Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities

Ruth Pauli; Gregor Kohls; Peter Tino; Jack C. Rogers; Sarah Baumann; Katharina Ackermann; Anka Bernhard; Anne Martinelli; Lucres Jansen; Helena Oldenhof; Karen Gonzalez-Madruga; Areti Smaragdi; Miguel Angel Gonzalez-Torres; Iñaki Kerexeta-Lizeaga; Cyril Boonmann; Linda Kersten; Aitana Bigorra; Amaia Hervas; Christina Stadler; Aranzazu Fernandez-Rivas; Arne Popma; Kerstin Konrad; Beate Herpertz-Dahlmann; Graeme Fairchild; Christine M. Freitag; Pia Rotshtein; Stephane A. De Brito

doi:10.1007/s00787-021-01893-5

Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities

Ruth Pauli, Gregor Kohls, Peter Tino, Jack C. Rogers, Sarah Baumann, Katharina Ackermann, Anka Bernhard, Anne Martinelli, Lucres Jansen, Helena Oldenhof, Karen Gonzalez-Madruga, Areti Smaragdi, Miguel Angel Gonzalez-Torres, Iñaki Kerexeta-Lizeaga, Cyril Boonmann, Linda Kersten, Aitana Bigorra, Amaia Hervas, Christina Stadler, Aranzazu Fernandez-RivasArne Popma, Kerstin Konrad, Beate Herpertz-Dahlmann, Graeme Fairchild, Christine M. Freitag, Pia Rotshtein, Stephane A. De Brito

Department of Psychology

Research output: Contribution to journal › Article › peer-review

4 Citations (SciVal)

Abstract

Conduct disorder (CD) with high levels of callous-unemotional traits (CD/HCU) has been theoretically linked to specific difficulties with fear and sadness recognition, in contrast to CD with low levels of callous-unemotional traits (CD/LCU). However, experimental evidence for this distinction is mixed, and it is unclear whether these difficulties are a reliable marker of CD/HCU compared to CD/LCU. In a large sample (N = 1263, 9–18 years), we combined univariate analyses and machine learning classifiers to investigate whether CD/HCU is associated with disproportionate difficulties with fear and sadness recognition over other emotions, and whether such difficulties are a reliable individual-level marker of CD/HCU. We observed similar emotion recognition abilities in CD/HCU and CD/LCU. The CD/HCU group underperformed relative to typically developing (TD) youths, but difficulties were not specific to fear or sadness. Classifiers did not distinguish between youths with CD/HCU versus CD/LCU (52% accuracy), although youths with CD/HCU and CD/LCU were reliably distinguished from TD youths (64% and 60%, respectively). In the subset of classifiers that performed well for youths with CD/HCU, fear and sadness were the most relevant emotions for distinguishing them from youths with CD/LCU and TD youths, respectively. We conclude that non-specific emotion recognition difficulties are common in CD/HCU, but are not reliable individual-level markers of CD/HCU versus CD/LCU. These findings highlight that a reduced ability to recognise facial expressions of distress should not be assumed to be a core feature of CD/HCU.

Original language	English
Pages (from-to)	589-600
Journal	European Child and Adolescent Psychiatry
Volume	32
Early online date	18 Oct 2021
DOIs	https://doi.org/10.1007/s00787-021-01893-5
Publication status	Published - 30 Apr 2023

Bibliographical note

Funding Information:
This study was conducted by the FemNAT-CD consortium (Neurobiology and Treatment of Adolescent Female Conduct Disorder: The Central Role of Emotion Processing, coordinator Christine M. Freitag). This collaborative project is funded by the European Commission under the 7th Framework Health Program, Grant Agreement no. 602407. Ruth Pauli was supported by the Biotechnology and Biological Sciences Research Council’s Midlands Integrative Biosciences Training Partnership (BBSRC MIBTP). During the writing of the manuscript, Stephane A. De Brito was supported by a short-term Invitational Fellowship from the Japanese Society for the Promotion of Science (JSPS - S19103) and an International Academic Fellowship from the Leverhulme Trust (IAF-2019-032).

Keywords

Callous-unemotional traits
Conduct disorder
Conduct problems
Emotion recognition
Machine learning

ASJC Scopus subject areas

Pediatrics, Perinatology, and Child Health
Developmental and Educational Psychology
Psychiatry and Mental health

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1007/s00787-021-01893-5Licence: CC BY

Cite this

Pauli, R., Kohls, G., Tino, P., Rogers, J. C., Baumann, S., Ackermann, K., Bernhard, A., Martinelli, A., Jansen, L., Oldenhof, H., Gonzalez-Madruga, K., Smaragdi, A., Gonzalez-Torres, M. A., Kerexeta-Lizeaga, I., Boonmann, C., Kersten, L., Bigorra, A., Hervas, A., Stadler, C., ... De Brito, S. A. (2023). Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities. European Child and Adolescent Psychiatry, 32, 589-600. https://doi.org/10.1007/s00787-021-01893-5

