Abstract
	The intracranial volume is made up by the cerebrospinal fluid (CSF) volume, the blood flow volume and the brain parenchyma. The inflow blood to the skull in systole temporary increases the intracranial volume. According to the Monroe-Kellie doctrine an increase in one volume should cause a decrease in one or both of the remaining two volumes in order to maintain volume constant. The imbalances that occur in this process of cerebral homeostasis have been linked to neurodegenerative and cerebrovascular diseases. Therefore, adequate methodologies in order to analyze the dynamic of the intracranial fluids (LCR and blood) are necessary.
	The cine sequences of phase contrast magnetic resonance imaging (PC-MRI) with cardiac synchronism allow to quantify the CSF and blood flows during a cardiac cycle. Flow measurement with PC-MRI is accurate and reproducible if an adequate acquisition protocol is used. The reproducibility and accuracy of the measures also depend on the use of adequate post-processing techniques that allow to segment the regions of interest (ROI), with great consistency and independency of the operator and to correct the residual systematic errors  caused by imperfect suppression of eddy currents and the contributing to the signal of small movements that the brain presents due to the transmission of the vascular pulse as well as the aliasing reflected as an abrupt change and opposed to the original sense of flow. These techniques also must take into account the errors associated with the partial volume effect (PVE), caused by the presence of stationary tissue and flow inside the voxels of the periphery of the region under study.
	The overall objective of this thesis was to develop a reproducible methodology to quantitatively evaluate the intracranial fluids dynamics within CSF spaces (aqueduct of Sylvius, prepontine cistern and subarachnoid space C2C3) and main cerebral blood vessels (carotids and vertebral arteries, jugular veins, straight and sagittal superior sinus) by means of PC-MRI. This methodology includes semiautomatic segmentation techniques by thresholding and K-means clustering that allow to delineate the ROI with great independency of the observer and to apply background and aliasing correction.
	Amplitude parameters (increase of flow volume in systole, mean flows, pulsatility, compliance indexes and supratentorial CSF) and temporal parameters (delay respect to the inflow blood to the brain) were quantified. The distribution of the pulsatibility of the CSF flow between the ventricular and subarachnoid spaces was measured as the ratio of aqueductal to C2C3 cervical stroke volume.
	The commitment between the acquisition protocol and the used methodology to calculate each one of the parameters has helped to establish reliable and reproducible normality ranges for each one of them. The evaluated control subjects were asymptomatic and they had no history of intracranial hypertension, neurological disease, vascular events or relevant risks factors allowing to generalize the measures. They were studied at the same hourly stripe to avoid influences of the circadian cycle. All acquisitions were carried with a 3.0T MR unit to avoid the influence of the field strength on parameters.
	In conclusion an adequate methodology was developed to carry out quick and reproducible quantitative blood and CSF flow analysis useful in clinical practice. This new approach to study cerebral fluid interactions should help to improve the understanding about the physiopathology of several cerebral diseases caused by CSF, blood and intracranial pressure alterations.

	Keywords- Phase contrast MRI, CSF, cerebral flow, segmentation, image processing, k-means