- Email: [email protected]
Viscoelastic properties of human cerebellum using magnetic resonance elastography
Journal of Biomechanics 44 (2011) 1909–1913
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Viscoelastic properties of human cerebellum using magnetic resonance elastography John Zhang a,b, Michael A. Green a, Ralph Sinkus c, Lynne E. Bilston a,d,n a
Neuroscience Research Australia, Barker Street, Randwick NSW 2031, Australia Faculty of Medicine University of New South Wales, Kensington, NSW 2052, Australia c INSERM U773, CRB3, Centre de Recherches Biome´dicales Bichat-Beaujon, Paris, France d Prince of Wales Clinical School, University of New South Wales, Kensington NSW 2052, Australia b
a r t i c l e i n f o
abstract
Article history: Accepted 19 April 2011
Background: The cerebellum has never been mechanically characterised, despite its physiological importance in the control of motion and the clinical prevalence of cerebellar pathologies. The aim of this study was to measure the linear viscoelastic properties of the cerebellum in human volunteers using Magnetic Resonance Elastography (MRE). Methods: Coronal plane brain 3D MRE data was performed on eight healthy adult volunteers, at 80 Hz, to compare the properties of cerebral and cerebellar tissues. The linear viscoelastic storage (G0 ) and loss moduli (G00 ) were estimated from the MRE wave images by solving the wave equation for propagation through an isotropic linear viscoelastic solid. Contributions of the compressional wave were removed via application of the curl-operator. Results: The storage modulus for the cerebellum was found to be significantly lower than that for the cerebrum, for both white and grey matter. Cerebrum: white matter (mean 7SD) G0 ¼ 2.41 7 0.23 kPa, grey matter G0 ¼2.34 70.22 kPa; cerebellum: white matter, G0 ¼1.85 70.18 kPa, grey matter G0 ¼ 1.77 7 0.24 kPa; cerebrum vs cerebellum, p o0.001. For the viscous behaviour, there were differences in between regions and also by tissue type, with the white matter being more viscous than grey matter and the cerebrum more viscous than the cerebellum. Cerebrum: white matter G00 ¼ 1.21 7 0.21 kPa, grey matter G00 ¼ 1.11 7 0.03 kPa; cerebellum: white matter G00 ¼ 1.1 7 0.23 kPa, grey matter G00 ¼ 0.94 70.17 kPa. Discussion: These data represent the first available data on the viscoelastic properties of cerebellum, which suggest that the cerebellum is less physically stiff than the cerebrum, possibly leading to a different response to mechanical loading. These data will be useful for modelling of the cerebellum for a range of purposes. & 2011 Elsevier Ltd. All rights reserved.
Keywords: Brain viscoelasticity Shear modulus Magnetic resonance imaging In vivo measurements
1. Introduction The viscoelastic properties of brain tissue play a role in a variety of neurological disorders and traumatic brain injury. It has been postulated that certain conditions, such as normal pressure hydrocephalus, are influenced by a change in the mechanical properties of the brain tissue (Pang and Altschuler, 1994; Dutta-Roy et al., 2008), and those changes in the viscoelasticity of the brain may be a marker for neurodegenerative conditions such as Alzheimer’s disease and multiple sclerosis (Kruse et al., 2008; Wuerfel et al., 2010). In addition, these parameters are vital for computational simulations such as Finite Element Analysis (FEA) of brain conditions, traumatic
n Corresponding author at: Neuroscience Research Australia, Barker St, Randwick, NSW 2031, Australia. E-mail address: [email protected] (L.E. Bilston).
0021-9290/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2011.04.034
brain injury, and surgical planning. To date, there have been a great number of rheological studies of the viscoelastic properties of the brain, generating data that vary by a considerable margin, reflecting the heterogeneity in methods employed by different research groups. Researchers have performed ex-vivo studies of cadaveric and animal brain tissues (e.g. (Bilston et al., 1997; Donnelly and Medige, 1997; Bilston et al., 2001; Miller and Chinzei, 2002; Hrapko et al., 2006)), measured the poroelastic properties (e.g. (Franceschini et al., 2006; Cheng and Bilston, 2007)), and developed many different types of constitutive models (e.g. (Bilston et al., 2001; Darvish and Crandall, 2001; Brands et al., 2004; Hrapko et al., 2006)). Brain tissue mechanical properties have been recently reviewed in detail elsewhere (Cheng et al., 2008). While the earliest studies only measured elastic properties of the brain (e.g. elastic shear modulus; McCracken et al., 2005), in recent years, the advent of Magnetic Resonance Elastography (MRE) has made possible the non-invasive measurement of brain viscoelasticity in living human
1910
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
subjects. This has led to several studies targeted at the viscoelastic properties of the brain (Green et al., 2008; Sack et al., 2008), which provide estimates of both the elastic (G0 ) and viscous (G00 ) components of the shear modulus. Results from MRE studies demonstrate a greater level of congruence, but some discrepancies in measured shear moduli persist due to differences in the methodology employed, particularly in the reconstruction algorithms used to extract the material properties from the MRE images. All previous MRE studies of the brain have exclusively focussed on the cerebrum, while no rheological measurements of the cerebellum (using any method) have been made. The cerebellum plays an important role in the control of motor functions; recent evidence suggests that it may also be involved in higher cognitive functions (Cantelmi et al., 2008; Puget et al., 2009). The cerebellum is sometimes affected in traumatic brain injury even when the initial impact is directed elsewhere in the brain, and mechanical forces caused by rapid deceleration are thought to be the cause of some of the pathological changes seen in the cerebellum (Gennarelli et al., 1982; Potts et al., 2009). Similarly, the biomechanics of the cerebellum is important in conditions where infratentorial herniation is a possibility. As for the cerebral hemispheres, it is possible that disease processes in the cerebellum might result in detectable mechanical changes. It is therefore desirable to investigate the viscoelastic properties of the cerebellum, in order to understand better how cerebellar tissues may respond to injuries and diseases. Current models of brain trauma typically assume that the cerebellum has similar mechanical properties to the remainder of the brain, in the absence of cerebellum mechanical properties data. The present study aims to characterize the viscoelastic properties of the living human cerebellum in young healthy subjects, and to compare these data to those from the cerebrum of the same subjects. We hypothesised that the cerebellum would be softer than the cerebrum, based on the fine microstructure of the cerebellum tissue.
