Your browser doesn't support javascript.
loading
Mostrar: 20 | 50 | 100
Resultados 1 - 20 de 38
Filtrar
1.
IEEE Trans Med Imaging ; 40(5): 1363-1376, 2021 May.
Artigo em Inglês | MEDLINE | ID: mdl-33507867

RESUMO

To better understand early brain development in health and disorder, it is critical to accurately segment infant brain magnetic resonance (MR) images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). Deep learning-based methods have achieved state-of-the-art performance; h owever, one of the major limitations is that the learning-based methods may suffer from the multi-site issue, that is, the models trained on a dataset from one site may not be applicable to the datasets acquired from other sites with different imaging protocols/scanners. To promote methodological development in the community, the iSeg-2019 challenge (http://iseg2019.web.unc.edu) provides a set of 6-month infant subjects from multiple sites with different protocols/scanners for the participating methods. T raining/validation subjects are from UNC (MAP) and testing subjects are from UNC/UMN (BCP), Stanford University, and Emory University. By the time of writing, there are 30 automatic segmentation methods participated in the iSeg-2019. In this article, 8 top-ranked methods were reviewed by detailing their pipelines/implementations, presenting experimental results, and evaluating performance across different sites in terms of whole brain, regions of interest, and gyral landmark curves. We further pointed out their limitations and possible directions for addressing the multi-site issue. We find that multi-site consistency is still an open issue. We hope that the multi-site dataset in the iSeg-2019 and this review article will attract more researchers to address the challenging and critical multi-site issue in practice.

2.
Ann Nucl Med ; 34(8): 545-548, 2020 Aug.
Artigo em Inglês | MEDLINE | ID: mdl-32451886

RESUMO

OBJECTIVE: We aimed to determine at what size melanoma metastases become detectable by PET/CT. METHODS: We reviewed a total of 293 whole-body PET/CT studies performed on 212 patients for staging of melanoma where there was an MRI within a month of the PET/CT. MR and PET/CT were reviewed independently by separate readers. RESULTS: PET/CT revealed an overall incidental true-positive rate of 1% on a per-patient basis, consistent with other studies, with 'hot' lesions (more avid than brain parenchyma) visible at smaller sizes than 'cold' lesions. CONCLUSIONS: PET/CT can detect metastatic melanoma lesions over about 2 cm in size, with hot lesions generally visible at smaller sizes.

3.
Semin Ultrasound CT MR ; 41(3): 275-283, 2020 Jun.
Artigo em Inglês | MEDLINE | ID: mdl-32448484

RESUMO

Radiologists add value in the setting of clinically suspected multiple sclerosis (MS) predominantly through knowledge of typical and atypical imaging findings, thus aiding the neurologist in differentiating multiple disease processes with similar clinical presentations which may mimic MS. This paper is meant as a guide to deepen understanding for the radiologist, enabling them to be a helpful member of the care team for MS patients.

4.
Semin Ultrasound CT MR ; 41(3): 284-295, 2020 Jun.
Artigo em Inglês | MEDLINE | ID: mdl-32448485

RESUMO

This article discusses mimics of multiple sclerosis (MS). Excluded in this discussion are neuromyelitis optica and vasculitis, discussed in other articles in this journal. Covered entities include posterior reversible encephalopathy syndrome, reversible vasoconstriction syndrome, acute disseminated encephalomyelitis, Sussac's Syndrome, and chronic idiopathic demyelinating polyneuropathy. There are also multiple infectious entities that mimic MS including; progressive multi-focal leukoencephalopathy (PML), Toxoplasmosis, Tuberculosis, Herpes Simplex Virus, Cytomegalovirus, Varicella zoster virus, Epstein Barr virus, Cryptococcus and Human immunodeficiency virus. In addition, there are leukoencephalopathies that can present in adulthood including Adrenoleukodystrophy, Metachromatic leukodystrophy, Cerebral autosomal dominant idiopathic leukoencephalopathy, Leigh's and Alexanders disease that could be mistaken for MS.

