traumas, emotion regulation difficulties, alexithymia and empathy

ABSTRACT

Introduction:

Facial emotion recognition (FER) is crucial for effective social competency, and problems in this skill are linked depression during adolescence. In this study, we aimed to find the rates of FER accuracy for negative (fearful, sad, angry, disgusted), positive (happy, surprised), and neutral emotions, and the possible predictors of FER skill for most confusing emotions.

Subjects and Methods:

A total of 67 drug-naive adolescents with depression (11 boys, 56 girls; 11–17 years) were recruited for the study. The facial emotion recognition test, childhood trauma questionnaire and basic empathy, difficulty of emotion regulation, and Toronto alexithymia scales were used.

Results:

The analysis demonstrated that adolescents have more difficulties in recognizing negative emotions when compared the positive ones. The most confusing emotion is fear (39.8% of fear was recognized as surprise). Boys have lower fear recognition skill than girls and higher childhood emotional abuse, physical abuse, emotional neglect, and difficulty in describing feelings to predict lower fear recognition skill. For sadness recognition skill, emotional neglect, difficulty in describing feelings, and depression severity were the negative predictors. Emotional empathy has a positive effect on disgust recognition skill.

Conclusion:

Our findings demonstrated that impairment of FER skill for negative emotions is associated with childhood traumas, emotion regulation difficulties, alexithymia, and empathy symptoms in adolescent depression.

INTRODUCTION

Understanding emotion signals, noticing and labeling them within self and from others are important skills for interpersonal relationships and are crucial for effective emotion regulation and social competency.[1,2] Conversely, problems in these processes have been linked with emotion dysregulation and increasing vulnerability to depression during adolescence.[3]

To date, a large body of research examined the biased facial emotion detection in adult depression.[4,5] Some of the results demonstrated lower sensitivity for happy emotions and a tendency to judge happiness as neutral, whereas others revealed deficits in identifying neutral emotions and confusing them with others.[6-8] Regarding the ability to identify negative emotions in adult depression, findings are more heterogeneous: Some studies demonstrated improved performance of depressed adults for negative emotions,[6,9] whereas others revealed an equal or even worse performance compared to healthy controls.[10,11] Results are similar for depressed adolescents. In a recent systematic review, three specific recognition deficits associated with adolescent depression: sensitivity to sadness, over-perceiving anger, and under-perceiving happiness.[12] Although previous studies demonstrated impaired recognition accuracy among depressed patients when compared to healthy controls, it is still unclear what are the rates of false recognitions for under-perceiving emotions among depressed adolescents. Also, our knowledge is insufficient about the differences between depressed adolescents who can or cannot recognize the most confusing emotions. To find the possible predictors of FER problems in adolescent depression, we reviewed the literature to date.

Previous studies have shown that several adverse situations including abuse and neglect in childhood can interfere in the process of recognizing emotions and lead to developing difficulties in predicting the consequences of behaviors, understanding and expressing the emotions, as well as in developing empathy and social skills among children and adolescents.[13-15] In a recent systematic review, results showed that abused children have a recognition bias for especially fear and anger expressions, and neglected children have a shorter reaction time for recognition of negative expressions.[13] Due to these results, a potential predictor could be the childhood abuse and neglect experience between the depressive adolescents who can or cannot recognize the emotions.

Several studies have reported that alexithymia, the difficulty in consciously experiencing, identifying, and describing emotions, is associated with deficits in the ability to facial emotion recognition of both positive and negative emotions, which may be linked to problems with empathy, emotion dysregulation, and theory of mind.[16-19] Several studies and meta-analysis have demonstrated a relationship between alexithymia and internalizing/externalizing symptoms among children and adolescents[20,21] Therefore, addressing alexithymia, emotion regulation problems, and empathy seems to be crucial when evaluating the differences between depressed adolescents who can or cannot recognize the emotions.

In the present study, we aimed to find the rates of false recognizing emotions among adolescents with depression. In addition, we aimed to determine whether alexithymia, empathy, and exposure to childhood traumas have a negative effect on the ability to recognize emotions, and if so, in which emotions this effect is more pronounced.

Our hypothesis was:

• Depressed adolescents who cannot recognize the emotions have higher alexithymia and lower empathy scores than others.

