An exploration of adolescent facial shape changes with age via multilevel partial least squares regression.
3D facial shape
Multilevel partial least squares
adolescence
Journal
Computer methods and programs in biomedicine
ISSN: 1872-7565
Titre abrégé: Comput Methods Programs Biomed
Pays: Ireland
ID NLM: 8506513
Informations de publication
Date de publication:
Mar 2021
Mar 2021
Historique:
received:
09
10
2020
accepted:
05
01
2021
pubmed:
24
1
2021
medline:
15
5
2021
entrez:
23
1
2021
Statut:
ppublish
Résumé
Multilevel statistical models represent the existence of hierarchies or clustering within populations of subjects (or shapes in this work). This is a distinct advantage over single-level methods that do not. Multilevel partial-least squares regression (mPLSR) is used here to study facial shape changes with age during adolescence in Welsh and Finnish samples comprising males and females. 3D facial images were obtained for Welsh and Finnish male and female subjects at multiple ages from 12 to 17 years old. 1000 3D points were defined regularly for each shape by using "meshmonk" software. A three-level model was used here, including level 1 (sex/ethnicity); level 2, all "subject" variations excluding sex, ethnicity, and age; and level 3, age. The mathematical formalism of mPLSR is given in an Appendix. Differences in facial shape between the ages of 12 and 17 predicted by mPLSR agree well with previous results of multilevel principal components analysis (mPCA); buccal fat is reduced with increasing age and features such as the nose, brow, and chin become larger and more distinct. Differences due to ethnicity and sex are also observed. Plausible simulated faces are predicted from the model for different ages, sexes and ethnicities. Our models provide good representations of the shape data by consideration of appropriate measures of model fit (RMSE and R Repeat measures in our dataset for the same subject at different ages can only be modelled indirectly at the lowest level of the model at discrete ages via mPCA. By contrast, mPLSR models age explicitly as a continuous covariate, which is a strong advantage of mPLSR over mPCA. These investigations demonstrate that multivariate multilevel methods such as mPLSR can be used to describe such age-related changes for dense 3D point data. mPLSR might be of much use in future for the prediction of facial shapes for missing persons at specific ages or for simulating shapes for syndromes that affect facial shape in new subject populations.
Sections du résumé
BACKGROUND AND OBJECTIVES
OBJECTIVE
Multilevel statistical models represent the existence of hierarchies or clustering within populations of subjects (or shapes in this work). This is a distinct advantage over single-level methods that do not. Multilevel partial-least squares regression (mPLSR) is used here to study facial shape changes with age during adolescence in Welsh and Finnish samples comprising males and females.
METHODS
METHODS
3D facial images were obtained for Welsh and Finnish male and female subjects at multiple ages from 12 to 17 years old. 1000 3D points were defined regularly for each shape by using "meshmonk" software. A three-level model was used here, including level 1 (sex/ethnicity); level 2, all "subject" variations excluding sex, ethnicity, and age; and level 3, age. The mathematical formalism of mPLSR is given in an Appendix.
RESULTS
RESULTS
Differences in facial shape between the ages of 12 and 17 predicted by mPLSR agree well with previous results of multilevel principal components analysis (mPCA); buccal fat is reduced with increasing age and features such as the nose, brow, and chin become larger and more distinct. Differences due to ethnicity and sex are also observed. Plausible simulated faces are predicted from the model for different ages, sexes and ethnicities. Our models provide good representations of the shape data by consideration of appropriate measures of model fit (RMSE and R
CONCLUSIONS
CONCLUSIONS
Repeat measures in our dataset for the same subject at different ages can only be modelled indirectly at the lowest level of the model at discrete ages via mPCA. By contrast, mPLSR models age explicitly as a continuous covariate, which is a strong advantage of mPLSR over mPCA. These investigations demonstrate that multivariate multilevel methods such as mPLSR can be used to describe such age-related changes for dense 3D point data. mPLSR might be of much use in future for the prediction of facial shapes for missing persons at specific ages or for simulating shapes for syndromes that affect facial shape in new subject populations.
Identifiants
pubmed: 33485077
pii: S0169-2607(21)00009-2
doi: 10.1016/j.cmpb.2021.105935
pmc: PMC7920996
mid: NIHMS1661256
pii:
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
105935Subventions
Organisme : NIDCR NIH HHS
ID : R01 DE027023
Pays : United States
Informations de copyright
Copyright © 2021 Elsevier B.V. All rights reserved.