2. Methods Eight healthy adult subjects (five males and three females, aged 22–43 years) with no history of neurological or psychiatric disorders volunteered for this study. Approval from the University of NSW Human Research Ethics Committee was obtained prior to data collection, and all subjects gave written informed consent. The subjects were fitted into a custom made oscillatory transducer, which produced a mechanical wave in the subjects’ brains through a mouthpiece. Details of these methods have been described previously (Green et al., 2008). The transducer was driven by a pulse generator triggered by the MRI spectrometer, in order to allow the synchronisation of the MRI pulse sequence with the mechanical wave. Imaging was conducted in the coronal plane, with the central slice located in the mid-region of the cerebellum. A Philips 3T scanner (Achieva 3TX. Philips Medical Systems, Best, The
Netherlands) was used. Imaging parameters were matrix 64 64, TE/TR¼50/700 ms, seven slices, eight time dynamics, FOV¼192 192, 3 mm slice thickness, and vibration frequency¼ 80 Hz. A matching T2 weighted anatomical image set (256 256) was also collected for ROI selection. The shear storage (G0 ) and loss moduli (G00 ), which represent the elastic and viscous components of the shear modulus, respectively, were then reconstructed from the 3D displacement field by solving the wave equation for propagation through an isotropic linear viscoelastic solid, using the curl operator to remove the contribution of the longitudinal waves. Full explanation of the theory behind the reconstruction and details of its implementation are given in Sinkus et al. (2005) and Green et al. (2008). The T2 anatomical images were used to manually select regions of interest (ROIs) consisting of the cerebellum or the parenchyma of the cerebral hemispheres, excluding the ventricles (see Fig. 2). ROIs were then copied geometrically to maps of G0 and G00 , and mean viscoelastic properties for each subject were calculated by taking the average of pixels contained within the ROIs. Averages and standard deviations across all subjects were then calculated from the individual subject data. Statistical comparisons were performed using the generalised linear model (GLM) method, with separate analyses for G0 and G00 . Comparisons were made for the brain region (cerebrum vs cerebellum) and grey and white matter. Potential interactions between region and tissue type were also tested. Significance was set at a ¼0.05.
3. Results A representative displacement field is shown in Fig. 1, demonstrating good wave penetration into the deep portions of the cerebral hemispheres and the cerebellum despite the inevitable reduction in amplitude between the outer edges of the cerebral hemispheres to the deep white matter from attenuation due to the viscoelasticity of brain tissue. Viscosity, in the context of wave propagation, has this effect of reducing wave amplitude as the wave propagates. This can be seen clearly in Figs. 1 and 2(c). Wave amplitude in the cerebellum was similar to the cerebral hemispheres. Representative viscoelastic data from the same subject are shown in Fig. 3, where one can differentiate between regions of high and low viscoelasticity, with low elasticity in the region of the ventricles as expected from a liquid. A summary of the storage and loss moduli and results of the statistical analysis are shown in Tables 1 and 2, respectively. For the storage modulus, the cerebellum was significantly softer than the cerebrum (GLM, p o0.001), but there were no significant differences between white and grey matter in either location. For the loss modulus (i.e. viscous properties), there were significant differences between the cerebrum and the cerebellum (p ¼0.037) and also between white and grey matter (p ¼0.04), but there was no significant interaction between region and white or grey matter (p4 0.5). In both regions, the white matter appeared to have a slightly higher loss modulus than the grey matter but this was not statistically significant. The means and standard deviations for each tissue type and region are plotted in Fig. 3.
Fig. 1. Representative displacement images from typical subject in mm: (a) displacement in X direction, (b) displacement in Y direction, and (c) displacement in Z direction. The wave amplitude decreases towards the centre of the brain as waves propagate inwards, but waves are clearly visible throughout the brain and cerebellum.
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
1911
Fig. 2. Representative viscoelastic data: (a) T2 weighted anatomical image, (b) the Storage Modulus (G0 , in kPa), and (c) the Loss Modulus (G00 ,in kPa). Spatial averages were generated for each ROI (examples shown in (a)) defined using T2 weighted images as shown (cyan ¼ grey matter, red ¼ white matter).
3.0
Table 2 Present data compared with previously published in vivo MRE data (mean7 SD); differences in the absolute values of the Shear Modulus reported could be due to differences in experimental methodology and reconstruction algorithms.
G' G"
Shear Moduli (kPa)
2.5
Frequency (Hz) G0 (kPa)
2.0 1.5 1.0 0.5 0.0 Cerebellum White
Cerebellum Grey
Cerebrum White
Cerebrum Grey
Region
Grey matter Human—present study Human—(Green et al., 2008) Human—(Kruse et al., 2008) Human—(McCracken et al., 2005)
80 90 100 80
2.34 7 0.22 1.11 70.03 3.017 0.10 2.50 70.20 5.22 7 0.23* 5.307 1.30*
White matter Human—present study Human—(Green et al., 2008) Human—(Kruse et al., 2008) Human—(McCracken et al., 2005)
80 90 100 80
2.41 7 0.23 1.21 70.21 2.707 0.1 2.50 70.20 13.6 7 0.66* 10.77 1.4*
White þGrey matter Human—(Hamhaber et al., 2007) Human—(Sack et al., 2009) Human—(Klatt et al., 2007)
83.33 62.5 62.5
3.507 1.33 2.017 0.23 0.84–2.28
0.80 70.13 0.57–2.96
n These data were obtained using an elastic model, and this value is the elastic shear modulus, G.
Fig. 3. Mean and standard deviation of G0 and G00 for the white and grey matter of the cerebellum and cerebrum (N ¼eight human subjects). Cerebellum is softer than the cerebrum (G0 , po 0.001, GLM). White matter is more viscous than grey matter (G00 ,GLM, p¼ 0.04), and the cerebrum is more viscous than the cerebellum (p ¼0.037).
