Machine Learning of Infant Spontaneous Movements for the Early Prediction of Cerebral Palsy: A Multi-Site Cohort Study.
cerebral palsy
general movement assessment
machine learning
premature infants
Journal
Journal of clinical medicine
ISSN: 2077-0383
Titre abrégé: J Clin Med
Pays: Switzerland
ID NLM: 101606588
Informations de publication
Date de publication:
18 Dec 2019
18 Dec 2019
Historique:
received:
22
10
2019
revised:
29
11
2019
accepted:
16
12
2019
entrez:
22
12
2019
pubmed:
22
12
2019
medline:
22
12
2019
Statut:
epublish
Résumé
Early identification of cerebral palsy (CP) during infancy will provide opportunities for early therapies and treatments. The aim of the present study was to present a novel machine-learning model, the Computer-based Infant Movement Assessment (CIMA) model, for clinically feasible early CP prediction based on infant video recordings. The CIMA model was designed to assess the proportion (%) of CP risk-related movements using a time-frequency decomposition of the movement trajectories of the infant's body parts. The CIMA model was developed and tested on video recordings from a cohort of 377 high-risk infants at 9-15 weeks corrected age to predict CP status and motor function (ambulatory vs. non-ambulatory) at mean 3.7 years age. The performance of the model was compared with results of the general movement assessment (GMA) and neonatal imaging. The CIMA model had sensitivity (92.7%) and specificity (81.6%), which was comparable to observational GMA or neonatal cerebral imaging for the prediction of CP. Infants later found to have non-ambulatory CP had significantly more CP risk-related movements (median: 92.8%, The CIMA model may be a clinically feasible alternative to observational GMA.
Sections du résumé
BACKGROUND
BACKGROUND
Early identification of cerebral palsy (CP) during infancy will provide opportunities for early therapies and treatments. The aim of the present study was to present a novel machine-learning model, the Computer-based Infant Movement Assessment (CIMA) model, for clinically feasible early CP prediction based on infant video recordings.
METHODS
METHODS
The CIMA model was designed to assess the proportion (%) of CP risk-related movements using a time-frequency decomposition of the movement trajectories of the infant's body parts. The CIMA model was developed and tested on video recordings from a cohort of 377 high-risk infants at 9-15 weeks corrected age to predict CP status and motor function (ambulatory vs. non-ambulatory) at mean 3.7 years age. The performance of the model was compared with results of the general movement assessment (GMA) and neonatal imaging.
RESULTS
RESULTS
The CIMA model had sensitivity (92.7%) and specificity (81.6%), which was comparable to observational GMA or neonatal cerebral imaging for the prediction of CP. Infants later found to have non-ambulatory CP had significantly more CP risk-related movements (median: 92.8%,
CONCLUSION
CONCLUSIONS
The CIMA model may be a clinically feasible alternative to observational GMA.
Identifiants
pubmed: 31861380
pii: jcm9010005
doi: 10.3390/jcm9010005
pmc: PMC7019773
pii:
doi:
Types de publication
Journal Article
Langues
eng
Subventions
Organisme : The Liaison Committee between the Central Norway Regional Health Authority and the Norwegian University of Science and Technology, Trondheim, Norway
ID : SO: 90056100
Déclaration de conflit d'intérêts
The authors declare no conflict of interest. Nor the funders had any role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish results. Colleen Peyton is a member of the Prechtl General Movement Trust speaker’s bureau.
Références
IEEE Trans Pattern Anal Mach Intell. 2019 Jul 17;:
pubmed: 31331883
Dev Med Child Neurol. 2015 Jan;57(1):29-36
pubmed: 25041565
Dev Med Child Neurol. 2018 May;60(5):438
pubmed: 29517109
Acta Paediatr. 2019 Oct;108(10):1817-1824
pubmed: 30883894
Arch Dis Child. 2000 Dec;83(6):475-80
pubmed: 11087279
J Clin Med. 2019 Oct 04;8(10):
pubmed: 31590221
IEEE Trans Neural Syst Rehabil Eng. 2016 Nov;24(11):1225-1234
pubmed: 27046852
Phys Ther. 2000 Oct;80(10):974-85
pubmed: 11002433
J Pediatr (Rio J). 2016 May-Jun;92(3 Suppl 1):S64-70
pubmed: 26997356
Dev Med Child Neurol. 2018 Jan;60(1):39-46
pubmed: 28832987
Dev Med Child Neurol. 1997 Apr;39(4):214-23
pubmed: 9183258
Physiol Meas. 2019 Feb 26;40(2):02TR02
pubmed: 30669130
Front Neurol. 2015 Jan 09;5:284
pubmed: 25620954
Conf Proc IEEE Eng Med Biol Soc. 2014;2014:3779-83
pubmed: 25570814
Pediatr Res. 2017 Oct;82(4):665-670
pubmed: 28745715
Dev Med Child Neurol. 2000 Dec;42(12):816-24
pubmed: 11132255
Dev Med Child Neurol Suppl. 2007 Feb;109:8-14
pubmed: 17370477
Early Hum Dev. 2014 Aug;90(8):387-92
pubmed: 24951073
J Pediatr. 2018 Dec;203:156-162
pubmed: 30244983
Early Hum Dev. 2009 May;85(5):297-302
pubmed: 19138831
Dev Med Child Neurol. 2013 Jun;55(6):509-19
pubmed: 23346889
PLoS One. 2014 May 12;9(5):e96944
pubmed: 24820185
J Clin Med. 2019 Oct 25;8(11):
pubmed: 31717717
IEEE Trans Neural Syst Rehabil Eng. 2012 Jul;20(4):605-14
pubmed: 22531824
Dev Med Child Neurol. 2010 Aug;52(8):773-8
pubmed: 20187882
Neurosci Biobehav Rev. 2007;31(8):1213-9
pubmed: 17561255
JAMA Pediatr. 2017 Sep 1;171(9):897-907
pubmed: 28715518