• Depressed adolescents who cannot recognize the emotions have more childhood abuse and neglect experience scores than others.

• Childhood abuse, neglect, and alexithymia have higher adverse effect on recognizing negative emotions when compared to positive or neutral ones.

• Empathy has positive effect on FER ability.

MATERIAL AND METHOD

The research was approved by the Ethics Committee of Ufuk University School of Medicine (Ethics Approval Number: 20180516/4). Adolescents who were referred to a psychiatry/child and adolescent psychiatry clinics were 11–18 years old, have normal intelligence levels, and are newly diagnosed with depression according to DSM-5 criteria were invited to the study. If they agreed to participate, written informed consent was obtained from them and their parents. Adolescents who have autism spectrum disorder, chronic medical or neurological disease, mental retardation, ADHD, visual or hearing impairment, treated with medication or therapy methods, or do not want to participate in the study were excluded. The study was carried out in Ufuk University and Gulhane Research and Training Hospital Psychiatry or Child and Adolescent Psychiatry outpatient clinics.

There were 11 boys and 56 girls aged between 11 and 17 years. The mean age of boys and girls was 14.94 ± 1.45; 14.72 ± 1.48, respectively. Demographic information form, the facial emotion recognition test, the basic empathy scale, difficulty of emotion regulation scale (DERS), Toronto alexithymia scale (TAS-20), and childhood trauma questionnaire were used, after permissions.

The facial emotion recognition test

We used Ekman and Friesen’s “Pictures of Facial Affect” test which includes the photos of four male and four female models (total 56 mixed photos) with happy, surprised, fearful, sad, angry, disgusted, and neutral facial expressions.[22] All photos were pressed on a white sheet (210 × 297 mm), and patients were asked to recognize facial emotional expressions within a distance of 45–60 cm. All participants were tested individually in a quiet room. In the beginning, we applied a trial section which composed of one photo for each emotion. In this section, we showed the photos in the same order, get the adolescent’s answer, and then told the true emotion of the photo. In the second part of the test, 49 photos were used. Among these photos, numbers of happy, sad, surprised, fearful, disgusted, angry, and neutral expressions were equal, overall ensuring that participants did not become familiarized to one specific emotional category. In this section, we did not give feedback to adolescents regarding the appropriateness of their response.

The basic empathy scale

This five-point Likert-type scale was developed by Jolliffe and Farrington (2006) to measure the cognitive and emotional empathy levels in the framework of four basic emotions (fear, sadness, anger, and happiness). It contains nine items for cognitive empathy and 11 items for emotional empathy. The Turkish validity and reliability study was completed on adolescents by Topçu et al. (2010).[23,24]

The difficulty of emotion regulation scale (DERS)

It was developed by Gratz and Roemer (2004).[25] The scale is composed of 36 items which are rated on a Likert-type scale, from 1 (almost never) to 5 (almost always). It includes six subscales which cover major dimensions of affect regulation from awareness to expression. The names of the subscales are: 1-Awareness: lack of awareness of emotional responses, 2-Clarity: lack of clarity of emotional responses, 3-Nonacceptance: nonacceptance of emotional responses, 4-Strategies: limited access to effective strategies, 5-Impulse: difficulties in controlling impulses when experiencing negative affect, and 6- Goals: difficulties in engaging goal-directed behavior when experiencing negative affect. The Turkish validity–reliability study was completed by Rugancı et al.[25]

Toronto alexithymia scale (TAS-20)

The scale was developed by Bagby, Parker, and Taylor.[26,27] It is a Likert-type scale from 1 = strongly disagree to 5 = strongly agree. It has three subscales: 1-Difficulty identifying feelings (7 items), 2-Difficulty describing feelings (5 items), and 3-Externally oriented thinking (8 items). The Turkish validity–reliability study was completed by Gülec et al.[28]

Beck depression inventory

BDI is a 21-item self-report inventory for the measurement of emotional, somatic, cognitive, and motivational symptoms in depression.[29,30] Items in the scale are rated between 0 and 3. The purpose of the scale is to determine the severity of depressive symptoms. The highest score is 63. The Turkish reliability and validity study was completed.[31]

