Table 2.

Studies examining emotion recognition in relation to CU traits

Study	Participants			CU measure (informant)	Emotion recognition			Main outcome specific to emotion recognition
	N (female)	Sample type	Age range (M)		Type of stimuli	Emotions	Measurement
Aghajani et al. (2021)	n = 81 (–) CD/LPE+  = 19 CD/LPE- = 31 *22% comorbid ADHD CG = 31	Forensic Community	15–19 (M CD/LPE ±  = 16.84 M HC = 17.02)	LPE proxy from the Remorseless, Callousness and Unemotionality subscales of the YPI (SR) CD/LPE+  = two or more items > 5	Facial	Fear Sad	RT and accuracy (% of correct attribution)	Controlling for age and IQ, but not ADHD: 1. There were no significant differences across groups in ER as measured by RT and % of correct attribution of other’s emotions
Bennett and Kerig (2014)	n = 417 (111) Primary CU (1) = 55 Secondary CU (2) = 76 Low CU (3) = 279	Forensic	12–18 (16.5) (M1 = 15.85 M2 = 16.06 M3 = 16.26)	ICU Total score (SR) CU > 26 CG < 26	Facial	Fear Anger Sad Shame Disgust	Accuracy (% of correct responses)	1. The acquired (secondary) CU group were more accurate in recognizing others disgust 2. Membership in the primary CU group was less than the acquired group as accuracy increased for recognition of disgust, but was more likely as accuracy increased for shame 3. In relation to the Low CU group, the primary CU were more accurate in the recognition of anger. There were no differences with the acquired CU group
Dadds et al. (2018)	n = 364 (102) *53% ODD/CD; 35.6% ADHD	Clinical	3–16 (8.93)	APSD/SDQ CU derived measure (Combined PR, TR, SR)	Facial	Happy Sad Anger Fear Disgust Neutral	Accuracy (Total score)	Controlling for the effects of ADHD medication: 1. Low CU traits were related with more accurate ER 2. The relationship between CU traits and ER differed according to maltreatment history; CU traits were associated with poorer recognition of those with zero or negligible history of maltreatment 3. Patterns of moderation effects of maltreatment were inconsistent across subgroups. For the group who self-reported maltreatment, CU traits were related with poor ER in those with lower anxiety 4. Both direct and moderated effects of CU on ER were not limited to fear or sadness. Negative correlations were to some extent evident for all the analyzed emotions, except happiness
De Ridder et al. (2016)	n = 55 (10)	Forensic	NR (14.8)	YPI CU subscale (SR)	Ecological	Anger Distress	Accuracy (Total score)	1. High CU adolescents were as accurate as low CU adolescents in inferring distress and anger in staff members 2. High CU adolescents notably overestimated the general intensity of both anger and distress 3. High CU adolescents perceived more anger in staff members when they were misbehaving, particularly breaking the rules
Ezpeleta et al. (2017)	n = 320 (155) CU+ /ODD+  = 51 CU+/ODD− = 24 CU−/ODD+  = 38 CU−/ODD− = 207 *8.7% comorbid ADHD (CG) = 207	Community	8 (–)	ICU Total score (TR)	Emoticons	Happy Sad Anger Fear Neutral	Accuracy RT	Controlling for sex, SES and comorbidities: 1. Children within the CU−/ODD+ and the CU+/ODD+ were less accurate than the CG in processing the information, specifically when the stimuli expressed happiness, fear or neutral 2. Children high on CU traits but low on ODD performed similarly to the CG with respect to both accuracy and RT with one exception: children high on CU traits responded faster to fear 3. Children within the CU+/ODD+ group differed in accuracy but not in RT in relation to the CG
Kahn et al. (2017)	112 (–)	Forensic	12–20 (15.5)	ICU Total score (SR)	Facial	Happy Sad Anger Disgust Fear Neutral	Accuracy	1. There were no significant main effects for CU traits or interactions with anxiety in predicting overall accuracy in the affective facial recognition task When predicting accuracy for independently identify the 6 analyzed emotions: 2. CU traits were positively associated with fear recognition accuracy at lower levels of anxiety 3. CU traits were related with poorer results in the recognition of disgust, with a trend suggesting that this relationship tend to occur at high levels of anxiety
