haut de page

mise à jour du
21 novembre 2002
1992; 35; 1; 10-5
 Cerebrospinal fluid flow:
physiology of respiration-related pulsations
Schroth G and U Klose
Department of neuroradiology University Tübingen Germany


Cardiac-related pulsations of the cerebrospinal fluid (CSF) can be investigated using ECG-gated FLASH magnetic resonance imaging (MRI). However, because 256 exitationts are required to obtain an image with a 256 x 256 matrix other components of the CSF flow are lost. By measurements on animals and humans it was established in the last century that CSF pressure also changes with respiration. However, the resulting changes in CSF flow have not been studied by X-ray contrast cinecisternography or cineventriculography nor by cine-MRI. We have studied respiration-related CSF dynamics by employing a new real-time MRI technique. [...]
Discussion : As early as 1886, Knollproved that pressure fluctuations in the cisterna magna of the rabbit changed not only with the heart beat, but also with respiration. Expiration was accompanied by an increase in CSF pressure, while a drop in pressure occurred with inspiration. After Quinke introduced spinal puncture into clinical practice at the end of the 19th century, Becher was able to confirm in humans the animal experiments showing the dependence of intracranial pressure fluctuations on heart beat and respiration. Since the pressure fluctuations were measured with an open water manometer, the results were relatively coarse; however, they were later confirmed by many other researchers.
Through precise measurements with membrane manometers, Ewig and Lullies showed a CSF pressure drop in the lumbar dural sac throughout inspiration, with a pressure increase during expiration, in thoracic respiration. During abdominal breathing an inspiratory increase and an expiratory decrease were observed.
Using ECG-gated FLASH-2D measurements it was possible to confirm that the cardiac-related oscillating CSF motion in the cervical spinal canal is superimposed on a bulk flow moving in separate channels, directed mainly downwards in the anterior subarachnoid space and upwards laterally. In addition, through modification of the FLASH-2D technique and with the help of RACE (real time acquisition and evaluation) it can be shown that the opposed components of CSF flow are not only anatomically separate, but alternate with time: caudad CSF flow in the anterior cervical subarachnoid space dominates during inspiration, whereas expiration is accompanied by an increase in cephalad flow. Caudad flow acceleration occurs; immediately on beginning inspiration, but ceases with abdominal compression.
In the aqueduct the cardiac-related CSF pulsation is affected in a similar way by respiration. The systolic pulsatile flow component downwards from the third to the fourth ventricle is increased during inspiration, but after a delay of 2-3 cardiac cycles (Fig. 10), whereas during the late phase of expiration backflow from the fourth to the third is accelerated. These observations can be interpreted as follows.
During thoracic inspiration, the spinal epidural veins are emptied, because of the inspiratory partial vacuum in the thoracic cavity; this leads to rapid caudad acceleration of CSF flow in the spinal canal. The increased inspiratory venous inflow to the heart does not reach the brain until passing through the lungs. This results in acceleration of the CSF pulsation in the aqueduct 2-3 cardiac cycles later. However, during predominantly abdominal breathing or a Valsalva manoeuvre outflow from the large veins caudal to the diaphragm is obstructed. This results in a volume-pressure increase in the epidural venous plexus of the thoracolumbar spinal canal, which can thus be discounted as a compliance system for craniospinal CSF pulsation.
Not all components of the "third circulation" in the spinal and intracranial subarachnoid space can be assessed, even with RACE. However, it shows that even minute changes in CSF flow following complex, physiological stimuli can be detected and explained.
In a patient examined late in the evening RACE showed reproducible changes in spinal CSF flow while yawning rapid flow in the anterior cervical subarachnoid space at the beginning of the yawn, ceasing after a few seconds. Using the saturation technique in this patient, who had asymmetrical internal jugular veins with small on the right and large on the left, blood flow in the vein was seen to increase significantly for several seconds.
Yawning begins with a deep thoracic inspiration, followed by an increase in intra-abdominal pressure after closure of the glottis, with simultaneous reflex opening of the mouth. The result of the initial deep thoracic inspiration is rapid caudad CSF flow in the spinal canal due to emptying of the epidural veins, ending after closure of the glottis and the increase in pressure in the abdomen. The inspiratory increase in blood volume blood reaches the intracranial system two or three heart beats later, after its passage through the lungs. Compliance of the craniospinal CSF system is stopped meanwhile, because of the rise in abdominal pressure. Therefore, only the expressible intracranial venous blood confers compliance, resulting in laster blood flow in the internal jugular vein. The slow course of the yawning reflex could be teleologically interpreted as ensuring the necessary delay between the emptying of the spinal veins and the arrival of the increased inspiratory blood volume in the cranial cavity two to three heart beats later. Its accompaniment by pleasurable sensations can be seen as an additional indication to the body that it must not cut short this physiologically necessary interval.