Childhood Trauma Questionnaire (CTQ). It is a 28-item self-report instrument developed by Bernstein for evaluating emotional, physical, and sexual abuse and physical and emotional neglect during childhood.[32] The Turkish reliability and validity study was completed.[33]

Statistical analysis

All statistical analyses were performed using the Statistical Package for Social Sciences for Windows (SPSS) version 22.0. We used haphazard sampling model. Demographic information was analyzed through descriptive statistics. Chi-square test was used for categorical variables. Kolmogorov–Smirnov test was used to test for normality. Descriptive analyses of FER scores of emotions were presented using means, standard deviations, and medians. We measured the correct answers for all emotions independently. Also we measured the displaced false answers ratio (e.g., how many fear emotions recognized as disgust?). Since the FER scores were normally distributed, the one-sample t-test was used to compare the differences of FER scores between emotion groups. A post hoc power analysis was conducted, and we found a moderate effect size. A multiple linear regression model was used to identify independent predictors of FER scores for most confusing emotions (fear, sadness anger, and disgust). The model fit was assessed using appropriate residual and goodness-of-fit statistics. A 5% type-I error level was used to infer statistical significance.

RESULTS

When we analyzed the accurate recognition scores for each emotion (the scores link between zero and seven), we found that some adolescents could not recognize any of the seven pictures of fear, sadness, anger, and disgust. Scores were significantly lower among these emotions when compared with happiness, surprise, and neutral. When we grouped the emotions as negative (fear, sadness, anger, and disgust) and positive (happiness, surprise), we found that adolescents have significantly impaired in recognition of negative emotions. The details of Mean ± SD, medians, and one-sample t-test significance scores for the emotions are presented in Table 1.

Table 1.

Accurate recognition scores of facial emotions

Accurate Recognition Scores	Mean±SD	Min-Max	Median	One-Sample T-Test, P
Negative Emotions
Fear	3.2±1.6	0-7	3	P <0.001 for fear- sadness, happiness, surprise, disgust, anger, neutral
Sadness	4.1±2.0	0-7	4	P <0.001 for sadness- happiness, anger, disgust, surprise, neutral; P=0.001 for sad–fearful
Anger	5.3±1.4	0-7	6	P <0.001 for anger- sadness, happiness, surprise, fear, neutral; P=0.760 for anger–disgust
Disgust	5.3±1.4	0-7	6	P <0.001 for disgust- sadness, happiness, surprise, fear, neutral; P=0.760 for disgust–anger
Positive Emotions
Happiness	6.8±0.5	3-7	7
Surprise	6.4±0.8	3-7	7
Neutral	6.3±0.8	4-7	7

Effect size d=0.4061005

When we explore the rates of false recognitions for each emotion, we found interesting results: The most confusing emotion was fear; 39.8% of fear was recognized as surprise. It means that there was a strong tendency to perceive a negative emotion as a positive one. On the other hand, we did not find such a tendency for other negative emotions. The second most confusing negative emotion was sadness. 16.6% of sadness was recognized as neutral, while 10.8% was recognized as disgust. The other important confusion was between two negative emotions: anger and disgust. 8.1% of anger was recognized as disgust, while 10.2% of disgust was recognized as anger. You can see the details in Table 2.

Table 2.

Rates of false recognitions between emotions

	Sadness	Happiness	Anger	Disgust	Surprise	Neutral	Fear
Sadness	-	0.2%	3.6%	10.8%	2.3%	16.6%	6.6%
Happiness	0.2%	-	0.4%	0.2%	0.6%	0.4%	0%
Anger	1.06%	0%	-	8.1%	5.7%	4.6%	3.4%
Disgust	1.2%	0%	10.2%	-	1.2%	1.9%	0.4%
Surprise	0.2%	0.6%	0.4%	1.06%	-	0.2%	5.7%
Neutral	1.9%	0.2%	3.4%	1.2%	1.2%	-	0.8%
Fear	1.9%	1.7%	2.3%	6.1%	39.8%	1.2%	-

To find the predictors of FER skill for negative emotions, we used linear regression analyses. Results are as follows:

Fear Recognition Skill: We found that there were negative relationships between fear recognition scores and male gender (Beta = -0.37, P = 0.04), childhood emotional abuse (Beta = -0.42, P = 0.03), physical abuse (Beta = -0.37, P = 0.02), emotional neglect (Beta = -0.37, P = 0.04), and difficulty describing feelings scores (Beta = -0.43, P = 0.03). It means that depressive adolescent boys have lower fear recognition skill than girls. And the adolescents who have higher childhood emotional abuse, physical abuse, emotional neglect, and higher difficulty in describing feelings have lower fear recognition skill [You can see the details in Table 3].