Pauli, R, Kohls, G, Tino, P, Rogers, JC, Baumann, S, Ackermann, K, Bernhard, A, Martinelli, A, Jansen, L, Oldenhof, H, Gonzalez-Madruga, K, Smaragdi, A, Gonzalez-Torres, MA, Kerexeta-Lizeaga, I, Boonmann, C, Kersten, L, Bigorra, A, Hervas, A, Stadler, C, Fernandez-Rivas, A, Popma, A, Konrad, K, Herpertz-Dahlmann, B, Fairchild, G, Freitag, CM, Rotshtein, P & De Brito, SA 2023, 'Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities', European Child and Adolescent Psychiatry, vol. 32, pp. 589-600. https://doi.org/10.1007/s00787-021-01893-5

@article{6e7019d65e104733becd330054d8a36c,

title = "Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities",

abstract = "Conduct disorder (CD) with high levels of callous-unemotional traits (CD/HCU) has been theoretically linked to specific difficulties with fear and sadness recognition, in contrast to CD with low levels of callous-unemotional traits (CD/LCU). However, experimental evidence for this distinction is mixed, and it is unclear whether these difficulties are a reliable marker of CD/HCU compared to CD/LCU. In a large sample (N = 1263, 9–18 years), we combined univariate analyses and machine learning classifiers to investigate whether CD/HCU is associated with disproportionate difficulties with fear and sadness recognition over other emotions, and whether such difficulties are a reliable individual-level marker of CD/HCU. We observed similar emotion recognition abilities in CD/HCU and CD/LCU. The CD/HCU group underperformed relative to typically developing (TD) youths, but difficulties were not specific to fear or sadness. Classifiers did not distinguish between youths with CD/HCU versus CD/LCU (52% accuracy), although youths with CD/HCU and CD/LCU were reliably distinguished from TD youths (64% and 60%, respectively). In the subset of classifiers that performed well for youths with CD/HCU, fear and sadness were the most relevant emotions for distinguishing them from youths with CD/LCU and TD youths, respectively. We conclude that non-specific emotion recognition difficulties are common in CD/HCU, but are not reliable individual-level markers of CD/HCU versus CD/LCU. These findings highlight that a reduced ability to recognise facial expressions of distress should not be assumed to be a core feature of CD/HCU.",

keywords = "Callous-unemotional traits, Conduct disorder, Conduct problems, Emotion recognition, Machine learning",

author = "Ruth Pauli and Gregor Kohls and Peter Tino and Rogers, {Jack C.} and Sarah Baumann and Katharina Ackermann and Anka Bernhard and Anne Martinelli and Lucres Jansen and Helena Oldenhof and Karen Gonzalez-Madruga and Areti Smaragdi and Gonzalez-Torres, {Miguel Angel} and I{\~n}aki Kerexeta-Lizeaga and Cyril Boonmann and Linda Kersten and Aitana Bigorra and Amaia Hervas and Christina Stadler and Aranzazu Fernandez-Rivas and Arne Popma and Kerstin Konrad and Beate Herpertz-Dahlmann and Graeme Fairchild and Freitag, {Christine M.} and Pia Rotshtein and {De Brito}, {Stephane A.}",

note = "Funding Information: This study was conducted by the FemNAT-CD consortium (Neurobiology and Treatment of Adolescent Female Conduct Disorder: The Central Role of Emotion Processing, coordinator Christine M. Freitag). This collaborative project is funded by the European Commission under the 7th Framework Health Program, Grant Agreement no. 602407. Ruth Pauli was supported by the Biotechnology and Biological Sciences Research Council{\textquoteright}s Midlands Integrative Biosciences Training Partnership (BBSRC MIBTP). During the writing of the manuscript, Stephane A. De Brito was supported by a short-term Invitational Fellowship from the Japanese Society for the Promotion of Science (JSPS - S19103) and an International Academic Fellowship from the Leverhulme Trust (IAF-2019-032). ",

year = "2023",

month = apr,

day = "30",

doi = "10.1007/s00787-021-01893-5",

language = "English",

volume = "32",

pages = "589--600",

journal = "European Child and Adolescent Psychiatry",

issn = "1018-8827",

publisher = "Springer",

}

TY - JOUR

T1 - Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities

AU - Pauli, Ruth

AU - Kohls, Gregor

AU - Tino, Peter

AU - Rogers, Jack C.