Kimonis et al. (2016)	n = 214 (100)	At-risk Community	3–6 (4.7)	ICU Total score (24 and 12 items), Callousness and Uncaring subscales (Combined PR, TR) UNSW CU	Facial static and dynamic	Anger Fear Happy Sad Pain (only in dynamic)	Accuracy	Controlling for CP and sociodemographics: 1. Children scoring high on ICU Total and Uncaring subscale scores were less accurate in recognizing anger, fear, happiness, and sadness 2. Callousness was only associated with poor fear and sadness recognition, but not after covaring with uncaring 3. Uncaring remained significantly associated with anger, happy and sad recognition after controlling for callousness 4. There was a significant effect of age, with older children being more accurate in emotion recognition
Klapwijk et al. (2016)	n = 79 (–) ASD = 23 CD/CU+ = 23 *34.78% comorbid ADHD CG = 33	Clinical Forensic Community	12–19 (MASD = 17 MCD/CU+  = 16.6 MCG = 17.1)	ICU Total score (SR) CU+  > 27 (median)	Facial	Anger Fear Neutral	Accuracy Congruency RT	1. The CD/CU+ group reacted faster than the CG 2. There were no differences across groups in the identification of other’s emotional expression 3. The CD/CU+ group reported more congruency for anger as compared to fearful emotions
Kohls et al. (2020a)	n = 1252 (796) CD = 542 (317) CG = 710 (479)	Clinical Forensic Community	9–18 (MCD = 14.4 MCG = 14.0)	YPI CU dimension (SR) ICU (PR)	Facial	Happy Sad Anger Fear Disgust Surprise	Accuracy (% correct)	1. There were higher levels of CU traits in the CD group, with no significant differences across boys and girls 2. Relative to the CG, youths with CD showed impaired emotion recognition, which was related to higher CU traits a more physical and proactive aggression 3. Elevated CU traits within the CD group were associated with overall emotion recognition impairments rather than deficits in particular emotions 3. Parent reported CU traits were not related with emotion recognition within the CD group 4.Overall, girls outperformed boys in emotion recognition
Kohls et al. (2020b)	n = 1252 (796) CD = 542 (317) CG = 710 (479)	Clinical Forensic Community	9–18 (MCD = 14.4 MCG = 14.0)	CU scale and LPE proxy from the Remorseless, Callousness and Unemotionality subscales of the YPI (SR) LPE+  = two or more items > 4	Facial	Happy Sad Anger Fear Disgust Surprise	Accuracy (% correct)	1. The subgroup of youths with CD with an emotion recognition deficit did not differ significantly from those without such deficit regarding the presence of the LPE specifier 2. Deficits in emotion processing, including emotion recognition, related to CD do not map neatly onto established DSM-5 subtypes, such as the CD+ LPE
Lui et al. (2016)	n = 103 (33)	At-risk	16–18 (16.93)	ICU Total score (SR)	Facial	Happy Sad Anger Disgust Fear Neutral	Accuracy (Total score)	1. CU traits were negatively correlated with emotion recognition 2. Emotion recognition did not mediate the association between CU traits and both affective and cognitive empathy
Martin-Key et al. (2017)	n = 77 (–) CD = 37 (CU+  = 20; CU− = 17) *46% comorbid ADHD CG = 40	Forensic Community	13–18 (M CD = 16.03) M CG = 16.20)	ICU Total score (SR) CD/CU+  > 30 (median split) CD/CU− ≤ 30	Facial dynamic	Happy Anger Sad Disgust Fear Surprise	Accuracy	1. There were no differences between CD groups in terms of ADHD 2. Participants in the CD group were less accurate than CG in the recognition of sadness, fear and disgust 3. Yet, no significant differences were found between CD/CU+ and CD/CU− groups, for any emotion 4. When CU traits were treated dimensionally, no significant correlations were found between CU traits and emotion recognition
Martin-Key et al. (2020)	n = (52) CD = 23 (CU+  = 10 CU− = 13) *13% comorbid ADHD CG = 29	Forensic Community	13–18 (MCD = 16.06 MCG = 16.22)	ICU Total score (SR) CD/CU+  > 28 (median split) CD/CU− ≤ 28	Facial dynamic	Anger Happy Sad Disgust Fear Surprise	Accuracy	1. There were no significant differences in emotion recognition between CD and CG, nor between CD/CU+ and CD/CU− groups 2. When CU traits were treated dimensionally, there were no significant correlations between CU traits in the CD group and emotion recognition for any of the six emotions
Milone et al. (2019)	n = 60 (–)	Clinical	11–17 (13.27)	APSD CU (Combined PR, SR)	Facial (eye region)	Happy Sad	Accuracy	1. The levels of CU traits were not related with abilities in emotion recognition 2. CD males high and low on CU traits did not differ in emotion recognition performance