Table 3.

Effects of demographic variables and childhood traumas, empathy, alexithymia, difficulties in emotion regulation scores on “fear” facial emotion recognition by linear regression analyses

	Unstandardized Coefficients		Standardized Coefficients Beta	P	Collinearity Statistics
	B	Std. Error			Tolerance	VIF
First Model
Age	0.122	0.427	0.106	0.783	0.262	3.820
Gender	-0.141	0.178	-0.271	0.449	0.315	3.178
CTQ Scores
Emotional Abuse	-0.409	0.333	-0.553	0.255	0.179	5.581
Physical Abuse	-0.266	0.205	-0.358	0.231	0.479	2.090
Physical Neglect	0.185	0.179	0.436	0.329	0.206	4.849
Emotional Neglect	-0.526	0.268	-1.132	0.085	0.110	9.110
Empathy Scale Scores
Cognitive Empathy	-0.134	0.184	-0.227	0.487	0.374	2.673
Emotional Empathy	0.182	0.174	0.496	0.326	0.162	6.177
Difficulty of Emotion Regulation Scale Scores
Awareness	0.179	0.192	0.417	0.379	0.182	5.494
Clarity	0.064	0.137	0.194	0.653	0.209	4.774
Nonacceptance	-0.257	0.184	-0.636	0.201	0.175	5.712
Strategy	0.108	0.124	0.384	0.410	0.186	5.372
Impulse	-0.223	0.145	-0.822	0.163	0.127	7.850
Goal	-0.060	0.205	-0.125	0.778	0.198	5.054
Toronto Alexithymia Scale Scores
Difficulty Identifying Feelings	0.066	0.126	0.211	0.615	0.224	4.456
Difficulty Describing Feelings	-0.472	0.226	-1.134	0.070	0.123	8.118
Externally Oriented Thinking	0.215	0.231	0.409	0.379	0.188	5.306
Beck Depression Scale Score	0.017	0.065	0.091	0.800	0.299	3.349
Final Model
Gender	-1.457	0.696	-0.376	0.048	0.782	1.279
Emotional Abuse	-0.313	0.137	-0.423	0.033	0.732	1.365
Physical Abuse	-0.280	0.120	-0.377	0.029	0.963	1.038
Emotional Neglect	-0.176	0.084	-0.379	0.047	0.778	1.286
Difficulty Describing Feelings	-0.180	0.079	-0.434	0.032	0.702	1.425

Sadness Recognition Skill: Similarly, childhood emotional neglect and difficulty in describing feelings have a negative effect on sadness recognition skill (Beta = -0.49, -0.45 and P = 0.003, 0.008; respectively). In addition, higher depression scores decrease sadness recognition (Beta = -0.40, P = 0.01). But interestingly, difficulties in engaging goal-directed behavior when experiencing negative affect scores (Goals subscale score) have a positive effect on this skill (Beta = 0.72, P < 0.0001) [You can see the details in Table 4].

Table 4.

Effects of demographic variables and childhood traumas, empathy, alexithymia, difficulties in emotion regulation scores on “sadness” facial emotion recognition by linear regression analyses

