ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates
Abstract Infant brain magnetic resonance imaging (MRI) is a promising approach for studying early neurodevelopment. However, segmenting small regions such as limbic structures is challenging due to their low inter-regional contrast and high curvature. MRI studies of the adult brain have successfully...
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2022-05-01
|
| Series: | Brain Informatics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s40708-022-00161-9 |
| _version_ | 1818207055735947264 |
|---|---|
| author | Yun Wang Fateme Sadat Haghpanah Xuzhe Zhang Katie Santamaria Gabriela Koch da Costa Aguiar Alves Elizabeth Bruno Natalie Aw Alexis Maddocks Cristiane S. Duarte Catherine Monk Andrew Laine Jonathan Posner program collaborators for Environmental influences on Child Health Outcomes |
| author_facet | Yun Wang Fateme Sadat Haghpanah Xuzhe Zhang Katie Santamaria Gabriela Koch da Costa Aguiar Alves Elizabeth Bruno Natalie Aw Alexis Maddocks Cristiane S. Duarte Catherine Monk Andrew Laine Jonathan Posner program collaborators for Environmental influences on Child Health Outcomes |
| author_sort | Yun Wang |
| collection | DOAJ |
| description | Abstract Infant brain magnetic resonance imaging (MRI) is a promising approach for studying early neurodevelopment. However, segmenting small regions such as limbic structures is challenging due to their low inter-regional contrast and high curvature. MRI studies of the adult brain have successfully applied deep learning techniques to segment limbic structures, and similar deep learning models are being leveraged for infant studies. However, these deep learning-based infant MRI segmentation models have generally been derived from small datasets, and may suffer from generalization problems. Moreover, the accuracy of segmentations derived from these deep learning models relative to more standard Expectation–Maximization approaches has not been characterized. To address these challenges, we leveraged a large, public infant MRI dataset (n = 473) and the transfer-learning technique to first pre-train a deep convolutional neural network model on two limbic structures: amygdala and hippocampus. Then we used a leave-one-out cross-validation strategy to fine-tune the pre-trained model and evaluated it separately on two independent datasets with manual labels. We term this new approach the Infant Deep learning SEGmentation Framework (ID-Seg). ID-Seg performed well on both datasets with a mean dice similarity score (DSC) of 0.87, a mean intra-class correlation (ICC) of 0.93, and a mean average surface distance (ASD) of 0.31 mm. Compared to the Developmental Human Connectome pipeline (dHCP) pipeline, ID-Seg significantly improved segmentation accuracy. In a third infant MRI dataset (n = 50), we used ID-Seg and dHCP separately to estimate amygdala and hippocampus volumes and shapes. The estimates derived from ID-seg, relative to those from the dHCP, showed stronger associations with behavioral problems assessed in these infants at age 2. In sum, ID-Seg consistently performed well on two different datasets with an 0.87 DSC, however, multi-site testing and extension for brain regions beyond the amygdala and hippocampus are still needed. |
| first_indexed | 2024-12-12T04:22:50Z |
| format | Article |
| id | doaj.art-d4481a55fb15434197165a2f7b2b6941 |
| institution | Directory Open Access Journal |
| issn | 2198-4018 2198-4026 |
| language | English |
| last_indexed | 2024-12-12T04:22:50Z |
| publishDate | 2022-05-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Brain Informatics |
| spelling | doaj.art-d4481a55fb15434197165a2f7b2b69412022-12-22T00:38:17ZengSpringerOpenBrain Informatics2198-40182198-40262022-05-019111110.1186/s40708-022-00161-9ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimatesYun Wang0Fateme Sadat Haghpanah1Xuzhe Zhang2Katie Santamaria3Gabriela Koch da Costa Aguiar Alves4Elizabeth Bruno5Natalie Aw6Alexis Maddocks7Cristiane S. Duarte8Catherine Monk9Andrew Laine10Jonathan Posner11program collaborators for Environmental influences on Child Health OutcomesDepartment of Psychiatry and Behavioral Sciences, Duke UniversityDepartment of Computer Science, University Of TorontoDepartment of Biomedical Engineering, Columbia UniversityNew York State Psychiatric InstituteNew York State Psychiatric InstituteNew York State Psychiatric InstituteNew York State Psychiatric InstituteDepartment of Radiology, Columbia UniversityNew York State Psychiatric InstituteNew York State