5.
Semin Ultrasound CT MR ; 41(3): 296-308, 2020 Jun.
Artigo em Inglês | MEDLINE | ID: mdl-32448486

RESUMO

This article discusses central nervous system vasculitis, a clinical and MRI mimic of multiple sclerosis (MS). There is a paucity of discussion of vasculitis in the radiology literature, and many MS neurologists believe that vasculitis is underdiagnosed. Therefore, the authors hope that the readers will find this paper increases their knowledge about CNS vasculitis and improves their ability to differentiate MS from vasculitis.

6.
Semin Ultrasound CT MR ; 41(3): 309-318, 2020 Jun.
Artigo em Inglês | MEDLINE | ID: mdl-32448487

RESUMO

The purpose of this paper is to serve as a template for greater understanding for the practicing radiologist about key steps to perform multimodality computer analysis of MRI images, specifically in multiple sclerosis patients. With this understanding, radiologists will be better equipped about how best to process and analyze MRI imaging data and obtain accurate quantitative information for MS patient evaluation. A secondary intent of this article is to improve radiologist understanding of how artificial intelligence will be employed in the future for better patient stratification, and for evaluation of response to therapy in both clinical care and drug trials.

7.
Clin Case Rep ; 7(10): 1957-1961, 2019 Oct.
Artigo em Inglês | MEDLINE | ID: mdl-31624617

RESUMO

Central nervous system (CNS)-relapsed mantle cell lymphoma (MCL) is a rare, aggressive non-Hodgkin lymphoma without a standard treatment. Ibrutinib has shown promising results for inducing remission in other non-Hodgkin lymphomas and may be considered as successful treatment for CNS-relapsed MCL in the future as well.

8.
Seizure ; 61: 128-134, 2018 Oct.
Artigo em Inglês | MEDLINE | ID: mdl-30138825

RESUMO

PURPOSE: Epilepsy surgery is the most successful method of treating medically unresponsive epilepsy, but carries a risk of morbidity. PET/MR is an emerging technique that increases detection of focal lesions whose resection may result in symptom remission. METHODS: Retrospective review of 74 focal epilepsy patients over a period of 3 years who had a PET/MR was performed following IRB permission and informed consent. 27 patients underwent surgery or RNS (responsive neurostimulator) placement. RESULTS: Hybrid PET-MR identified new anatomic or functional lesions in 10 patients not identified with standalone 3 T MR. Of the 27 patients who underwent focal surgery (19) or RNS placement (8), 24 showed improvement (Engel's I-III), 2 did not (Engel's IV), and one had an RNS explanted due to infection. MR and PET were read by 2 separate neuroradiologists and nuclear medicine physicians, respectively. Modalities were evaluated in terms of ability to detect the correct lobe and side for a focal lesion whose resection improved symptoms. Prior standalone MR exhibited 71-77% sensitivity and 0% specificity (as there were only 2 nonresponders), MR associated with PET/MR had 68-71% sensitivity and 0-50% specificity (depending on whether a lesion was seen on one of the nonresponders), and PET had 68-71% sensitivity and 25-33% specificity. Using either PET or MR to identify a focal lesion, PET/MR had sensitivity of 78-82% and specificity 0-50%. CONCLUSIONS: PET-MR provides additional sensitivity when used as two combined modalities for detecting possible epileptic foci.