	Unstandardized Coefficients		Standardized Coefficients Beta	p	Collinearity Statistics
	B	Std. Error			Tolerance	VIF
First Model
Age	0.119	0.418	0.098	0.783	0.262	3.820
Gender	0.019	0.174	0.034	0.917	0.315	3.178
CTQ Scores
Emotional Abuse	0.023	0.326	0.030	0.945	0.179	5.581
Physical Abuse	0.125	0.201	0.158	0.551	0.479	2.090
Physical Neglect	0.027	0.175	0.061	0.880	0.206	4.849
Emotional Neglect	-0.178	0.262	-0.362	0.515	0.110	9.110
Empathy Scale Scores
Cognitive Empathy	-0.032	0.180	-0.051	0.864	0.374	2.673
Emotional Empathy	-0.036	0.171	-0.092	0.838	0.162	6.177
Difficulty of Emotion Regulation Scale Scores
Awareness	-0.036	0.188	-0.078	0.854	0.182	5.494
Clarity	-0.026	0.134	-0.075	0.851	0.209	4.774
Nonacceptance	0.127	0.180	0.296	0.502	0.175	5.712
Strategy	0.038	0.121	0.129	0.760	0.186	5.372
Impulse	0.010	0.142	0.035	0.945	0.127	7.850
Goal	0.375	0.201	0.740	0.098	0.198	5.054
Toronto Alexithymia Scale Scores
Difficulty Identifying Feelings	0.069	0.123	0.208	0.590	0.224	4.456
Difficulty Describing Feelings	-0.258	0.221	-0.584	0.278	0.123	8.118
Externally Oriented Thinking	-0.131	0.226	-0.235	0.578	0.188	5.306
Beck Depression Scale Score	-0.054	0.063	-0.274	0.420	0.299	3.349
Final Model
Emotional Neglect	-0.243	0.073	-0.493	0.003	0.771	1.297
Goal	0.368	0.075	0.726	<.0001	0.763	1.311
Difficulty Describing Feelings	-0.202	0.069	-0.458	0.008	0.685	1.459
Depression Score	-0.080	0.032	-0.407	0.019	0.652	1.534

Anger Recognition Skill: Like in sadness, goals score has a positive effect on anger recognition skills (Beta = 0.43, P = 0.02) [You can see the details in Table 5].

Table 5.

Effects of demographic variables and childhood traumas, empathy, alexithymia, difficulties in emotion regulation scores on “anger” facial emotion recognition by linear regression analyses

	Unstandardized Coefficients		Standardized Coefficients Beta	P	Collinearity Statistics
	B	Std. Error			Tolerance	VIF
First Model
Age	-0.156	0.428	-0.142	0.725	0.262	3.820
Gender	-0.039	0.178	-0.078	0.832	0.315	3.178
CTQ Scores
Emotional Abuse	-0.254	0.334	-0.359	0.468	0.179	5.581
Physical Abuse	-0.073	0.206	-0.103	0.731	0.479	2.090
Physical Neglect	0.114	0.179	0.280	0.541	0.206	4.849
Emotional Neglect	-0.400	0.268	-0.899	0.174	0.110	9.110
Empathy Scale Scores
Cognitive Empathy	-0.238	0.185	-0.421	0.233	0.374	2.673
Emotional Empathy	0.204	0.175	0.579	0.276	0.162	6.177
Difficulty of Emotion Regulation Scale Scores
Awareness	-0.046	0.193	-0.111	0.819	0.182	5.494
Clarity	-0.060	0.138	-0.189	0.676	0.209	4.774
Nonacceptance	-0.194	0.185	-0.502	0.323	0.175	5.712
Strategy	-0.172	0.124	-0.639	0.204	0.186	5.372
Impulse	-0.179	0.145	-0.690	0.252	0.127	7.850
Goal	0.269	0.205	0.586	0.227	0.198	5.054
Toronto Alexithymia Scale Scores
Difficulty Identifying Feelings	0.082	0.126	0.275	0.532	0.224	4.456
Difficulty Describing Feelings	-0.123	0.227	-0.310	0.601	0.123	8.118
Externally Oriented Thinking	0.503	0.232	0.998	0.062	0.188	5.306
Beck Depression Scale Score	0.025	0.065	0.143	0.705	0.299	3.349
Final Model
Goal	0.198	0.081	0.433	0.021	1.000	1.000

Disgust Recognition Skill: Emotional empathy has a positive effect on disgust recognition skill (Beta = 0.52, P = 0.01) [You can see the details in Table 6].

Table 6.

