ABSTRACT. Cerebrospinal fluid (CSF) spaces include ventricles and cerebral and spinal subarachnoid spaces. CSF motion is a combined effect of CSF production rate and superimposed cardiac pulsations. Knowledge of CSF dynamics has benefited considerably from the development of phase-contrast (PC) MRI. There are several disorders such as communicating and non-communicating hydrocephalus, Chiari malformation, syringomyelic cyst and arachnoid cyst that can change the CSF dynamics. The aims of this pictorial review are to outline the PC MRI technique, CSF physiology and cerebrospinal space anatomy, to describe a group of congenital and acquired disorders that can alter the CSF dynamics, and to assess the use of PC MRI in the assessment of various central nervous system abnormalities. During the last two decades, flow-sensitive MRI techniques have been increasingly applied to quantitatively and qualitatively assess cerebrospinal fluid (CSF) flow dynamics [1]. CSF flow MRI can be used to discriminate between communicating hydrocephalus and non-communicating hydrocephalus, to localise the level of obstruction in obstructive hydrocephalus, to determine whether arachnoid cysts communicate with the subarachnoid space, to differentiate between arachnoid cysts and subarachnoid space, to discriminate between syringomyelia and cystic myelomalacia, and to evaluate flow patterns of posterior fossa cystic malformations. This imaging method can also provide significant information in pre-operative evaluation of Chiari 1 malformation and normal pressure hydrocephalus and post-operative follow-up of patients with neuroendoscopic third ventriculostomy (NTV) and ventriculoperitoneal (VP) shunt [1][2][3][4][5][6][7][8][9][10][11]. In this pictorial review, we emphasise phase-contrast (PC) MRI technique, CSF physiology and cerebrospinal space anatomy, congenital and acquired disorders that can alter the cerebrospinal fluid dynamics, and the use of PC MRI in the assessment of various central nervous system (CNS) abnormalities. CSF anatomy and physiologyCSF comprises all intracerebral ventricles, spinal and brain subarachnoid spaces, such as cisterns and sulci, and the central canal of the spinal cord. The rate of CSF formation in humans is about 0.3-0.4 ml min 21 (about 500 ml day 21 ). Total CSF volume is 90-150 ml in adults and 10-60 ml in neonates. Potential sites of CSF origin include the choroid plexus, parenchyma of the brain and the spinal cord, and ependymal lining of the ventricles [12].The portion of the fluid formed in the lateral ventricles escapes by the foramen of Monro into the third ventricle and then via the aqueduct into the fourth ventricle. A little CSF occurs in the central canal of the spinal cord and may be added to the intraventricular supply. From the fourth ventricle the fluid pours into the subarachnoid spaces through the medial foramen of Magendie and the two lateral foramina of Luschka. There is no functional communication between the cerebral ventricles and the subarachnoid spaces in any region except from th...

No abstract

No abstract

We propose an adversarial training procedure for learning a causal implicit generative model for a given causal graph. We show that adversarial training can be used to learn a generative model with true observational and interventional distributions if the generator architecture is consistent with the given causal graph. We consider the application of generating faces based on given binary labels where the dependency structure between the labels is preserved with a causal graph. This problem can be seen as learning a causal implicit generative model for the image and labels. We devise a two-stage procedure for this problem. First we train a causal implicit generative model over binary labels using a neural network consistent with a causal graph as the generator. We empirically show that Wasserstein GAN can be used to output discrete labels. Later we propose two new conditional GAN architectures, which we call CausalGAN and CausalBEGAN. We show that the optimal generator of the CausalGAN, given the labels, samples from the image distributions conditioned on these labels. The conditional GAN combined with a trained causal implicit generative model for the labels is then an implicit causal generative network over the labels and the generated image. We show that the proposed architectures can be used to sample from observational and interventional image distributions, even for interventions which do not naturally occur in the dataset.

No abstract

No abstract

Most FLLs are benign ones such as hemangiomas and cysts, which can be readily and practically characterized only by using visual assessment of DWIs without requiring time-consuming conventional and dynamic contrast-enhanced imaging sequences. Some benign lesions that are falsely interpreted as malignant can be further characterized by using conventional and contrast-enhanced MR studies.

Extrahepatic abdominal hydatid lesions have nearly identical imaging features, including the presence of cyst wall calcification, daughter cysts, and membrane detachment. The combinations of radiologic and serologic tests especially in patients living in the endemic areas contribute to the diagnosis. Despite their rarity, being familiar with the spectrum of radiologic findings in these unusual sites is helpful to improve diagnostic accuracy.