Assuntos
Encéfalo/diagnóstico por imagem , Epilepsia Resistente a Medicamentos/diagnóstico por imagem , Imagem por Ressonância Magnética/métodos , Tomografia por Emissão de Pósitrons/métodos , Adulto , Epilepsia Resistente a Medicamentos/mortalidade , Epilepsia Resistente a Medicamentos/fisiopatologia , Epilepsia Resistente a Medicamentos/terapia , Eletroencefalografia , Feminino , Humanos , Processamento de Imagem Assistida por Computador , Masculino , Estudos Retrospectivos , Resultado do Tratamento
9.
Clin Immunol ; 197: 45-53, 2018 12.
Artigo em Inglês | MEDLINE | ID: mdl-30149119

RESUMO

IL-11 induced differentiation and expansion of Th17 cells in patients with early relapsing-remitting multiple sclerosis (RRMS). In mice with relapsing-remitting experimental autoimmune encephalomyelitis (RREAE), IL-11 exacerbated disease, induced demyelination in the central nervous system (CNS), increased the percentage of IL-17A+CD4+ Th17 cells in the CNS in the early acute phase, and up-regulated serum IL-17A levels and the percentage of IL-17A+CD4+ Th17 cells in lymph nodes, and IFN-γ+CD4+ T cells in spinal cord in the RR phase. IL-11 antagonist suppressed RREAE disease activities, inhibited IL-17A+CD4+ cell infiltration and demyelination in the CNS, and decreased the percentage of IL-17A+CD4+ T cells in peripheral blood mononuclear cells and ICAM1+CD4+ T cells in brain and SC. Diffusion Tensor Imaging indicated that IL-11 antagonist inhibited demyelination in several brain regions. We conclude that by suppressing Th17 cell-mediated neuroinflammation and demyelination, IL-11 antagonist can be further studied as a potential selective and early therapy for RRMS.


Assuntos
Encéfalo/diagnóstico por imagem , Encefalomielite Autoimune Experimental/imunologia , Interleucina-11/antagonistas & inibidores , Medula Espinal/diagnóstico por imagem , Células Th17/imunologia , Animais , Encéfalo/imunologia , Imagem de Tensor de Difusão , Inflamação , Interleucina-11/imunologia , Subunidade alfa de Receptor de Interleucina-11 , Leucócitos Mononucleares , Camundongos , Esclerose Múltipla Recidivante-Remitente , Proteínas Recombinantes de Fusão , Medula Espinal/imunologia
10.
Med Image Anal ; 43: 10-22, 2018 Jan.
Artigo em Inglês | MEDLINE | ID: mdl-28961451

RESUMO

Hippocampal subfields play important roles in many brain activities. However, due to the small structural size, low signal contrast, and insufficient image resolution of 3T MR, automatic hippocampal subfields segmentation is less explored. In this paper, we propose an automatic learning-based hippocampal subfields segmentation method using 3T multi-modality MR images, including structural MRI (T1, T2) and resting state fMRI (rs-fMRI). The appearance features and relationship features are both extracted to capture the appearance patterns in structural MR images and also the connectivity patterns in rs-fMRI, respectively. In the training stage, these extracted features are adopted to train a structured random forest classifier, which is further iteratively refined in an auto-context model by adopting the context features and the updated relationship features. In the testing stage, the extracted features are fed into the trained classifiers to predict the segmentation for each hippocampal subfield, and the predicted segmentation is iteratively refined by the trained auto-context model. To our best knowledge, this is the first work that addresses the challenging automatic hippocampal subfields segmentation using relationship features from rs-fMRI, which is designed to capture the connectivity patterns of different hippocampal subfields. The proposed method is validated on two datasets and the segmentation results are quantitatively compared with manual labels using the leave-one-out strategy, which shows the effectiveness of our method. From experiments, we find a) multi-modality features can significantly increase subfields segmentation performance compared to those only using one modality; b) automatic segmentation results using 3T multi-modality MR images could be partially comparable to those using 7T T1 MRI.