Moore et al. (2019)	n = 1214 (656) (607 twin pairs, 251 MZ, 356 DZ)	Community	9.1–20.8 (14.1)	ICU Total score, Uncaring/Callous and Unemotional subscales (PR)	Facial	Happy Sad Anger Fear Disgust Surprise	Accuracy: Unbiased hit rate (UBHR)	1. The uncaring/callous dimension was significantly associated with impaired recognition of happiness, sadness, fear, surprise and disgust 2. The unemotional dimension was significantly associated with improved recognition of surprise and disgust 3. The total ICU score was significantly associated with impaired recognition of sadness 4. The relationship between uncaring/callousness and deficits in distress cues recognition (i.e., fear and sadness) was entirely accounted by shared genetic influences
Muñoz (2009)	n = 55 (–)	At-risk	8–16 (11.8)	ICU Total score and subscales (SR)	Facial and body posture	Happy Sad Fear Anger Surprise Disgust (only in faces)	Accuracy	1.Youths who were CU+ traits had fewer correct responses to fear and anger faces 2.Deficit in labelling fear faces and fear body postures among youths scoring highly on CU traits, although accuracy was lower for more types of emotions than fear Deficits specific to fear for CU+  3.Boys who were poor at labelling both the fear faces and fear body postures had the highest levels of total CU traits
O’Kearney et al. (2017)	n = 124 (45) *ODD = 43 (31) CG = 17 (14) *55.4% comorbid ADHD	Clinical Community	4–8 (MODD = 5.9 MCG = 5.9)	ICU Total score (PR) ODD/CU+  ≥ 67% (median split) ODD/CU− < 67%	Facial	Happy Sad Anger Fear	Accuracy	1.The were no significant differences between groups for emotion recognition
O’Kearney et al. (2020)	n = 74 (31) ODD = 74	Clinical	4–8 (MODD = 5.9)	ICU Total score (PR)	Facial	Happy Sad Anger Fear	Accuracy	1.CU had a small negative first-order association with the ability to understand mixed emotions but not with the other emotions abilities 2. Only at lower levels of affect arousal/dysregulation there is a notable association between higher CU and poorer emotion recognition
Pauli et al. (2021)	n = 1263 (536) CD/CU+  = 248 (153) CD/CU− = 230 (130) CG = 785 (523)	Clinical Forensic Community	9–18	ICU Total score (SR) CD/CU+  ≥ 39 (median split) CD/CU− < 29	Facial	Anger Disgust Fear Happy Sad Surprise	Accuracy (% correct)	1.The CD/CU+ group performed significantly worse than de GC group, but the CD/CU− group did not differ from either group 2.Recognition accuracy was better for high-intensity than for low- intensity emotions and GC and CD/CU− groups benefited more from the increased expression intensity than the CD/CU+  3.Group differences in emotion recognition abilities were not driven by difficulties with specific emotions such as fear and sadness 4. Female youths significantly outperformed male youths, as evidenced by a main effect of sex
Peticlerc et al. (2019)	n = 1005 (–) (504 twin pairs, 209 MZ, 295 DZ)	Community	6–7	Four items ad hoc 1. “he/she didn’t seem to feel guilty after misbehaving” 2.” he/she has been insensitive to others' feelings” 3.” his/her emotions appear superficial” 4.” has not kept his/her promises”	Facial	Happy Sad Anger Fear	Accuracy	Controlling for other problems (physical aggression, ADHD and depressive symptoms), and for the recognition of other emotions: 1.There was a relationship between CU traits, measured in kindergarten and first grade, and poor emotional recognition of fear, measured in first grade 2. The relationship between CU traits and deficits in fear recognition was genetically influenced 3. The relationship between CU traits and deficits in the recognition of sadness did not hold after controlling for other problems and the recognition of other emotions. No genetic influence was reflected in this relationship, although evidence was provided for the influence of the non-shared environment
Rehder et al. (2017)	n = 761 (657) EA = 446 CP+ CU = 20 CG = 442 AA = 315 CP+ CU = 16 CG = 291	Community at risk	6–7	ICU Empathic-prosocial and callous subscales (PR)	Facial	Anger Sad Happy Fear	Accuracy (overall and specific emotions)	Controlling for primary caregivers’ years of education and children’s age: 1.Differences among typical children (CG), children with CP-only and children with CP + CU were moderated by child race and family income for overall emotion recognition, and by child race for specific emotion recognition 2. Only among EA children, and children within families with higher income-to-needs ratio (i.e., no extreme poverty), CG showed better accuracy for overall emotion recognition than children in the CP only and CP + CU groups 3. Only among EA children, CG showed better recognition for happy faces than children in the CP only and CP + CU group 4. Children in the CP only and CP + CU groups often did not perform differently on emotion recognition accuracy