Effects of demographic variables and childhood traumas, empathy, alexithymia, difficulties in emotion regulation scores on “disgust” facial emotion recognition by linear regression analyses

	Unstandardized Coefficients		Standardized Coefficients Beta	P	Collinearity Statistics
	B	Std. Error			Tolerance	VIF
First Model
Age	0.330	0.388	0.337	0.419	0.262	3.820
Gender	-0.148	0.161	-0.333	0.384	0.315	3.178
CTQ Scores
Emotional Abuse	-0.270	0.302	-0.428	0.398	0.179	5.581
Physical Abuse	0.222	0.186	0.349	0.268	0.479	2.090
Physical Neglect	0.180	0.162	0.495	0.300	0.206	4.849
Emotional Neglect	0.000	0.243	-0.001	0.999	0.110	9.110
Empathy Scale Scores
Cognitive Empathy	0.332	0.167	0.657	0.083	0.374	2.673
Emotional Empathy	0.145	0.158	0.462	0.386	0.162	6.177
Difficulty of Emotion Regulation Scale Scores
Awareness	-0.057	0.174	-0.155	0.752	0.182	5.494
Clarity	0.274	0.125	0.975	0.059	0.209	4.774
Nonacceptance	-0.026	0.167	-0.076	0.880	0.175	5.712
Strategy	0.061	0.113	0.255	0.603	0.186	5.372
Impulse	0.020	0.131	0.085	0.885	0.127	7.850
Goal	-0.058	0.186	-0.143	0.762	0.198	5.054
Toronto Alexithymia Scale Scores
Difficulty Identifying Feelings	-0.144	0.114	-0.539	0.243	0.224	4.456
Difficulty Describing Feelings	-0.045	0.205	-0.127	0.832	0.123	8.118
Externally Oriented Thinking	-0.073	0.210	-0.162	0.738	0.188	5.306
Beck Depression Scale Score	0.003	0.059	0.018	0.963	0.299	3.349
Final Model
Physical Abuse	0.238	0.122	0.375	0.062	0.816	1.226
Emotional Empathy	0.165	0.060	0.525	0.011	0.816	1.226

DISCUSSION

Our study is one of the first studies that examine the facial emotion recognition (FER) skill in depressed adolescents by addressing childhood traumas, emotion regulation difficulties, empathy, and alexithymia. Our results provided that depressed adolescents have more difficulties in recognizing negative emotions (fear, sadness, anger, and disgust) when compared the positive ones. The most confusing emotion is fear (39.8% of fear was recognized as surprise). Boys have lower fear recognition skill than girls and higher childhood emotional abuse, physical abuse, emotional neglect, and higher difficulty in describing feelings to predict lower fear recognition skill. Emotional neglect and difficulty in describing feelings have a negative effect on sadness recognition skill, too. On the other hand, emotional empathy has a positive effect on disgust recognition skill. In addition, we found a positive relationship between sadness and anger recognition skills and DERS Goals subscale which measures difficulties in engaging goal directed behavior when experiencing negative affect.

In the discussion section, we will address our results in six subheadings in the light of previous studies.

1. Facial Emotion Recognition Accuracy

A review which includes the results of 40 studies about FER in adult depression (2010) demonstrated that depressive individuals have a negative response bias toward sadness; they tend to evaluate positive (happy), neutral, or ambiguous facial expressions as more sad or less happy compared with controls. They also have increased vigilance and selective attention toward sad expressions while having reduced general or emotion-specific recognition accuracy.[5] On the other hand, a recent meta-analysis which examines 22 independent samples (2015) indicated that depressive samples have impaired FER for anger, disgust, fear, happiness, and surprise, but not sadness.[34] In addition, a study which examined the FER skills of adolescents in three groups (major depression, anorexia nervosa and healthy controls) found no FER impairments among clinical groups. Also, they found that anorexia nervosa group was more accurate than the control group in recognizing afraid facial expressions and more accurate than the major depression group in recognizing happy, sad, and afraid expressions.[35] Our findings are consistent with the previous studies which demonstrated impaired recognition accuracy and inconsistent with the negative response bias (increased sensitivity to negative emotions) to facial expressions in depressed individuals. We found that depressed adolescents have significantly decreased FER accuracy in negative emotions, especially for fear and sadness. We did not find an effect of age, inconsistent with recent studies (we included the adolescents between 11 and 17 years). A recent study which assessed 478 children aged 6–16 years (by using the Ekman–Friesen Pictures of Facial Affect and controlled the effect of IQ) demonstrated that FER skills of happiness, surprise, fear, and disgust improve with increasing age, but there was little or no change in sad and angry expressions.[36] Another study that examined associations between psychological and somatic problems and FER in early adolescence (age 12) showed that higher anxiety, depression, and somatization scores were associated with lower FER skill for “angry” while higher for “fearful” expressions. On the other hand, they found no association between mental health problems and FER skills of sad expressions.[37] For future researches, we suggest including both early and mid-late adolescence samples in the same study to understand the effect of age on FER skills of emotions among depressive adolescents.