Assuntos
Hipocampo/anatomia & histologia , Imagem por Ressonância Magnética/métodos , Humanos
11.
Environ Res ; 150: 411-422, 2016 10.
Artigo em Inglês | MEDLINE | ID: mdl-27376929

RESUMO

Children's air pollution exposures are associated with systemic and brain inflammation and the early hallmarks of Alzheimer's disease (AD). The Apolipoprotein E (APOE) 4 allele is the most prevalent genetic risk for AD, with higher risk for women. We assessed whether gender, BMI, APOE and metabolic variables in healthy children with high exposures to ozone and fine particulate matter (PM2.5) influence cognition. The Wechsler Intelligence Scale for Children (WISC-R) was administered to 105 Mexico City children (12.32±5.4 years, 69 APOE 3/3 and 36 APOE 3/4). APOE 4v 3 children showed decrements on attention and short-term memory subscales, and below-average scores in Verbal, Performance and Full Scale IQ. APOE 4 females had higher BMI and females with normal BMI between 75-94% percentiles had the highest deficits in Total IQ, Performance IQ, Digit Span, Picture Arrangement, Block Design and Object Assembly. Fasting glucose was significantly higher in APOE 4 children p=0.006, while Gender was the main variable accounting for the difference in insulin, HOMA-IR and leptin (p<.05). Gender, BMI and APOE influence children's cognitive responses to air pollution and glucose is likely a key player. APOE 4 heterozygous females with >75% to <94% BMI percentiles are at the highest risk of severe cognitive deficits (1.5-2SD from average IQ). Young female results highlight the urgent need for gender-targeted health programmes to improve cognitive responses. Multidisciplinary intervention strategies could provide paths for prevention or amelioration of female air pollution targeted cognitive deficits and possible long-term AD progression.


Assuntos
Poluentes Atmosféricos/análise , Apolipoproteína E4/genética , Cognição , Ozônio/análise , Material Particulado/análise , Adolescente , Glicemia/análise , Índice de Massa Corporal , Criança , Cidades , Exposição Ambiental/efeitos adversos , Exposição Ambiental/análise , Feminino , Humanos , Masculino , México , Escalas de Wechsler
13.
Mach Learn Med Imaging ; 10019: 229-236, 2016 Oct.
Artigo em Inglês | MEDLINE | ID: mdl-28603791

RESUMO

Hippocampal subfields play important and divergent roles in both memory formation and early diagnosis of many neurological diseases, but automatic subfield segmentation is less explored due to its small size and poor image contrast. In this paper, we propose an automatic learning-based hippocampal subfields segmentation framework using multi-modality 3TMR images, including T1 MRI and resting-state fMRI (rs-fMRI). To do this, we first acquire both 3T and 7T T1 MRIs for each training subject, and then the 7T T1 MRI are linearly registered onto the 3T T1 MRI. Six hippocampal subfields are manually labeled on the aligned 7T T1 MRI, which has the 7T image contrast but sits in the 3T T1 space. Next, corresponding appearance and relationship features from both 3T T1 MRI and rs-fMRI are extracted to train a structured random forest as a multi-label classifier to conduct the segmentation. Finally, the subfield segmentation is further refined iteratively by additional context features and updated relationship features. To our knowledge, this is the first work that addresses the challenging automatic hippocampal subfields segmentation using 3T routine T1 MRI and rs-fMRI. The quantitative comparison between our results and manual ground truth demonstrates the effectiveness of our method. Besides, we also find that (a) multi-modality features significantly improved subfield segmentation performance due to the complementary information among modalities; (b) automatic segmentation results using 3T multimodality images are partially comparable to those on 7T T1 MRI.

