Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs

Sara Ranjbar, Kyle W. Singleton, Lee Curtin, Susan Christine Massey, Andrea Hawkins-Daarud, Pamela R. Jackson, Kristin R. Swanson

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Fluid intelligence (Gf) has been defined as the ability to reason and solve previously unseen problems. Links to Gf have been found in magnetic resonance imaging (MRI) sequences such as functional MRI and diffusion tensor imaging. As part of the Adolescent Brain Cognitive Development Neurocognitive Prediction Challenge 2019, we sought to predict Gf in children aged 9–10 from T1-weighted (T1 W) MRIs. The data included atlas–aligned volumetric T1 W images, atlas–defined segmented regions, age, and sex for 3739 subjects used for training and internal validation and 415 subjects used for external validation. We trained sex-specific convolutional neural net (CNN) and random forest models to predict Gf. For the convolutional model, skull-stripped volumetric T1 W images aligned to the SRI24 brain atlas were used for training. Volumes of segmented atlas regions along with each subject’s age were used to train the random forest regressor models. Performance was measured using the mean squared error (MSE) of the predictions. Random forest models achieved lower MSEs than CNNs. Further, the external validation data had a better MSE for females than males (60.68 vs. 80.74), with a combined MSE of 70.83. Our results suggest that predictive models of Gf from volumetric T1 W MRI features alone may perform better when trained separately on male and female data. However, the performance of our models indicates that more information is necessary beyond the available data to make accurate predictions of Gf.

Original languageEnglish (US)
Title of host publicationAdolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings
EditorsKilian M. Pohl, Ehsan Adeli, Wesley K. Thompson, Marius George Linguraru
PublisherSpringer
Pages150-157
Number of pages8
ISBN (Print)9783030319007
DOIs
StatePublished - Jan 1 2019
Event1st Challenge in Adolescent Brain Cognitive Development Neurocognitive Prediction, ABCD-NP 2019, held in conjunction with the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2019 - Shenzhen, China
Duration: Oct 13 2019Oct 13 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11791 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference1st Challenge in Adolescent Brain Cognitive Development Neurocognitive Prediction, ABCD-NP 2019, held in conjunction with the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2019
CountryChina
CityShenzhen
Period10/13/1910/13/19

Fingerprint

Magnetic resonance imaging
Brain
Random Forest
Fluid
Fluids
Mean Squared Error
Atlas
Magnetic Resonance Imaging
Magnetic resonance
Prediction
Predict
Neural Nets
Model
Diffusion tensor imaging
Functional Magnetic Resonance Imaging
Imaging techniques
Predictive Model
Tensor
Imaging
Intelligence

Keywords

  • Deep learning
  • Fluid intelligence
  • Sex differences

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Ranjbar, S., Singleton, K. W., Curtin, L., Massey, S. C., Hawkins-Daarud, A., Jackson, P. R., & Swanson, K. R. (2019). Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs. In K. M. Pohl, E. Adeli, W. K. Thompson, & M. G. Linguraru (Eds.), Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings (pp. 150-157). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11791 LNCS). Springer. https://doi.org/10.1007/978-3-030-31901-4_18

Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs. / Ranjbar, Sara; Singleton, Kyle W.; Curtin, Lee; Massey, Susan Christine; Hawkins-Daarud, Andrea; Jackson, Pamela R.; Swanson, Kristin R.

Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings. ed. / Kilian M. Pohl; Ehsan Adeli; Wesley K. Thompson; Marius George Linguraru. Springer, 2019. p. 150-157 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11791 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ranjbar, S, Singleton, KW, Curtin, L, Massey, SC, Hawkins-Daarud, A, Jackson, PR & Swanson, KR 2019, Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs. in KM Pohl, E Adeli, WK Thompson & MG Linguraru (eds), Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11791 LNCS, Springer, pp. 150-157, 1st Challenge in Adolescent Brain Cognitive Development Neurocognitive Prediction, ABCD-NP 2019, held in conjunction with the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2019, Shenzhen, China, 10/13/19. https://doi.org/10.1007/978-3-030-31901-4_18
Ranjbar S, Singleton KW, Curtin L, Massey SC, Hawkins-Daarud A, Jackson PR et al. Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs. In Pohl KM, Adeli E, Thompson WK, Linguraru MG, editors, Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings. Springer. 2019. p. 150-157. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-31901-4_18
Ranjbar, Sara ; Singleton, Kyle W. ; Curtin, Lee ; Massey, Susan Christine ; Hawkins-Daarud, Andrea ; Jackson, Pamela R. ; Swanson, Kristin R. / Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs. Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings. editor / Kilian M. Pohl ; Ehsan Adeli ; Wesley K. Thompson ; Marius George Linguraru. Springer, 2019. pp. 150-157 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{3318485c2c474c06a910cdccfca58268,
title = "Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs",
abstract = "Fluid intelligence (Gf) has been defined as the ability to reason and solve previously unseen problems. Links to Gf have been found in magnetic resonance imaging (MRI) sequences such as functional MRI and diffusion tensor imaging. As part of the Adolescent Brain Cognitive Development Neurocognitive Prediction Challenge 2019, we sought to predict Gf in children aged 9–10 from T1-weighted (T1 W) MRIs. The data included atlas–aligned volumetric T1 W images, atlas–defined segmented regions, age, and sex for 3739 subjects used for training and internal validation and 415 subjects used for external validation. We trained sex-specific convolutional neural net (CNN) and random forest models to predict Gf. For the convolutional model, skull-stripped volumetric T1 W images aligned to the SRI24 brain atlas were used for training. Volumes of segmented atlas regions along with each subject’s age were used to train the random forest regressor models. Performance was measured using the mean squared error (MSE) of the predictions. Random forest models achieved lower MSEs than CNNs. Further, the external validation data had a better MSE for females than males (60.68 vs. 80.74), with a combined MSE of 70.83. Our results suggest that predictive models of Gf from volumetric T1 W MRI features alone may perform better when trained separately on male and female data. However, the performance of our models indicates that more information is necessary beyond the available data to make accurate predictions of Gf.",
keywords = "Deep learning, Fluid intelligence, Sex differences",
author = "Sara Ranjbar and Singleton, {Kyle W.} and Lee Curtin and Massey, {Susan Christine} and Andrea Hawkins-Daarud and Jackson, {Pamela R.} and Swanson, {Kristin R.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-31901-4_18",
language = "English (US)",
isbn = "9783030319007",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "150--157",
editor = "Pohl, {Kilian M.} and Ehsan Adeli and Thompson, {Wesley K.} and Linguraru, {Marius George}",
booktitle = "Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings",