Psychiatric InstituteDepartment of Biomedical Engineering, Columbia UniversityDepartment of Psychiatry and Behavioral Sciences, Duke UniversityAbstract Infant brain magnetic resonance imaging (MRI) is a promising approach for studying early neurodevelopment. However, segmenting small regions such as limbic structures is challenging due to their low inter-regional contrast and high curvature. MRI studies of the adult brain have successfully applied deep learning techniques to segment limbic structures, and similar deep learning models are being leveraged for infant studies. However, these deep learning-based infant MRI segmentation models have generally been derived from small datasets, and may suffer from generalization problems. Moreover, the accuracy of segmentations derived from these deep learning models relative to more standard Expectation–Maximization approaches has not been characterized. To address these challenges, we leveraged a large, public infant MRI dataset (n = 473) and the transfer-learning technique to first pre-train a deep convolutional neural network model on two limbic structures: amygdala and hippocampus. Then we used a leave-one-out cross-validation strategy to fine-tune the pre-trained model and evaluated it separately on two independent datasets with manual labels. We term this new approach the Infant Deep learning SEGmentation Framework (ID-Seg). ID-Seg performed well on both datasets with a mean dice similarity score (DSC) of 0.87, a mean intra-class correlation (ICC) of 0.93, and a mean average surface distance (ASD) of 0.31 mm. Compared to the Developmental Human Connectome pipeline (dHCP) pipeline, ID-Seg significantly improved segmentation accuracy. In a third infant MRI dataset (n = 50), we used ID-Seg and dHCP separately to estimate amygdala and hippocampus volumes and shapes. The estimates derived from ID-seg, relative to those from the dHCP, showed stronger associations with behavioral problems assessed in these infants at age 2. In sum, ID-Seg consistently performed well on two different datasets with an 0.87 DSC, however, multi-site testing and extension for brain regions beyond the amygdala and hippocampus are still needed.https://doi.org/10.1186/s40708-022-00161-9Deep learningSegmentationInfant neuroimagingConvolutional neural networksHippocampusAmygdala |
| spellingShingle | Yun Wang Fateme Sadat Haghpanah Xuzhe Zhang Katie Santamaria Gabriela Koch da Costa Aguiar Alves Elizabeth Bruno Natalie Aw Alexis Maddocks Cristiane S. Duarte Catherine Monk Andrew Laine Jonathan Posner program collaborators for Environmental influences on Child Health Outcomes ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates Brain Informatics Deep learning Segmentation Infant neuroimaging Convolutional neural networks Hippocampus Amygdala |
| title | ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates |
| title_full | ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates |
| title_fullStr | ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates |
| title_full_unstemmed | ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates |
| title_short | ID-Seg: an infant deep learning-based segmentation framework to improve limbic structure estimates |
| title_sort | id seg an infant deep learning based segmentation framework to improve limbic structure estimates |
| topic | Deep learning Segmentation Infant neuroimaging Convolutional neural networks Hippocampus Amygdala |
| url | https://doi.org/10.1186/s40708-022-00161-9 |
| work_keys_str_mv | AT yunwang idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT fatemesadathaghpanah idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT xuzhezhang idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT katiesantamaria idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT gabrielakochdacostaaguiaralves idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT elizabethbruno idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT natalieaw idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT alexismaddocks idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT cristianesduarte idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT catherinemonk idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT andrewlaine idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT jonathanposner idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates AT programcollaboratorsforenvironmentalinfluencesonchildhealthoutcomes idseganinfantdeeplearningbasedsegmentationframeworktoimprovelimbicstructureestimates |