Hypothalamus fully automatic segmentation from MR images using a U-Net based architecture

Rodrigues, Lívia; Rezende, Tiago Marques de; Zanesco, Ariane Maria; Hernández, Ana María; França, Marcondes Cavalcante; Rittner, Letícia

doi:10.1117/12.2542585

Cited by 5 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Indeed, this experiment demonstrates that the presented method: ( i ) segments the posterior and tubular subregions with the same level of precision as the whole hypothalamus, and ( ii ) significantly outperforms MAS (thoroughly validated and widely used in neuroimaging) for the whole hypothalamus and all subunits, while running orders of magnitude faster at test time. In comparison with a recent deep learning approach for whole hypothalamus segmentation ( Rodrigues et al., 2020 ), our model shows an improvement of 0.07 in Dice coefficient. Although these results are not directly comparable due to differences in datasets, the improvement may be because of our more aggressive data augmentation scheme, including: linear and elastic transformations, bias field corruption, and intensity augmentation.…”

Section: Discussionmentioning

confidence: 74%

“…Automated algorithms have been introduced to tackle these problems, as they do not require human intervention and enable reproducible segmentations of large datasets. However, very few automated strategies have been proposed to segment the whole hypothalamus in structural MRI scans ( D’Haese, Duay, Merchant, Macq, Dawant, 2003 , Orbes-Arteaga, Cárdenas-Peña, Álvarez, Orozco, Castellanos-Dominguez, 2015 , Rodrigues, Rezende, Zanesco, Hernandez, Franca, Rittner, 2020 , Thomas, Beyer, Lewe, Zhang, Schindler, Schönknecht, Stumvoll, Villringer, Witte, 2019 ), and no automated method exists – to the best of our knowledge – for hypothalamic nuclei segmentation.…”

Section: Introductionmentioning

confidence: 99%

“…Modern automated image segmentation techniques rely on deep neural networks ( Kamnitsas et al., 2017 ), which have been recently applied to segment the whole hypothalamus (as opposed to hypothalamic subunits) with relatively high accuracy ( Rodrigues et al., 2020 ). By learning a set of convolutional kernels, convolutional neural networks (CNN) can effectively capture highly non-linear distributions between inputs and outputs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automated segmentation of the hypothalamus and associated subunits in brain MRI

et al. 2020

View full text Add to dashboard Cite

Highlights A publicly available deep learning tool to segment the hypothalamus and its subunits. Our tool outperforms inter-rater accuracy and approaches intra-rater precision level. It can robustly generalise to unseen heterogeneous datasets. It yields a rejection rate of less than 1% in a QC analysis performed on 675 scans. It detects subtle subunit-specific hypothalamic atrophy in Alzheimer’s Disease.

show abstract

Section: Discussionmentioning

confidence: 74%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Automated segmentation of the hypothalamus and associated subunits in brain MRI

et al. 2020

View full text Add to dashboard Cite

show abstract

Advances in MRI-Based Anatomy of the Human Hypothalamus and Effects of the Hypothalamic Neuropeptide Oxytocin on Brain BOLD Signals

Mueller

Spindler

Caspers

et al. 2021

Masterclass in Neuroendocrinology

View full text Add to dashboard Cite

AI-assisted quantification of hypothalamic atrophy in amyotrophic lateral sclerosis by convolutional neural network-based automatic segmentation

Vernikouskaya,

Müller,

Roselli

et al. 2023

Sci Rep

View full text Add to dashboard Cite

The hypothalamus is a small structure of the brain with an essential role in metabolic homeostasis, sleep regulation, and body temperature control. Some neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS) and dementia syndromes are reported to be related to hypothalamic volume alterations. Despite its crucial role in human body regulation, neuroimaging studies of this structure are rather scarce due to work-intensive operator-dependent manual delineations from MRI and lack of automated segmentation tools. In this study we present a fully automatic approach based on deep convolutional neural networks (CNN) for hypothalamic segmentation and volume quantification. We applied CNN of U-Net architecture with EfficientNetB0 backbone to allow for accurate automatic hypothalamic segmentation in seconds on a GPU. We further applied our approach for the quantification of the normalized hypothalamic volumes to a large neuroimaging dataset of 432 ALS patients and 112 healthy controls (without the ground truth labels). Using the automated volumetric analysis, we could reproduce hypothalamic atrophy findings associated with ALS by detecting significant volume differences between ALS patients and controls at the group level. In conclusion, a fast and unbiased AI-assisted hypothalamic quantification method is introduced in this study (whose acceptance rate based on the outlier removal strategy was estimated to be above 95%) and made publicly available for researchers interested in the conduction of hypothalamus studies at a large scale.

show abstract

Hypothalamus fully automatic segmentation from MR images using a U-Net based architecture

Cited by 5 publications

References 0 publications

Automated segmentation of the hypothalamus and associated subunits in brain MRI

Automated segmentation of the hypothalamus and associated subunits in brain MRI

Advances in MRI-Based Anatomy of the Human Hypothalamus and Effects of the Hypothalamic Neuropeptide Oxytocin on Brain BOLD Signals

AI-assisted quantification of hypothalamic atrophy in amyotrophic lateral sclerosis by convolutional neural network-based automatic segmentation

Contact Info

Product

Resources

About