14.
Seizure ; 31: 1-4, 2015 Sep.
Artigo em Inglês | MEDLINE | ID: mdl-26362368

RESUMO

PURPOSE: We aim to evaluate the utility/improved accuracy of hybrid PET/MR compared to current practice separate 3T MRI and PET-CT imaging for localization of seizure foci. METHOD: In a pilot study, twenty-nine patients undergoing epilepsy surgery evaluation were imaged using PET/MR. This subject group had 29 previous clinical 3T MRI as well as 12 PET-CT studies. Prior clinical PET and MR images were read sequentially while the hybrid PET/MR was concurrently read. RESULTS: The median interval between hybrid PET/MR and prior imaging studies was 5 months (range 1-77 months). In 24 patients, there was no change in the read between the clinical exams and hybrid PET/MR while new anatomical or functional lesions were identified by hybrid PET/MR in 5 patients without significant clinical change. Four new anatomical MR lesions were seen with concordant PET findings. The remaining patient revealed a new abnormal PET lesion without an MR abnormality. All new PET/MR lesions were clinically significant with concordant EEG and/or SPECT results as potential epileptic foci. CONCLUSION: Our initial hybrid PET-MRI experience increased diagnostic yields for detection of potential epileptic lesions. This may be due to the unique advantage of improved co-registration and simultaneous review of both structural and functional data.


Assuntos
Epilepsia Resistente a Medicamentos/diagnóstico por imagem , Epilepsia Resistente a Medicamentos/patologia , Imagem por Ressonância Magnética/métodos , Imagem Multimodal/métodos , Tomografia por Emissão de Pósitrons/métodos , Adulto , Encéfalo/diagnóstico por imagem , Encéfalo/patologia , Encéfalo/fisiopatologia , Encéfalo/cirurgia , Diagnóstico , Epilepsia Resistente a Medicamentos/fisiopatologia , Epilepsia Resistente a Medicamentos/cirurgia , Eletroencefalografia , Humanos , Projetos Piloto , Cuidados Pré-Operatórios/métodos , Estudos Retrospectivos , Convulsões/diagnóstico por imagem , Convulsões/patologia , Convulsões/fisiopatologia , Convulsões/cirurgia , Tomografia Computadorizada de Emissão de Fóton Único
15.
J Alzheimers Dis ; 48(4): 1065-75, 2015.
Artigo em Inglês | MEDLINE | ID: mdl-26402110

RESUMO

Severe air pollution exposures produce systemic, respiratory, myocardial, and brain inflammation and Alzheimer's disease (AD) hallmarks in clinically healthy children. We tested whether hippocampal metabolite ratios are associated with contrasting levels of air pollution, APOE, and body mass index (BMI) in paired healthy children and one parent sharing the same APOE alleles. We used 1H-MRS to interrogate bilateral hippocampal single-voxel in 57 children (12.45 ± 3.4 years) and their 48 parents (37.5 ± 6.78 years) from a low pollution city versus Mexico City (MC). NAA/Cr, Cho/Cr, and mI/Cr metabolite ratios were analyzed. The right hippocampus NAA/Cr ratio was significantly different between cohorts (p = 0.007). The NAA/Cr ratio in right hippocampus in controls versus APOE ɛ4 MC children and in left hippocampus in MC APOE ɛ4 parents versus their children was significantly different after adjusting for age, gender, and BMI (p = 0.027 and 0.01, respectively). The NAA/Cr ratio is considered reflective of neuronal density/functional integrity/loss of synapses/higher pTau burden, thus a significant decrease in hippocampal NAA/Cr ratios may constitute a spectral marker of early neurodegeneration in young urbanites. Decreases in NAA/Cr correlate well with cognitive function, behavioral symptoms, and dementia severity; thus, since the progression of AD starts decades before clinical diagnosis, our findings support the hypothesis that under chronic exposures to fine particulate matter and ozone above the standards, neurodegenerative processes start in childhood and APOE ɛ4 carriers are at higher risk. Gene and environmental factors are critical in the development of AD and the identification and neuroprotection of young urbanites at high risk must become a public health priority.