Table 1 Means and standard deviations for the storage and loss moduli by region and tissue type. Region
Tissue type
G0 (kPa)
G00 (kPa)
Cerebellum
Grey matter White matter Grey matter White matter
1.77 70.24 1.85 70.18 2.34 70.22 2.41 70.23
0.94 7 0.17 1.10 7 0.23 1.11 7 0.03 1.21 7 0.21
Cerebrum
G00 (kPa)
4. Discussion These results represent the first available dataset on the viscoelastic properties of the cerebellum. They suggest that the cerebellum is softer than the cerebral hemispheres, but that there is no difference in viscosity between the two brain regions. The moduli for the cerebrum are comparable to the published data from the literature. Green et al. (2008) found the G0 and G00 of cerebral grey matter to be 3.1 and 2.5 kPa, respectively, and the G0 and G00 of cerebral white matter to be 2.7 and 2.5 kPa, using a
similar reconstruction technique at an oscillatory frequency of 90 Hz, which is slightly higher than that used in the current study. Higher frequency would be expected to increase the shear modulus estimates slightly due to the frequency dependence of brain mechanical properties (Cheng et al., 2008). Other MRE studies of brain tissue include McCracken et al. (2005), who determined the elastic shear moduli of grey and white matter to be 5.3 and 10.7 kPa, respectively, at 80 Hz; and Kruse et al. (2008), who determined the elastic shear moduli of grey and white matter to be 5.2 and 13.6 kPa, respectively, at 100 Hz. Some MRE studies have not distinguished between grey and white matter; these include Sack et al. (2009) who found the G0 and G00 of brain tissue to be 2.01 and 0.8 kPa, respectively, at 62.5 Hz; and Klatt et al. (2007) who used a number of different viscoelastic models to calculate the G0 and G00 of brain tissue to be between 0.84 and 2.28, and 0.57 and 2.96 kPa, respectively, at 62.5 Hz. Lastly, Hamhaber et al. (2007) found G0 ¼3.5 kPa at 83 Hz. Present data is compared to these results in Table 2. There is no prior published data on the viscoelastic properties of cerebellum. While the cerebrum data are at the lower end of the published range and consistent with the data from the Berlin group (Klatt et al. 2007; Hamhaber et al. 2007) there are differences in reconstruction technique that might account for the higher values obtained by the Mayo Clinic group (McCracken et al. 2005; Kruse et al. 2008). The McCracken et al. (2005) and Kruse et al. (2008) studies use an
1912
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
analytical technique that tends to overestimate the elastic shear modulus in regions of low signal-to-noise ratio, such as the deeper brain structures. This can be seen in their data (see Figure 4 in (Kruse et al., 2008)). Also that technique does not estimate the viscoelastic properties, as the tissue is assumed to be purely elastic without any loss component. The finding that the cerebellum is slightly ( 23–24%) softer than the cerebrum is consistent with our initial hypothesis. One possible explanation for this lies in the delicate, branching ultrastructure of cerebellum, including the molecular layer and granular cell layers. This can be readily observed in histology of the cerebellum (Young and Wheater, 2006). There are also proportionally many fewer glial cells in the cerebellum than the cerebral cortex, and thus the cells may be less tightly bound together (Azevedo et al., 2009) and all of these factors may influence the mechanical properties, although the details of why this might be the case are not known. The apparent differences in viscous properties do not have an obvious explanation, and further studies are needed to confirm whether this is a real difference or a chance finding. There are several limitations of the study to keep in mind when considering these results. Firstly, due to the fine structure of the cerebellum, it was not possible to perfectly separate white and grey matter in much of this region without partial volume effects, due to the voxel size (3 mm) required for good signal quality. These effects were minimised as much as possible by careful ROI selection. Moreover, due to the very fine structure of the sulci of the cerebellum, it is not possible to be completely certain that there is not a small amount of CSF in the cerebellum regions of interest for grey matter. This is unlikely to account for the differences in observed mechanical properties as this affected only the grey matter, and similar differences were seen in the deeper white matter where this is unlikely to be a significant effect. This is also true for the cerebral hemispheres, but to a lesser extent, since the sulci are larger and thus easier to exclude from the regions of interest. Lastly, the method employed in this study assumes the brain to be isotropic in nature, despite of evidence that white matter at least should be treated as anisotropic (Prange and Margulies, 2002). Methods of analysing anisotropic tissues in MRE are still under development (Green et al., 2009). Despite these limitations, these results represent the first available data on the viscoelastic properties of the cerebellum. They suggest that the cerebellum is less physically stiff than the cerebrum, and these data may assist in other research efforts to understand the brain’s response to mechanical loading, including studies of injury and structural neurological disorders. Further research is needed to confirm these findings, and to investigate why the cerebellum is less stiff than the cerebrum, and whether it responds differently to the same forces.
5. Conclusion MR elastographic measurements on eight healthy adult subjects were analysed to compare the viscoelasticity of the cerebrum and the cerebellum. It was found that G0 (elastic shear modulus) of the cerebellum is lower than that of the cerebral hemispheres. The results suggest that the cerebellum may respond differently to mechanical loading than the cerebrum.
Conflict of interest statement None declared.
Acknowledgements The authors would like to thank the staff of the Neuroscience Research Australia imaging centre for their assistance with imaging experiments. This research was funded by a discovery Grant from the Australian Research Council. Lynne Bilston is supported by an NHMRC senior research fellowship.