}

TY - GEN

T1 - Sex Differences in Predicting Fluid Intelligence of Adolescent Brain from T1-Weighted MRIs

AU - Ranjbar, Sara

AU - Singleton, Kyle W.

AU - Curtin, Lee

AU - Massey, Susan Christine

AU - Hawkins-Daarud, Andrea

AU - Jackson, Pamela R.

AU - Swanson, Kristin R.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Fluid intelligence (Gf) has been defined as the ability to reason and solve previously unseen problems. Links to Gf have been found in magnetic resonance imaging (MRI) sequences such as functional MRI and diffusion tensor imaging. As part of the Adolescent Brain Cognitive Development Neurocognitive Prediction Challenge 2019, we sought to predict Gf in children aged 9–10 from T1-weighted (T1 W) MRIs. The data included atlas–aligned volumetric T1 W images, atlas–defined segmented regions, age, and sex for 3739 subjects used for training and internal validation and 415 subjects used for external validation. We trained sex-specific convolutional neural net (CNN) and random forest models to predict Gf. For the convolutional model, skull-stripped volumetric T1 W images aligned to the SRI24 brain atlas were used for training. Volumes of segmented atlas regions along with each subject’s age were used to train the random forest regressor models. Performance was measured using the mean squared error (MSE) of the predictions. Random forest models achieved lower MSEs than CNNs. Further, the external validation data had a better MSE for females than males (60.68 vs. 80.74), with a combined MSE of 70.83. Our results suggest that predictive models of Gf from volumetric T1 W MRI features alone may perform better when trained separately on male and female data. However, the performance of our models indicates that more information is necessary beyond the available data to make accurate predictions of Gf.

AB - Fluid intelligence (Gf) has been defined as the ability to reason and solve previously unseen problems. Links to Gf have been found in magnetic resonance imaging (MRI) sequences such as functional MRI and diffusion tensor imaging. As part of the Adolescent Brain Cognitive Development Neurocognitive Prediction Challenge 2019, we sought to predict Gf in children aged 9–10 from T1-weighted (T1 W) MRIs. The data included atlas–aligned volumetric T1 W images, atlas–defined segmented regions, age, and sex for 3739 subjects used for training and internal validation and 415 subjects used for external validation. We trained sex-specific convolutional neural net (CNN) and random forest models to predict Gf. For the convolutional model, skull-stripped volumetric T1 W images aligned to the SRI24 brain atlas were used for training. Volumes of segmented atlas regions along with each subject’s age were used to train the random forest regressor models. Performance was measured using the mean squared error (MSE) of the predictions. Random forest models achieved lower MSEs than CNNs. Further, the external validation data had a better MSE for females than males (60.68 vs. 80.74), with a combined MSE of 70.83. Our results suggest that predictive models of Gf from volumetric T1 W MRI features alone may perform better when trained separately on male and female data. However, the performance of our models indicates that more information is necessary beyond the available data to make accurate predictions of Gf.

KW - Deep learning

KW - Fluid intelligence

KW - Sex differences

UR - http://www.scopus.com/inward/record.url?scp=85075658459&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85075658459&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-31901-4_18

DO - 10.1007/978-3-030-31901-4_18

M3 - Conference contribution

AN - SCOPUS:85075658459

SN - 9783030319007

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 150

EP - 157

BT - Adolescent Brain Cognitive Development Neurocognitive Prediction - 1st Challenge, ABCD-NP 2019, held in Conjunction with MICCAI 2019, Proceedings

A2 - Pohl, Kilian M.

A2 - Adeli, Ehsan

A2 - Thompson, Wesley K.

A2 - Linguraru, Marius George

PB - Springer

ER -