Assuntos
Poluição do Ar/efeitos adversos , Doença de Alzheimer/metabolismo , Apolipoproteína E4/genética , Ácido Aspártico/análogos & derivados , Creatina/metabolismo , Hipocampo/metabolismo , Adolescente , Adulto , Doença de Alzheimer/epidemiologia , Doença de Alzheimer/genética , Ácido Aspártico/metabolismo , Índice de Massa Corporal , Criança , Estudos de Coortes , Família , Feminino , Humanos , Masculino , México/epidemiologia , Estudos Prospectivos , Espectroscopia de Prótons por Ressonância Magnética , População Urbana , Proteínas tau/metabolismo
16.
Epilepsy Behav Case Rep ; 4: 52-5, 2015.
Artigo em Inglês | MEDLINE | ID: mdl-26288757

RESUMO

Mesial temporal sclerosis (MTS) is a well-recognized cause of intractable epilepsy; however, coexistence with focal cortical dysplasia (FCD) is less common. Middle fossa epidermoid cysts are rare and may involve the temporal lobe. Most epidermoids are clinically silent, slow-growing, and seldom associated with overt symptomatology, including seizures. We describe a patient with multiple comorbidities including left MTS and a large epidermoid cyst involving the left quadrigeminal plate cistern compressing upon the cerebellar vermis and tail of the left hippocampus, resulting in refractory left temporal lobe epilepsy. The patient underwent left anterior temporal lobectomy. The surgical pathology demonstrated a third pathological finding of left temporal FCD type Ia. The patient has been seizure-free since the surgery. This case provides additional information with regard to the understanding of epileptogenicity and surgical planning in patients with MTS and epidermoid cysts.

17.
Ann Otol Rhinol Laryngol ; 124(9): 681-90, 2015 Sep.
Artigo em Inglês | MEDLINE | ID: mdl-25766964

RESUMO

OBJECTIVES: To determine if spatial orientation of the cochlea within the temporal bone is related to age or sensorineural hearing loss (SNHL) and describe the implications for cochlear implantation. METHODS: Five angles of cochlear orientation were determined from computed tomography (CT) imaging of the temporal bones in adults with (n = 55) and without (n = 27) sensorineural hearing loss (SNHL) and children with (n = 45) and without (n = 12) SNHL: facial recess versus basal turn, posterior semicircular canal versus basal turn, round window versus basal turn (axial view), round window versus basal turn (coronal view), and the cochlear axis versus the mastoid facial nerve. RESULTS: All angles showed substantial variation between subjects and between ears. The angles between the round window and basal turn (coronal view) and the posterior semicircular canal and basal turn were significantly correlated with age for all subjects with SNHL (r = 0.22, P = .002 and r = 0.15, P = .03, respectively). Patients with SNHL had significantly more acute angles (46.6° vs 55.8°) between the round window versus basal turn (axial orientation) compared to controls (P < .001). CONCLUSIONS: Cochlear orientation within the temporal bone changes with age and the degree of SNHL. These results suggest that the approach to the round window for electrode insertion might differ between children and adults.


Assuntos
Implante Coclear/métodos , Nervo Facial/anatomia & histologia , Perda Auditiva Neurossensorial , Janela da Cóclea , Canais Semicirculares , Adulto , Fatores Etários , Idoso , Anatomia Comparada , Criança , Pré-Escolar , Feminino , Perda Auditiva Neurossensorial/patologia , Perda Auditiva Neurossensorial/cirurgia , Humanos , Masculino , Pessoa de Meia-Idade , Janela da Cóclea/anatomia & histologia , Janela da Cóclea/patologia , Canais Semicirculares/anatomia & histologia , Canais Semicirculares/patologia , Osso Temporal/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos
18.
J Neuroimmunol ; 278: 108-11, 2015 Jan 15.
Artigo em Inglês | MEDLINE | ID: mdl-25595259

RESUMO

IMPORTANCE: This is the first report of neuromyelitis optica spectrum disorder (NMOSD) associated with dermatomyositis (DM). REPORT: A 40year-old Caucasian female presented with 6months of worsening fatigue, rash, acute weakness worse in her lower extremities, and urinary retention. She was found to have both NMOSD and anti-melanoma differentiation-associated gene (MDA)5 positive DM with interstitial lung disease (ILD). She was treated aggressively and she regained her ability to ambulate. CONCLUSION: We recommend considering NMOSD in the differential diagnosis of patients with DM and other autoimmune disorders that also present with clinical signs of myelopathy.