AU - Baumann, Sarah

AU - Ackermann, Katharina

AU - Bernhard, Anka

AU - Martinelli, Anne

AU - Jansen, Lucres

AU - Oldenhof, Helena

AU - Gonzalez-Madruga, Karen

AU - Smaragdi, Areti

AU - Gonzalez-Torres, Miguel Angel

AU - Kerexeta-Lizeaga, Iñaki

AU - Boonmann, Cyril

AU - Kersten, Linda

AU - Bigorra, Aitana

AU - Hervas, Amaia

AU - Stadler, Christina

AU - Fernandez-Rivas, Aranzazu

AU - Popma, Arne

AU - Konrad, Kerstin

AU - Herpertz-Dahlmann, Beate

AU - Fairchild, Graeme

AU - Freitag, Christine M.

AU - Rotshtein, Pia

AU - De Brito, Stephane A.

N1 - Funding Information: This study was conducted by the FemNAT-CD consortium (Neurobiology and Treatment of Adolescent Female Conduct Disorder: The Central Role of Emotion Processing, coordinator Christine M. Freitag). This collaborative project is funded by the European Commission under the 7th Framework Health Program, Grant Agreement no. 602407. Ruth Pauli was supported by the Biotechnology and Biological Sciences Research Council’s Midlands Integrative Biosciences Training Partnership (BBSRC MIBTP). During the writing of the manuscript, Stephane A. De Brito was supported by a short-term Invitational Fellowship from the Japanese Society for the Promotion of Science (JSPS - S19103) and an International Academic Fellowship from the Leverhulme Trust (IAF-2019-032).

PY - 2023/4/30

Y1 - 2023/4/30

N2 - Conduct disorder (CD) with high levels of callous-unemotional traits (CD/HCU) has been theoretically linked to specific difficulties with fear and sadness recognition, in contrast to CD with low levels of callous-unemotional traits (CD/LCU). However, experimental evidence for this distinction is mixed, and it is unclear whether these difficulties are a reliable marker of CD/HCU compared to CD/LCU. In a large sample (N = 1263, 9–18 years), we combined univariate analyses and machine learning classifiers to investigate whether CD/HCU is associated with disproportionate difficulties with fear and sadness recognition over other emotions, and whether such difficulties are a reliable individual-level marker of CD/HCU. We observed similar emotion recognition abilities in CD/HCU and CD/LCU. The CD/HCU group underperformed relative to typically developing (TD) youths, but difficulties were not specific to fear or sadness. Classifiers did not distinguish between youths with CD/HCU versus CD/LCU (52% accuracy), although youths with CD/HCU and CD/LCU were reliably distinguished from TD youths (64% and 60%, respectively). In the subset of classifiers that performed well for youths with CD/HCU, fear and sadness were the most relevant emotions for distinguishing them from youths with CD/LCU and TD youths, respectively. We conclude that non-specific emotion recognition difficulties are common in CD/HCU, but are not reliable individual-level markers of CD/HCU versus CD/LCU. These findings highlight that a reduced ability to recognise facial expressions of distress should not be assumed to be a core feature of CD/HCU.

AB - Conduct disorder (CD) with high levels of callous-unemotional traits (CD/HCU) has been theoretically linked to specific difficulties with fear and sadness recognition, in contrast to CD with low levels of callous-unemotional traits (CD/LCU). However, experimental evidence for this distinction is mixed, and it is unclear whether these difficulties are a reliable marker of CD/HCU compared to CD/LCU. In a large sample (N = 1263, 9–18 years), we combined univariate analyses and machine learning classifiers to investigate whether CD/HCU is associated with disproportionate difficulties with fear and sadness recognition over other emotions, and whether such difficulties are a reliable individual-level marker of CD/HCU. We observed similar emotion recognition abilities in CD/HCU and CD/LCU. The CD/HCU group underperformed relative to typically developing (TD) youths, but difficulties were not specific to fear or sadness. Classifiers did not distinguish between youths with CD/HCU versus CD/LCU (52% accuracy), although youths with CD/HCU and CD/LCU were reliably distinguished from TD youths (64% and 60%, respectively). In the subset of classifiers that performed well for youths with CD/HCU, fear and sadness were the most relevant emotions for distinguishing them from youths with CD/LCU and TD youths, respectively. We conclude that non-specific emotion recognition difficulties are common in CD/HCU, but are not reliable individual-level markers of CD/HCU versus CD/LCU. These findings highlight that a reduced ability to recognise facial expressions of distress should not be assumed to be a core feature of CD/HCU.

KW - Callous-unemotional traits

KW - Conduct disorder

KW - Conduct problems

KW - Emotion recognition

KW - Machine learning

UR - http://www.scopus.com/inward/record.url?scp=85117237156&partnerID=8YFLogxK

U2 - 10.1007/s00787-021-01893-5

DO - 10.1007/s00787-021-01893-5

M3 - Article

AN - SCOPUS:85117237156

SN - 1018-8827

VL - 32

SP - 589

EP - 600

JO - European Child and Adolescent Psychiatry

JF - European Child and Adolescent Psychiatry

ER -

Machine learning classification of conduct disorder with high versus low levels of callous-unemotional traits based on facial emotion recognition abilities

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this