Abstract : Cardiac-related motion of the cerebrospinal fluid (CSF) was investigated by analysis of the velocity-dependent phase of CSF protons and flow-dependent signal enhancement in magnitude images using ECG-gated FLASH sequences. In the cerebral aqueduct, CSF flow from the third to the fourth ventricle begins 200 ms after the R-wave of the ECG and simulates an arterial pulse wave pattern. It lasts about 60% of the cardiac cycle and is followed by backflow from the fourth to the third ventricle, which is slower and shorter. In the spinal canal, oscillating caudad motion precedes flow from the third to the fourth ventricle by about 50-100 ms and is superimposed on a bulk flow, which moves simultaneously in opposite directions in separate subarachnoid channels; it is directed mainly caudally in the anterior cervical subarachnoid space.
Schroth, G. and U. Klose (1992). "Cerebrospinal fluid flow; Physiology of cardiac-related pulsation." Neuroradiology 35(1): 1-9. Abstract :
Cerebrospinal fluid (CSF) flow in the cerebral aqueduct and spinal canal was analysed using real-time magnetic resonance imaging measurement techniques. Respiration-induced rhythmic modulation of the cardiac-related oscillating CSF pulsation in the cerebral aqueduct and spinal canal was found. Deep inspiration was immediately followed by a marked increase in downward CSF flow in the cervical spinal canal, whereas a delay of about two heart beats was seen before downward flow from the third to the fourth ventricle increased. This pattern was also detected during yawning and was followed by a marked increase of blood flow in the internal jugular vein.
Darko Oreskovic and all The formation and circulation of cerebrospinal fluid inside the cat brain ventricles: a fact or an illusion?Neuroscience Letters 2002;327; p103-106
Formation and circulation of cerebrospinal fluid (CSF) have been studied in the isolated brain ventricles of anesthetized cats by a new approach and under direct observation. A plastic cannula was introduced into the aqueduct of Sylvius through the vermis cerebelli and the outflow of CSFfrom the cannula was used asthe CSF formation and circulation index. During the 60 min of observation at a physiological CSF pressure not a single drop of CSF escaped out of the end of the cannula. This indicatesthatCSF netformation and circulation insidethe brain ventricles, proposed by classical hypothesis regarding CSF dynamics, should be at least re-evaluated.
In conclusion, in spite of the fact that in our model animals were under anaesthesia and that the fourth ventricle was excluded from the investigation, the classical hypothesis of the CSF formation and circulation cannot be upheld, because our results clearly show that the net CSF formation does not take place inside the other brain ventricles and that CSF does not circulate as a slow river from the ventricles to the subarachnoid space. It is very well known (and commonly applied in routine clinical practice) that the increase of blood osmolality can decrease the CSF (intracranial) pressure, simply by extraction of fluid from nervous tissue. Our experimental results clearly revealed that the increase in CSF osmolality will subsequently lead to the increase of the CSF volume inside the brain ventricles. This drives to a conclusion that the osmolality represents one of the major determinants of fluid exchange in intracranial pressure. It seems that the control of the CSF volume is under the influence of hydrostatic and osmotic forces between the CSF systern and the surrounding tissue and that the CSF volume will be changed, depending on the prevalence of those forces, caused by (patho)physiological reasons inside or outside the CSF system. Anyhow, the results presented in this article call for a new approach to the physiology and pathology of CSF and we feel that time has come to re-evaluate the classical hypothesis of the CSF hydrodynamies to abandon this hypothesis as the traditional framework of thinking.
Bergsneider, M. Evolving concepts of cerebrospinal fluid physiology. Neurosurg Clin N Am (2001) 12(4): 631-8.
This article reviews the basic known functions of cerebrospinal fluid (CSF). The traditional concepts of CSF production and absorption are reviewed and recent challenges to these concepts are discussed. MR imaging studies have begun to elucidate the complex interaction between pulsatile CSF movement, bulk CSF flow, and intracranial compliance. An understanding of a variety of disorders, including hydrocephalus and Chiari malformations, continue to evolve as knowledge of CSF physiology is increased.
-Legendre R, Piéron H. De la propriété hypnotoxique des humeurs développée au cours d'une veille prolongée C.R. Société de Biologie de Paris 1912; 70; 210-212
-Bouyssou M, Tricoire J Experimental detection of a cervical arousal mechanism of yawning, enhancing intracerebral (CSF) fluid pressure J Dental Res 1985; 64; 721
 -Lepp FH Remarques sur la signification physiologique du bâillement Bull Group Int Rech Sci Stomtol Odontol 1982; 25; 251-290
-Nolman B Yawning, cerebral fluid and the lymphatic pump 2006
-Oreskovic D; Klarica M; Vukic M The formation and circulation of cerebrospinal fluid inside the cat brain ventricles : a fact or an illusion ? Neuroscience letters 2002; 327; 103-106
-Patra P, Gunness TK, Robert R Physiologic variations of the internal jugular vein surface, role of the omohyoid muscle, a preliminary echographic study Surg Radiol Anat 1988; 10; 2; 107-12
-Schniter E The evolution of yawning : why do we yawn and why is it contagious ? thèse 2001
-Schroth G, Klose U Cerebrospinal fluid flow; Physiology of respiration-related pulsations. Neuroradiology 1992; 35; 1; 10-15
-Walusinski O Prostaglandines, adénosine, sommeil & bâillements 2004
-Woodbury R, B Abretj Influence of dying gasps, yawns and sighs on blood pressure and blood flow Am J Physiol 1944; 142; 721-726