References Azevedo, F.A.C., Carvalho, L.R.B., Grinberg, L.T., Farfel, J.M., Ferretti, R.E.L., Leite, R.E.P., Filho, W.J., Lent, R., Herculano-Houzel, S., 2009. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. The Journal of Comparative Neurology 513 (5), 532–541. Bilston, L.E., Liu, Z., Phan-Thien, N., 1997. Linear viscoelastic properties of bovine brain tissue in shear. Biorheology 34 (6), 377–385. Bilston, L.E., Liu, Z., Phan-Thien, N., 2001. Large strain behaviour of brain tissue in shear: some experimental data and differential constitutive model. Biorheology 38 (4), 335–345. Brands, D.W.A., Peters, G.W.M., Bovendeerd, P.H.M., 2004. Design and numerical implementation of a 3-d non-linear viscoelastic constitutive model for brain tissue during impact. Journal of Biomechanics 37 (1), 127–134. Cantelmi, D., Schweizer, T., Cusimano, M., 2008. Role of the cerebellum in the neurocognitive sequelae of treatment of tumours of the posterior fossa: an update. The Lancet Oncology 9 (6), 569–576. Cheng, S., Bilston, L.E., 2007. Unconfined compression of white matter. Journal of Biomechanics 40 (1), 117–124. Cheng, S., Clarke, E.C., Bilston, L.E., 2008. Rheological properties of the tissues of the central nervous system: a review. Medical Engineering and Physics 30 (10), 1318–1337. Darvish, K.K., Crandall, J.R., 2001. Nonlinear viscoelastic effects in oscillatory shear deformation of brain tissue. Medical Engineering and Physics 23 (9), 633–645. Donnelly, B.R., Medige, J., 1997. Shear properties of human brain tissue. Journal of Biomechanical Engineering 119 (4), 423–432. Dutta-Roy, T., Wittek, A., Miller, K., 2008. Biomechanical modelling of normal pressure hydrocephalus. Journal of Biomechanics 41 (10), 2263–2271. Franceschini, G., Bigoni, D., Regitnig, P., Holzapfel, G.A., 2006. Brain tissue deforms similarly to filled elastomers and follows consolidation theory. Journal of the Mechanics and Physics of Solids 54 (12), 2592–2620. Gennarelli, T.A., Thibault, L.E., Adams, J.H., Graham, D.I., Thompson, C.J., Marcincin, R.P., 1982. Diffuse axonal injury and traumatic coma in the primate. Annals of Neurology 12 (6), 564–574. Green, M.A., Bilston, L.E., Sinkus, R., 2008. In vivo brain viscoelastic properties measured by magnetic resonance elastography. NMR in Biomedicine 21 (7), 755–764. Green, M.A., Bilston, L.E., Van Houten, E., Sinkus, R., 2009. In-vivo brain viscoelastic anisotropic properties using DTI and MR-elastography. In: Proceedings of the 17th ISMRM, Honolulu, HI. Hamhaber, U., Sack, I., Papazoglou, S., Rump, J., Klatt, D., Braun, J., 2007. Threedimensional analysis of shear wave propagation observed by in vivo magnetic resonance elastography of the brain. Acta Biomaterialia 3 (1), 127–137. Hrapko, M., Van Dommelen, J.a.W., Peters, G.W.M., Wismans, J.S.H.M., 2006. The mechanical behaviour of brain tissue: large strain response and constitutive modelling. Biorheology 43 (5), 623–636. Klatt, D., Hamhaber, U., Asbach, P., Braun, J., Sack, I., 2007. Noninvasive assessment of the rheological behavior of human organs using multifrequency mr elastography: a study of brain and liver viscoelasticity. Physics in Medicine and Biology 52 (24), 72–81. Kruse, S.A., Rose, G.H., Glaser, K.J., Manduca, A., Felmlee, J.P., Jack Jr., C.R., Ehman, R.L., Kruse, S.A., Rose, G.H., Glaser, K.J., Manduca, A., Felmlee, J.P., Jack Jr., C.R., Ehman, R.L., 2008. Magnetic resonance elastography of the brain. Neuroimage 39 (1), 231–237. McCracken, P.J., Manduca, A., Felmlee, J., Ehman, R.L., 2005. Mechanical transientbased magnetic resonance elastography. Magnetic Resonance in Medicine 53 (3), 628–639. Miller, K., Chinzei, K., 2002. Mechanical properties of brain tissue in tension. Journal of Biomechanics 35 (4), 483–490. Pang, D., Altschuler, E., 1994. Low-pressure hydrocephalic state and viscoelastic alterations in the brain. Neurosurgery 35 (4), 643–655. Potts, M., Adwanikar, H., Noble-Haeusslein, L., 2009. Models of traumatic cerebellar injury. The Cerebellum 8 (3), 211–221. Prange, M.T., Margulies, S.S., 2002. Regional, directional, and age-dependent properties of the brain undergoing large deformation. Journal of Biomechanical Engineering 124 (2), 244–252. Puget, S., Boddaert, N., Viguier, D., Kieffer, V., Bulteau, C., Garnett, M., Callu, D., Sainte-Rose, C., Kalifa, C., Dellatolas, G., 2009. Injuries to inferior vermis and dentate nuclei predict poor neurological and neuropsychological outcome in children with malignant posterior fossa tumors. Cancer 115 (6), 1338–1347.
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
Sack, I., Beierbach, B., Hamhaber, U., Klatt, D., Braun, J., Sack, I., Beierbach, B., Hamhaber, U., Klatt, D., Braun, J., 2008. Non-invasive measurement of brain viscoelasticity using magnetic resonance elastography. NMR in Biomedicine 21 (3), 265–271. Sack, I., Beierbach, B., Wuerfel, J., Klatt, D., Hamhaber, U., Papazoglou, S., Martus, P., Braun, J., Sack, I., Beierbach, B., Wuerfel, J., Klatt, D., Hamhaber, U., Papazoglou, S., Martus, P., Braun, J., 2009. The impact of aging and gender on brain viscoelasticity. Neuroimage 46 (3), 652–657.
1913
Sinkus, R., Tanter, M., Xydeas, T., Catheline, S., Bercoff, J., Fink, M., 2005. Viscoelastic shear properties of in vivo breast lesions measured by MR elastography. Magnetic Resonance Imaging 23 (2), 159–165. Wuerfel, J., Paul, F., Beierbach, B., Hamhaber, U., Klatt, D., Papazoglou, S., Zipp, F., Martus, P., Braun, J., Sack, I., 2010. MR-elastography reveals degradation of tissue integrity in multiple sclerosis. NeuroImage 49 (3), 2520–2525. Young, B., Wheater, P., 2006. Wheater’s Functional Histology: A Text and Colour Atlas. Churchill Livingstone, Philadelphia.