Déclaration de conflit d'intérêts
Declaration of Competing Interest There are no conflicts of interest relating to this research.
Références
Br J Oral Surg. 1975 Jul;13(1):1-26
pubmed: 807232
Comput Biol Med. 1992 May;22(3):181-8
pubmed: 1617952
J Forensic Sci. 2012 Mar;57(2):318-27
pubmed: 22073932
Nature. 2015 Mar 19;519(7543):309-314
pubmed: 25788095
Orthod Craniofac Res. 2006 Feb;9(1):31-7
pubmed: 16420272
Angle Orthod. 1975 Jan;45(1):12-25
pubmed: 1054930
PLoS One. 2009;4(5):e5472
pubmed: 19424496
Toxicol Rep. 2015 Jun 18;2:957-960
pubmed: 28962435
Forensic Sci Int Genet. 2014 Nov;13:208-16
pubmed: 25194685
Am J Orthod Dentofacial Orthop. 2007 Oct;132(4):439-49
pubmed: 17920496
Forensic Sci Int. 2015 Mar;248:33-40
pubmed: 25576677
Pesqui Odontol Bras. 2003 Jan-Mar;17(1):24-8
pubmed: 12908055
Eur J Orthod. 1999 Apr;21(2):137-43
pubmed: 10327737
Proc Biol Sci. 2014 Oct 7;281(1792):
pubmed: 25122232
Eur J Hum Genet. 2014 Nov;22(11):1268-71
pubmed: 24398794
Am J Med Genet. 1993 Oct 1;47(5):748-52
pubmed: 8267006
PLoS Genet. 2014 Mar 20;10(3):e1004224
pubmed: 24651127
Hum Mol Genet. 2013 Sep 15;22(18):3807-17
pubmed: 23704328
Front Genet. 2018 Oct 16;9:462
pubmed: 30386375
Sci Rep. 2019 Apr 15;9(1):6085
pubmed: 30988365
Orthod Craniofac Res. 2012 May;15(2):92-102
pubmed: 22515185
Eur J Orthod. 2014 Apr;36(2):133-9
pubmed: 22492465
Am J Orthod. 1979 Mar;75(3):264-81
pubmed: 285609
Angle Orthod. 2000 Dec;70(6):471-83
pubmed: 11138651
Nat Genet. 2018 Mar;50(3):414-423
pubmed: 29459680
J Oral Maxillofac Surg. 2012 Oct;70(10):e553-62
pubmed: 22990101
Lancet. 1973 Nov 3;302(7836):999-1001
pubmed: 4127281
Eur J Orthod. 2020 Nov 3;42(5):517-524
pubmed: 31748803
Comput Methods Programs Biomed. 2016 Jun;129:149-59
pubmed: 26830379
BMJ Open. 2013 May 28;3(5):
pubmed: 23793675
Am J Orthod Dentofacial Orthop. 1998 Sep;114(3):283-91
pubmed: 9743133
Am J Orthod. 1983 Jan;83(1):5-18
pubmed: 6571773
Eur J Orthod. 2015 Feb;37(1):1-12
pubmed: 24152377
JAMA Pediatr. 2017 Aug 1;171(8):771-780
pubmed: 28586842
BMJ Open. 2015 Sep 08;5(9):e009027
pubmed: 26351193
Angle Orthod. 1972 Oct;42(4):319-38
pubmed: 4507149
Am J Orthod Dentofacial Orthop. 2005 Jul;128(1):45-56
pubmed: 16027625
Comput Methods Programs Biomed. 2020 May;188:105272
pubmed: 31865094
PLoS One. 2016 Sep 01;11(9):e0162250
pubmed: 27584156
BMJ Open. 2017 Aug 3;7(7):e015410
pubmed: 28775124
Sci Rep. 2018 Mar 19;8(1):4771
pubmed: 29556038
Arch Dis Child. 1955 Aug;30(152):391-5
pubmed: 13249628
Eur J Orthod. 2003 Dec;25(6):599-608
pubmed: 14700266
Eur J Orthod. 2001 Jun;23(3):275-85
pubmed: 11471270
Drug Alcohol Depend. 2019 Apr 1;197:42-47
pubmed: 30772781
JAMA. 1991 Apr 17;265(15):1961-7
pubmed: 2008025