Assuntos
Dermatomiosite/diagnóstico , Neuromielite Óptica/fisiopatologia , Adulto , Anticorpos/sangue , RNA Helicases DEAD-box/imunologia , Dermatomiosite/sangue , Feminino , Humanos , Helicase IFIH1 Induzida por Interferon , Imagem por Ressonância Magnética , Pele/patologia , Medula Espinal/patologia
19.
Artigo em Inglês | MEDLINE | ID: mdl-27019875

RESUMO

Aberrant development of the human brain during the first year after birth is known to cause critical implications in later stages of life. In particular, neuropsychiatric disorders, such as attention deficit hyperactivity disorder (ADHD), have been linked with abnormal early development of the hippocampus. Despite its known importance, studying the hippocampus in infant subjects is very challenging due to the significantly smaller brain size, dynamically varying image contrast, and large across-subject variation. In this paper, we present a novel method for effective hippocampus segmentation by using a multi-atlas approach that integrates the complementary multimodal information from longitudinal T1 and T2 MR images. In particular, considering the highly heterogeneous nature of the longitudinal data, we propose to learn their common feature representations by using hierarchical multi-set kernel canonical correlation analysis (CCA). Specifically, we will learn (1) within-time-point common features by projecting different modality features of each time point to its own modality-free common space, and (2) across-time-point common features by mapping all time-point-specific common features to a global common space for all time points. These final features are then employed in patch matching across different modalities and time points for hippocampus segmentation, via label propagation and fusion. Experimental results demonstrate the improved performance of our method over the state-of-the-art methods.

20.
Med Image Comput Comput Assist Interv ; 17(Pt 2): 308-15, 2014.
Artigo em Inglês | MEDLINE | ID: mdl-25485393

RESUMO

Accurate segmentation of the hippocampus from infant MR brain images is a critical step for investigating early brain development. Unfortunately, the previous tools developed for adult hippocampus segmentation are not suitable for infant brain images acquired from the first year of life, which often have poor tissue contrast and variable structural patterns of early hippocampal development. From our point of view, the main problem is lack of discriminative and robust feature representations for distinguishing the hippocampus from the surrounding brain structures. Thus, instead of directly using the predefined features as popularly used in the conventional methods, we propose to learn the latent feature representations of infant MR brain images by unsupervised deep learning. Since deep learning paradigms can learn low-level features and then successfully build up more comprehensive high-level features in a layer-by-layer manner, such hierarchical feature representations can be more competitive for distinguishing the hippocampus from entire brain images. To this end, we apply Stacked Auto Encoder (SAE) to learn the deep feature representations from both T1- and T2-weighed MR images combining their complementary information, which is important for characterizing different development stages of infant brains after birth. Then, we present a sparse patch matching method for transferring hippocampus labels from multiple atlases to the new infant brain image, by using deep-learned feature representations to measure the interpatch similarity. Experimental results on 2-week-old to 9-month-old infant brain images show the effectiveness of the proposed method, especially compared to the state-of-the-art counterpart methods.


Assuntos
Envelhecimento/patologia , Envelhecimento/fisiologia , Inteligência Artificial , Hipocampo/anatomia & histologia , Hipocampo/crescimento & desenvolvimento , Imagem por Ressonância Magnética/métodos , Reconhecimento Automatizado de Padrão/métodos , Humanos , Interpretação de Imagem Assistida por Computador/métodos , Lactente , Recém-Nascido , Reprodutibilidade dos Testes , Sensibilidade e Especificidade
SELEÇÃO DE REFERÊNCIAS
DETALHE DA PESQUISA
...