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Viscoelastic properties of human cerebellum using magnetic resonance elastography John Zhang a,b, Michael A. Green a, Ralph Sinkus c, Lynne E. Bilston a,d,n a
Neuroscience Research Australia, Barker Street, Randwick NSW 2031, Australia Faculty of Medicine University of New South Wales, Kensington, NSW 2052, Australia c INSERM U773, CRB3, Centre de Recherches Biome´dicales Bichat-Beaujon, Paris, France d Prince of Wales Clinical School, University of New South Wales, Kensington NSW 2052, Australia b
a r t i c l e i n f o
abstract
Article history: Accepted 19 April 2011
Background: The cerebellum has never been mechanically characterised, despite its physiological importance in the control of motion and the clinical prevalence of cerebellar pathologies. The aim of this study was to measure the linear viscoelastic properties of the cerebellum in human volunteers using Magnetic Resonance Elastography (MRE). Methods: Coronal plane brain 3D MRE data was performed on eight healthy adult volunteers, at 80 Hz, to compare the properties of cerebral and cerebellar tissues. The linear viscoelastic storage (G0 ) and loss moduli (G00 ) were estimated from the MRE wave images by solving the wave equation for propagation through an isotropic linear viscoelastic solid. Contributions of the compressional wave were removed via application of the curl-operator. Results: The storage modulus for the cerebellum was found to be significantly lower than that for the cerebrum, for both white and grey matter. Cerebrum: white matter (mean 7SD) G0 ¼ 2.41 7 0.23 kPa, grey matter G0 ¼2.34 70.22 kPa; cerebellum: white matter, G0 ¼1.85 70.18 kPa, grey matter G0 ¼ 1.77 7 0.24 kPa; cerebrum vs cerebellum, p o0.001. For the viscous behaviour, there were differences in between regions and also by tissue type, with the white matter being more viscous than grey matter and the cerebrum more viscous than the cerebellum. Cerebrum: white matter G00 ¼ 1.21 7 0.21 kPa, grey matter G00 ¼ 1.11 7 0.03 kPa; cerebellum: white matter G00 ¼ 1.1 7 0.23 kPa, grey matter G00 ¼ 0.94 70.17 kPa. Discussion: These data represent the first available data on the viscoelastic properties of cerebellum, which suggest that the cerebellum is less physically stiff than the cerebrum, possibly leading to a different response to mechanical loading. These data will be useful for modelling of the cerebellum for a range of purposes. & 2011 Elsevier Ltd. All rights reserved.
Keywords: Brain viscoelasticity Shear modulus Magnetic resonance imaging In vivo measurements
1. Introduction The viscoelastic properties of brain tissue play a role in a variety of neurological disorders and traumatic brain injury. It has been postulated that certain conditions, such as normal pressure hydrocephalus, are influenced by a change in the mechanical properties of the brain tissue (Pang and Altschuler, 1994; Dutta-Roy et al., 2008), and those changes in the viscoelasticity of the brain may be a marker for neurodegenerative conditions such as Alzheimer’s disease and multiple sclerosis (Kruse et al., 2008; Wuerfel et al., 2010). In addition, these parameters are vital for computational simulations such as Finite Element Analysis (FEA) of brain conditions, traumatic
n Corresponding author at: Neuroscience Research Australia, Barker St, Randwick, NSW 2031, Australia. E-mail address: [email protected] (L.E. Bilston).
0021-9290/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2011.04.034
brain injury, and surgical planning. To date, there have been a great number of rheological studies of the viscoelastic properties of the brain, generating data that vary by a considerable margin, reflecting the heterogeneity in methods employed by different research groups. Researchers have performed ex-vivo studies of cadaveric and animal brain tissues (e.g. (Bilston et al., 1997; Donnelly and Medige, 1997; Bilston et al., 2001; Miller and Chinzei, 2002; Hrapko et al., 2006)), measured the poroelastic properties (e.g. (Franceschini et al., 2006; Cheng and Bilston, 2007)), and developed many different types of constitutive models (e.g. (Bilston et al., 2001; Darvish and Crandall, 2001; Brands et al., 2004; Hrapko et al., 2006)). Brain tissue mechanical properties have been recently reviewed in detail elsewhere (Cheng et al., 2008). While the earliest studies only measured elastic properties of the brain (e.g. elastic shear modulus; McCracken et al., 2005), in recent years, the advent of Magnetic Resonance Elastography (MRE) has made possible the non-invasive measurement of brain viscoelasticity in living human
1910
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
subjects. This has led to several studies targeted at the viscoelastic properties of the brain (Green et al., 2008; Sack et al., 2008), which provide estimates of both the elastic (G0 ) and viscous (G00 ) components of the shear modulus. Results from MRE studies demonstrate a greater level of congruence, but some discrepancies in measured shear moduli persist due to differences in the methodology employed, particularly in the reconstruction algorithms used to extract the material properties from the MRE images. All previous MRE studies of the brain have exclusively focussed on the cerebrum, while no rheological measurements of the cerebellum (using any method) have been made. The cerebellum plays an important role in the control of motor functions; recent evidence suggests that it may also be involved in higher cognitive functions (Cantelmi et al., 2008; Puget et al., 2009). The cerebellum is sometimes affected in traumatic brain injury even when the initial impact is directed elsewhere in the brain, and mechanical forces caused by rapid deceleration are thought to be the cause of some of the pathological changes seen in the cerebellum (Gennarelli et al., 1982; Potts et al., 2009). Similarly, the biomechanics of the cerebellum is important in conditions where infratentorial herniation is a possibility. As for the cerebral hemispheres, it is possible that disease processes in the cerebellum might result in detectable mechanical changes. It is therefore desirable to investigate the viscoelastic properties of the cerebellum, in order to understand better how cerebellar tissues may respond to injuries and diseases. Current models of brain trauma typically assume that the cerebellum has similar mechanical properties to the remainder of the brain, in the absence of cerebellum mechanical properties data. The present study aims to characterize the viscoelastic properties of the living human cerebellum in young healthy subjects, and to compare these data to those from the cerebrum of the same subjects. We hypothesised that the cerebellum would be softer than the cerebrum, based on the fine microstructure of the cerebellum tissue.