2. Empathy

As known, accurate recognition of emotional expressions is an initial step to empathic responding. Empathy has two separate types: cognitive and emotional. These dimensions are considered independent; for example, deficient cognitive empathy can coexist with elevated emotional empathy. These independent dimensions may differentially relate to FER skills, but unfortunately, there are a few studies to date which examined this relationship among adolescents with depression or comorbid depressive symptoms. A study with adolescents and young adults examined the types of empathy (emotional emphatic concern vs. empathy quotient which includes cognitive empathy and social skills) and their relationship to expression presentation speed and emotion type. They found that both empathic concern and the empathy quotient were related to accuracy at the brief exposure, but only the empathy quotient’s social skill construct was related to accuracy at the long exposure of emotion expressions. Besides, empathic concern was unrelated to fear recognition, whereas the empathy quotient was highly related.[38] This study provided evidence for the relationship between cognitive empathy and fear recognition skill. In another study with adolescents, the relationship between callous-unemotional traits and cognitive affective empathy were examined by addressing the potential role of affective perspective-taking and facial emotion recognition. Results demonstrated that callous-unemotional traits were negatively associated with cognitive and affective empathy, and the association between callous-unemotional traits and affective empathy was partially mediated by affective perspective-taking, but they did not find a relationship between FER and empathy types.[39] To our knowledge, for the first time, our results demonstrated a positive relationship between emotional empathy and disgust recognition skill among depressed adolescents. To date, many other studies provided that disgust recognition deficit correlated with reduced insula functions in depression.[40-43] Also, insula has a key role in the processing and understanding social emotions and empathizing.[44] It could be possible that in adolescence depression, impairment of the insula is not severe as in adulthood and therapeutic training of emotional empathy for this group could have implications in disgust FER skill. This hypothesis should be examined in future studies.

3. Difficulties in Emotion Regulation

Although the relationship between FER and emotion regulation in depressive patients is not clear in the literature, many studies have shown that depressive patients have insufficient emotion regulation skills; also study results suggested that failures in emotion regulation, when faced with negative affect resulting from the experience of a negative event, may be an important factor in the development of depressive symptoms.[5,34] In this study, we found a positive relationship between sadness and anger recognition skills and DERS Goals subscale which measures difficulties in engaging goal-directed behavior when experiencing negative affect. This could be associated with the relationship between negative response bias (higher recognition in sadness and anger) and the type of emotion regulation difficulties. Supporting this possibility, a recent study with adolescents demonstrated that if an adolescent experienced higher-than-usual sadness or anger on a particular day, then they also experienced higher than usual depressive or aggressive symptoms, and this link mediated with emotion dysregulation.[45] Also, a study demonstrated that greater sadness dysregulation was uniquely and significantly associated with depression and social anxiety but not aggression, whereas greater anger dysregulation was associated with aggressive behavior but not depression and anxiety in a youth sample. These studies and our result provided preliminary evidence for the relationship between sadness and anger FER skills and emotion regulation among adolescents, but larger-sample studies are needed to determine the mechanisms underlying this relationship.[5,9]

4. Alexithymia

Alexithymia accompanies approximately half of depressive patients and is a vulnerability factor for developing mental disorders. Yet, some studies suggested that alexithymia and depression are distinct constructs, and alexithymia predisposes individuals to a negative emotional valence.[46-48] In addition, it has been shown that individuals with alexithymia have inadequate FER skills.[49] In this study, we found a relationship between “difficulty describing feelings,” which is one of the alexithymia subscales, and fear–sadness FER skills in depressive adolescents. This finding supports the view in the literature that alexithymia can play an important role in emotion regulation, incompatible emotion processing, and inadequate thinking ability due to the difficulty in defining and explaining the emotional state of the self and the others.[46,50,51] Due to our results, the relationship between alexithymia and FER skills was limited to fear and sadness, so we could speculate that the mediator role of alexithymia on FER skills of depressed adolescents is particularly related to negative emotions.