Schwenck et al. (2011)	n = 192 (–) ASD = 55 CD = 70 (CD/CU+  = 36 CD/CU− = 34) *44.8% comorbid ADHD CG = 67	Clinical	6–17 (12.3)	ICU Total score (PR) CD/CU+  ≥ 32 (median split) CD/CU− < 32	Facial morphed	Anger Happy Sad Fear Disgust	Accuracy RT	1. There were no significant differences between CD groups (CU+/CU−) in emotion recognition based on RT and the number of correctly identified emotions 2. Removing ADHD children receiving medical treatment did not change the pattern of results
Schwenck et al. (2014)	n = (64) CP = 32 (CP/CU+  = 16; CP/CU− = 16) *62.6% comorbid ADHD CG = 32	Clinical	8–16 (13.23)	ICU Total score (PR) CP/CU+  > 35.5 (median split) CP/CU− < 35.5	Facial morphed	Anger Happy Sad Fear Disgust	Accuracy RT	There were no significant differences in the analyzed variables between children with and without comorbid ADHD Regarding RT: 1. Girls with CP/CU− reacted more slowly than CG to faces developing happy, sad, and fearful expressions 2. Girls with CP/CU+ did not differ from CP/CU− and CG Regarding accuracy: 3. Girls with CP/CU+ recognized fearful expressions better than the other two groups 4. Girls with CP/CU− recognized sad expressions less often than CG 5. Girls with CP/CU+ mistook sad faces more commonly as disgusted faces than CG; whilst girls with CP/CU− identified fearful faces as disgusted more often than the CP/CU+ group
White et al. (2016)	n = 337 (185) (High LC = 107; Low LC- = 230; High PI = 96; Low PI = 241)	At-risk	3–7 (4.82)	MAP-DB Low Concern, Punishment Insensitivity (PR) High/Low groups ≥ /≤ 80th percentile	Facial	Fear Happy Anger Neutral	Accuracy Latency (mean RT)	Controlling for age, sex, temper loss, aggression, impulsivity, and the control block In terms of Accuracy: 1. The High LC group was less accurate in identifying fearful faces than the low LC; the accuracy rates did not differ for angry or happy faces 2. There were no differences between High and Low PI groups 3. Female were more accurately to emotional facial expressions than males In terms of Latency: 3. No significant differences were observed between High/Low LC and PI groups
Wolf and Muñoz (2014)	n = 50 (–)	At-risk	11–16 (14.3)	YPI CU subscale ICU Total score (SR)	Facial and Body postures Dynamic	Fear Anger Happy Disgust Sad Pain	Accuracy	Controlling for age and violent delinquency: 1. Fearful facial and bodily expressions were unrelated with CU traits 2. For the YPI-CU, there were deficits in the recognition of pain in faces, often misidentified as sadness and disgust, and anger in postures, often misidentified as happiness and disgust 3. For the ICU, results showed a significant enhancement in the recognition of anger in faces and disgust in postures
Woodworth and Waschbusch (2007)	n = 73 (14) CP = 32 CP/CU = 24 CG = 18 *84.9% comorbid ADHD	Clinical Community	7–12.78 (9.81)	APSD CU (Combined PR, TR) CU+  ≥ 67 (median split) CU− ≤ 63	Facial	Anger Disgust Fear Happy Sad Surprise	Accuracy	Controlling age, sex, IQ and ADHD 1. Children with higher CU scores were less accurate in labelling sad affect than children with lower CU scores 2. Children with higher CU scores were more accurate in labelling fear than children with lower CU scores, but this main effect was qualified by the CP-CU interaction trend 3. Children with high CP but low CU traits were less accurate than other children in interpreting fearful facial emotions

AA African American, ADHD attention deficit hyperactivity disorder, APSD antisocial process screening device, ASD autism spectrum disorder, CD conduct disorder, CG comparison group, CP conduct problems, CU callous-unemotional, DZ dizygotic, EA European American, ER emotion recognition, GM grandiose-manipulative, HC healthy control, ICU inventory of callous-unemotional traits, INS impulsive-need of stimulation, IQ intelligence quotient, LC low concern, LPE limited prosocial emotions, M media, MAP-DB multidimensional assessment profile of disruptive behavior, MZ monozygotic, NR not reported, ODD oppositional defiant disorder, PI punishment insensitivity, PR parent-reported, RT reaction time, SES socioeconomic status, SDQ strengths and difficulties questionnaire, SR self-reported, TR teacher-reported, UBHR unbiased hit rate, UNSW University of New South Wales, YPI youth psychopathic traits inventory