2. Methods Eight healthy adult subjects (five males and three females, aged 22–43 years) with no history of neurological or psychiatric disorders volunteered for this study. Approval from the University of NSW Human Research Ethics Committee was obtained prior to data collection, and all subjects gave written informed consent. The subjects were fitted into a custom made oscillatory transducer, which produced a mechanical wave in the subjects’ brains through a mouthpiece. Details of these methods have been described previously (Green et al., 2008). The transducer was driven by a pulse generator triggered by the MRI spectrometer, in order to allow the synchronisation of the MRI pulse sequence with the mechanical wave. Imaging was conducted in the coronal plane, with the central slice located in the mid-region of the cerebellum. A Philips 3T scanner (Achieva 3TX. Philips Medical Systems, Best, The
Netherlands) was used. Imaging parameters were matrix 64 64, TE/TR¼50/700 ms, seven slices, eight time dynamics, FOV¼192 192, 3 mm slice thickness, and vibration frequency¼ 80 Hz. A matching T2 weighted anatomical image set (256 256) was also collected for ROI selection. The shear storage (G0 ) and loss moduli (G00 ), which represent the elastic and viscous components of the shear modulus, respectively, were then reconstructed from the 3D displacement field by solving the wave equation for propagation through an isotropic linear viscoelastic solid, using the curl operator to remove the contribution of the longitudinal waves. Full explanation of the theory behind the reconstruction and details of its implementation are given in Sinkus et al. (2005) and Green et al. (2008). The T2 anatomical images were used to manually select regions of interest (ROIs) consisting of the cerebellum or the parenchyma of the cerebral hemispheres, excluding the ventricles (see Fig. 2). ROIs were then copied geometrically to maps of G0 and G00 , and mean viscoelastic properties for each subject were calculated by taking the average of pixels contained within the ROIs. Averages and standard deviations across all subjects were then calculated from the individual subject data. Statistical comparisons were performed using the generalised linear model (GLM) method, with separate analyses for G0 and G00 . Comparisons were made for the brain region (cerebrum vs cerebellum) and grey and white matter. Potential interactions between region and tissue type were also tested. Significance was set at a ¼0.05.
3. Results A representative displacement field is shown in Fig. 1, demonstrating good wave penetration into the deep portions of the cerebral hemispheres and the cerebellum despite the inevitable reduction in amplitude between the outer edges of the cerebral hemispheres to the deep white matter from attenuation due to the viscoelasticity of brain tissue. Viscosity, in the context of wave propagation, has this effect of reducing wave amplitude as the wave propagates. This can be seen clearly in Figs. 1 and 2(c). Wave amplitude in the cerebellum was similar to the cerebral hemispheres. Representative viscoelastic data from the same subject are shown in Fig. 3, where one can differentiate between regions of high and low viscoelasticity, with low elasticity in the region of the ventricles as expected from a liquid. A summary of the storage and loss moduli and results of the statistical analysis are shown in Tables 1 and 2, respectively. For the storage modulus, the cerebellum was significantly softer than the cerebrum (GLM, p o0.001), but there were no significant differences between white and grey matter in either location. For the loss modulus (i.e. viscous properties), there were significant differences between the cerebrum and the cerebellum (p ¼0.037) and also between white and grey matter (p ¼0.04), but there was no significant interaction between region and white or grey matter (p4 0.5). In both regions, the white matter appeared to have a slightly higher loss modulus than the grey matter but this was not statistically significant. The means and standard deviations for each tissue type and region are plotted in Fig. 3.
Fig. 1. Representative displacement images from typical subject in mm: (a) displacement in X direction, (b) displacement in Y direction, and (c) displacement in Z direction. The wave amplitude decreases towards the centre of the brain as waves propagate inwards, but waves are clearly visible throughout the brain and cerebellum.
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
1911
Fig. 2. Representative viscoelastic data: (a) T2 weighted anatomical image, (b) the Storage Modulus (G0 , in kPa), and (c) the Loss Modulus (G00 ,in kPa). Spatial averages were generated for each ROI (examples shown in (a)) defined using T2 weighted images as shown (cyan ¼ grey matter, red ¼ white matter).
3.0
Table 2 Present data compared with previously published in vivo MRE data (mean7 SD); differences in the absolute values of the Shear Modulus reported could be due to differences in experimental methodology and reconstruction algorithms.
G' G"
Shear Moduli (kPa)
2.5
Frequency (Hz) G0 (kPa)
2.0 1.5 1.0 0.5 0.0 Cerebellum White
Cerebellum Grey
Cerebrum White
Cerebrum Grey
Region
Grey matter Human—present study Human—(Green et al., 2008) Human—(Kruse et al., 2008) Human—(McCracken et al., 2005)
80 90 100 80
2.34 7 0.22 1.11 70.03 3.017 0.10 2.50 70.20 5.22 7 0.23* 5.307 1.30*
White matter Human—present study Human—(Green et al., 2008) Human—(Kruse et al., 2008) Human—(McCracken et al., 2005)
80 90 100 80
2.41 7 0.23 1.21 70.21 2.707 0.1 2.50 70.20 13.6 7 0.66* 10.77 1.4*
White þGrey matter Human—(Hamhaber et al., 2007) Human—(Sack et al., 2009) Human—(Klatt et al., 2007)
83.33 62.5 62.5
3.507 1.33 2.017 0.23 0.84–2.28
0.80 70.13 0.57–2.96
n These data were obtained using an elastic model, and this value is the elastic shear modulus, G.
Fig. 3. Mean and standard deviation of G0 and G00 for the white and grey matter of the cerebellum and cerebrum (N ¼eight human subjects). Cerebellum is softer than the cerebrum (G0 , po 0.001, GLM). White matter is more viscous than grey matter (G00 ,GLM, p¼ 0.04), and the cerebrum is more viscous than the cerebellum (p ¼0.037).