5. Childhood Traumas

To date, many studies had been shown that maltreated children and adolescents have deficits in processing emotional expressions as well as presenting difficulties in identifying, expressing, and recognizing emotions. In detail, a recent review demonstrated that abused children have a recognition bias for negative expressions, especially fear and anger while neglected children have a shorter reaction time for them.[13]

Due to our results, emotional neglect has a negative effect on fear and sadness FER skills, while emotional and physical abuse also have a negative effect on fear FER skill. Our results are consistent with the results of studies suggesting that emotional abuse could be a determinant for the development of depression in maltreated children.[52] However, contrary to our findings, in many research, maltreatment types had been associated with the tendency of higher recognition for fear and sadness.[53,54] This inconsistency suggested that depression could be an important predictor in the relationship between maltreatment and FER skills, especially on fear and sadness. It could be possible that increased negative emotion recognition tends to impair interpersonal relationships of maltreated children throughout their development, but on the other hand, it could be a necessity and coping mechanism related to survival in a hostile environment. And the maltreated adolescents who could not use this mechanism would have higher risk for depression. Supporting this hypothesis, an adult study that examined depressive patients and healthy controls with and without childhood maltreatment showed that healthy individuals with a history of childhood maltreatment made significantly more errors in recognizing when compared to healthy nonabused and depressed abused individuals. Also resilient individuals with a history of childhood maltreatment but who have not developed depression show an absence of a fear bias, which may help explain why they do not manifest depressive symptoms, despite their experiences of childhood maltreatment.[55] Also a recent study which examined the effect of violence and neglect on fear learning in young adults demonstrated that threat exposure increased autonomic reactivity while neglect exposure attenuated so adverse experiences may be linked to impairments in fear and safety learning, but the type of impairment may differ with the type of adversity.[56] The impairments in fear recognition should be examined in future adolescent studies comparing with depression and healthy controls.

6. Role of Gender

We found that gender affects only fear recognition skill among adolescents with depression (boys have lower fear recognition skill than girls). This result is inconsistent with the previous studies with nonclinical child and adolescent samples. In 2015, which addressed the development of emotion recognition abilities through childhood and adolescence, by testing the effect of age, pubertal stage and gender found that girls have more accuracy in recognizing surprise, disgust, and anger but not fear.[36] This difference may be related to the effect of depression on FER ability. A recent meta-analysis demonstrated that depressed participants were less accurate in recognizing facial expressions of emotion when depression was more severe.[57] In our sample, we did not find an effect of depression severity on FER abilities, but our sample consisted of adolescents who had been diagnosed with depression by a psychiatrist and were scheduled for treatment so it is possible to detect a different gender effect when compared with the normal population samples or samples with subthreshold symptoms. Supporting this possibility, in a recent study, fear FER accuracy was found to be higher in patients with more severe depression who had suicidal ideation compared to those who did not.[58] In order to better understand the gender effect on fear FER accuracy, we think that case–control studies are needed to compare it with the control group.

Strengths and limitations

Conducting the research in a drug-naive adolescent sample and evaluation of FER for negative positive and neutral emotions are the strengths of our study. But collecting the data by self-report scales, using the haphazard sampling method, moderate effect size, and the absence of a healthy control group are the limitations of the study. Also computerized test could be more appropriate than manual for recognizing FER.

CONCLUSION

In conclusion, our study demonstrated that depressive adolescents have higher FER impairment in negative emotions when compared to positive and neutral ones. The most confusing emotion is fear, approximately one in third fear emotion recognized as surprise, in other words, a negative emotion confused with a positive one. Childhood adversities and alexithymia symptoms have a negative effect on fear and sadness FER skill in depressed adolescents. Therefore, more follow-up studies are required to understand the etiology of FER problems in adolescents.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.