Table 1 Means and standard deviations for the storage and loss moduli by region and tissue type. Region
Tissue type
G0 (kPa)
G00 (kPa)
Cerebellum
Grey matter White matter Grey matter White matter
1.77 70.24 1.85 70.18 2.34 70.22 2.41 70.23
0.94 7 0.17 1.10 7 0.23 1.11 7 0.03 1.21 7 0.21
Cerebrum
G00 (kPa)
4. Discussion These results represent the first available dataset on the viscoelastic properties of the cerebellum. They suggest that the cerebellum is softer than the cerebral hemispheres, but that there is no difference in viscosity between the two brain regions. The moduli for the cerebrum are comparable to the published data from the literature. Green et al. (2008) found the G0 and G00 of cerebral grey matter to be 3.1 and 2.5 kPa, respectively, and the G0 and G00 of cerebral white matter to be 2.7 and 2.5 kPa, using a
similar reconstruction technique at an oscillatory frequency of 90 Hz, which is slightly higher than that used in the current study. Higher frequency would be expected to increase the shear modulus estimates slightly due to the frequency dependence of brain mechanical properties (Cheng et al., 2008). Other MRE studies of brain tissue include McCracken et al. (2005), who determined the elastic shear moduli of grey and white matter to be 5.3 and 10.7 kPa, respectively, at 80 Hz; and Kruse et al. (2008), who determined the elastic shear moduli of grey and white matter to be 5.2 and 13.6 kPa, respectively, at 100 Hz. Some MRE studies have not distinguished between grey and white matter; these include Sack et al. (2009) who found the G0 and G00 of brain tissue to be 2.01 and 0.8 kPa, respectively, at 62.5 Hz; and Klatt et al. (2007) who used a number of different viscoelastic models to calculate the G0 and G00 of brain tissue to be between 0.84 and 2.28, and 0.57 and 2.96 kPa, respectively, at 62.5 Hz. Lastly, Hamhaber et al. (2007) found G0 ¼3.5 kPa at 83 Hz. Present data is compared to these results in Table 2. There is no prior published data on the viscoelastic properties of cerebellum. While the cerebrum data are at the lower end of the published range and consistent with the data from the Berlin group (Klatt et al. 2007; Hamhaber et al. 2007) there are differences in reconstruction technique that might account for the higher values obtained by the Mayo Clinic group (McCracken et al. 2005; Kruse et al. 2008). The McCracken et al. (2005) and Kruse et al. (2008) studies use an
1912
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
analytical technique that tends to overestimate the elastic shear modulus in regions of low signal-to-noise ratio, such as the deeper brain structures. This can be seen in their data (see Figure 4 in (Kruse et al., 2008)). Also that technique does not estimate the viscoelastic properties, as the tissue is assumed to be purely elastic without any loss component. The finding that the cerebellum is slightly ( 23–24%) softer than the cerebrum is consistent with our initial hypothesis. One possible explanation for this lies in the delicate, branching ultrastructure of cerebellum, including the molecular layer and granular cell layers. This can be readily observed in histology of the cerebellum (Young and Wheater, 2006). There are also proportionally many fewer glial cells in the cerebellum than the cerebral cortex, and thus the cells may be less tightly bound together (Azevedo et al., 2009) and all of these factors may influence the mechanical properties, although the details of why this might be the case are not known. The apparent differences in viscous properties do not have an obvious explanation, and further studies are needed to confirm whether this is a real difference or a chance finding. There are several limitations of the study to keep in mind when considering these results. Firstly, due to the fine structure of the cerebellum, it was not possible to perfectly separate white and grey matter in much of this region without partial volume effects, due to the voxel size (3 mm) required for good signal quality. These effects were minimised as much as possible by careful ROI selection. Moreover, due to the very fine structure of the sulci of the cerebellum, it is not possible to be completely certain that there is not a small amount of CSF in the cerebellum regions of interest for grey matter. This is unlikely to account for the differences in observed mechanical properties as this affected only the grey matter, and similar differences were seen in the deeper white matter where this is unlikely to be a significant effect. This is also true for the cerebral hemispheres, but to a lesser extent, since the sulci are larger and thus easier to exclude from the regions of interest. Lastly, the method employed in this study assumes the brain to be isotropic in nature, despite of evidence that white matter at least should be treated as anisotropic (Prange and Margulies, 2002). Methods of analysing anisotropic tissues in MRE are still under development (Green et al., 2009). Despite these limitations, these results represent the first available data on the viscoelastic properties of the cerebellum. They suggest that the cerebellum is less physically stiff than the cerebrum, and these data may assist in other research efforts to understand the brain’s response to mechanical loading, including studies of injury and structural neurological disorders. Further research is needed to confirm these findings, and to investigate why the cerebellum is less stiff than the cerebrum, and whether it responds differently to the same forces.
5. Conclusion MR elastographic measurements on eight healthy adult subjects were analysed to compare the viscoelasticity of the cerebrum and the cerebellum. It was found that G0 (elastic shear modulus) of the cerebellum is lower than that of the cerebral hemispheres. The results suggest that the cerebellum may respond differently to mechanical loading than the cerebrum.
Conflict of interest statement None declared.
Acknowledgements The authors would like to thank the staff of the Neuroscience Research Australia imaging centre for their assistance with imaging experiments. This research was funded by a discovery Grant from the Australian Research Council. Lynne Bilston is supported by an NHMRC senior research fellowship.
References Azevedo, F.A.C., Carvalho, L.R.B., Grinberg, L.T., Farfel, J.M., Ferretti, R.E.L., Leite, R.E.P., Filho, W.J., Lent, R., Herculano-Houzel, S., 2009. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. The Journal of Comparative Neurology 513 (5), 532–541. Bilston, L.E., Liu, Z., Phan-Thien, N., 1997. Linear viscoelastic properties of bovine brain tissue in shear. Biorheology 34 (6), 377–385. Bilston, L.E., Liu, Z., Phan-Thien, N., 2001. Large strain behaviour of brain tissue in shear: some experimental data and differential constitutive model. Biorheology 38 (4), 335–345. Brands, D.W.A., Peters, G.W.M., Bovendeerd, P.H.M., 2004. Design and numerical implementation of a 3-d non-linear viscoelastic constitutive model for brain tissue during impact. Journal of Biomechanics 37 (1), 127–134. Cantelmi, D., Schweizer, T., Cusimano, M., 2008. Role of the cerebellum in the neurocognitive sequelae of treatment of tumours of the posterior fossa: an update. The Lancet Oncology 9 (6), 569–576. Cheng, S., Bilston, L.E., 2007. Unconfined compression of white matter. Journal of Biomechanics 40 (1), 117–124. Cheng, S., Clarke, E.C., Bilston, L.E., 2008. Rheological properties of the tissues of the central nervous system: a review. Medical Engineering and Physics 30 (10), 1318–1337. Darvish, K.K., Crandall, J.R., 2001. Nonlinear viscoelastic effects in oscillatory shear deformation of brain tissue. Medical Engineering and Physics 23 (9), 633–645. Donnelly, B.R., Medige, J., 1997. Shear properties of human brain tissue. Journal of Biomechanical Engineering 119 (4), 423–432. Dutta-Roy, T., Wittek, A., Miller, K., 2008. Biomechanical modelling of normal pressure hydrocephalus. Journal of Biomechanics 41 (10), 2263–2271. Franceschini, G., Bigoni, D., Regitnig, P., Holzapfel, G.A., 2006. Brain tissue deforms similarly to filled elastomers and follows consolidation theory. Journal of the Mechanics and Physics of Solids 54 (12), 2592–2620. Gennarelli, T.A., Thibault, L.E., Adams, J.H., Graham, D.I., Thompson, C.J., Marcincin, R.P., 1982. Diffuse axonal injury and traumatic coma in the primate. Annals of Neurology 12 (6), 564–574. Green, M.A., Bilston, L.E., Sinkus, R., 2008. In vivo brain viscoelastic properties measured by magnetic resonance elastography. NMR in Biomedicine 21 (7), 755–764. Green, M.A., Bilston, L.E., Van Houten, E., Sinkus, R., 2009. In-vivo brain viscoelastic anisotropic properties using DTI and MR-elastography. In: Proceedings of the 17th ISMRM, Honolulu, HI. Hamhaber, U., Sack, I., Papazoglou, S., Rump, J., Klatt, D., Braun, J., 2007. Threedimensional analysis of shear wave propagation observed by in vivo magnetic resonance elastography of the brain. Acta Biomaterialia 3 (1), 127–137. Hrapko, M., Van Dommelen, J.a.W., Peters, G.W.M., Wismans, J.S.H.M., 2006. The mechanical behaviour of brain tissue: large strain response and constitutive modelling. Biorheology 43 (5), 623–636. Klatt, D., Hamhaber, U., Asbach, P., Braun, J., Sack, I., 2007. Noninvasive assessment of the rheological behavior of human organs using multifrequency mr elastography: a study of brain and liver viscoelasticity. Physics in Medicine and Biology 52 (24), 72–81. Kruse, S.A., Rose, G.H., Glaser, K.J., Manduca, A., Felmlee, J.P., Jack Jr., C.R., Ehman, R.L., Kruse, S.A., Rose, G.H., Glaser, K.J., Manduca, A., Felmlee, J.P., Jack Jr., C.R., Ehman, R.L., 2008. Magnetic resonance elastography of the brain. Neuroimage 39 (1), 231–237. McCracken, P.J., Manduca, A., Felmlee, J., Ehman, R.L., 2005. Mechanical transientbased magnetic resonance elastography. Magnetic Resonance in Medicine 53 (3), 628–639. Miller, K., Chinzei, K., 2002. Mechanical properties of brain tissue in tension. Journal of Biomechanics 35 (4), 483–490. Pang, D., Altschuler, E., 1994. Low-pressure hydrocephalic state and viscoelastic alterations in the brain. Neurosurgery 35 (4), 643–655. Potts, M., Adwanikar, H., Noble-Haeusslein, L., 2009. Models of traumatic cerebellar injury. The Cerebellum 8 (3), 211–221. Prange, M.T., Margulies, S.S., 2002. Regional, directional, and age-dependent properties of the brain undergoing large deformation. Journal of Biomechanical Engineering 124 (2), 244–252. Puget, S., Boddaert, N., Viguier, D., Kieffer, V., Bulteau, C., Garnett, M., Callu, D., Sainte-Rose, C., Kalifa, C., Dellatolas, G., 2009. Injuries to inferior vermis and dentate nuclei predict poor neurological and neuropsychological outcome in children with malignant posterior fossa tumors. Cancer 115 (6), 1338–1347.
J. Zhang et al. / Journal of Biomechanics 44 (2011) 1909–1913
Sack, I., Beierbach, B., Hamhaber, U., Klatt, D., Braun, J., Sack, I., Beierbach, B., Hamhaber, U., Klatt, D., Braun, J., 2008. Non-invasive measurement of brain viscoelasticity using magnetic resonance elastography. NMR in Biomedicine 21 (3), 265–271. Sack, I., Beierbach, B., Wuerfel, J., Klatt, D., Hamhaber, U., Papazoglou, S., Martus, P., Braun, J., Sack, I., Beierbach, B., Wuerfel, J., Klatt, D., Hamhaber, U., Papazoglou, S., Martus, P., Braun, J., 2009. The impact of aging and gender on brain viscoelasticity. Neuroimage 46 (3), 652–657.
1913
Sinkus, R., Tanter, M., Xydeas, T., Catheline, S., Bercoff, J., Fink, M., 2005. Viscoelastic shear properties of in vivo breast lesions measured by MR elastography. Magnetic Resonance Imaging 23 (2), 159–165. Wuerfel, J., Paul, F., Beierbach, B., Hamhaber, U., Klatt, D., Papazoglou, S., Zipp, F., Martus, P., Braun, J., Sack, I., 2010. MR-elastography reveals degradation of tissue integrity in multiple sclerosis. NeuroImage 49 (3), 2520–2525. Young, B., Wheater, P., 2006. Wheater’s Functional Histology: A Text and Colour Atlas. Churchill Livingstone, Philadelphia.