Biexponential diffusion attenuation in the rat spinal cord: Computer simulations based on anatomic images of axonal architecture

Abstract

Water diffusion in neurological tissues is known to possess multicomponent diffusion behavior. The fractions of fast and slow apparent diffusion components have often been attributed to the volume fractions of extracellular space (ECS) and intracellular space (ICS) although diffusion fractions are at variance with the tissue compartment volume ratios. In this article this puzzle was examined with a finite difference diffusion simulation model on the basis of optical images from sectioned rat spinal cord. Here the results show that assignment of fractions obtained from biexponential fits of fast and slow diffusion attenuation to ECS and ICS volume ratios is not correct. Rather, the observed multicomponent diffusion behavior is caused by motional restriction and limited intercompartmental water exchange in that at long diffusion times diffusion attenuation is shown to become monoexponential. Although the measured apparent diffusion fractions also depend on T 2 relaxation time of water protons in the various compartments, the sensitivity to T 2 is small and thus Biexponential NMR diffusion attenuation of water has been reported in both the brain (1-4) and the spinal cord (5-7), and the phenomenon appears to be inherent to all neural tissues. The observation has invariably been attributed to the presence of two different fractions corresponding to fast and slow apparent diffusion, associated with the volume fractions of extracellular spaces (ECS) and intracellular space (ICS). However, the discrepancies between the actual ECS and ICS volume ratios and the calculated fractions of fast and slow diffusion components remain unresolved.Biexponential water diffusion in both normal and ischemic rat brains was observed by Niendorf et al. (1). Based on the results of cell swelling and shrinking experiments and findings from electrical impedance measurements, the authors observed changes in the fractions of fast and slow diffusion components which were correlated with changes in the volume fractions of the ECS and ICS. However, the two fractions were found to deviate significantly from those of intra-and extracellular compartment volumes.Ignoring effects of internal gradients and their interaction with the external diffusion-sensitizing gradients, the differences between intra-and extracellular T 2 relaxation times, partial volume averaging, and water exchange between the intra-and extracellular compartments were considered to explain this discrepancy. Biexponential diffusion behavior in rat brain was described by Buckley et al. (3), who found that cell swelling induced by ouabain resulted in an increase in the fraction corresponding to the slow component. The mismatch between the expected apparent diffusion component fractions and physiological compartments was attributed to differences in T 2 relaxation times between ECS and ICS. Very recently, Inglis et al. (7) reported diffusion tensor data for fast and slow diffusion in rat spinal cords. Two components were observed in both gray matter and white matter, and the slow com...

Ex vivo diffusion tensor imaging and quantitative tractography of the rat spinal cord during long‐term recovery from moderate spinal contusion

“…In addition, DWIs had ␦ ϭ 3 ms, ⌬ ϭ 15 ms, and b-values ranging from 493.5-500.7s/mm 2 (target b-value ϭ 500 s/mm 2 ) depending on orientation of diffusion sensitizing gradients with respect to imaging gradients. A b-value of 500 s/mm 2 was chosen to provide diffusion measurements in the fast diffusion compartment exclusively (30). A total of 24 axial DTI images were obtained throughout the length of the spinal cord (ϳ7 cm).…”

High-Fidelity Meshes from Tissue Samples for Diffusion MRI Simulations

“…The most common numerical methods for synthesizing diffusion MRI data are based on finitedifference approaches e.g. [8] and Monte-Carlo methods e.g. [9,10,11].…”

“…Other work [8,11] base the simulations in models derived from light microscopy images for a better approximation of the tissue. Lipinski [11] was the first to use histologic images in combination with two-dimensional Monte-Carlo simulations to study the diffusion signal.…”

“…The tissue model is based on a rough segmentation of a digitized light microscopy image of white matter tissue. More recently, Chin et al [8] use a finite-difference approach and construct a tissue model by replicating a light microscopy image in the third dimension. These numerical models are an improvement on the simplified geometric models, yet, the singleslice microscopy images are low-resolution and the models describe the tissue only in two dimensions.…”

“…This paper presents a method for constructing three-dimensional tissue models from a stack of high-resolution Confocal Laser Scanning Microscopy (CLSM) images. We capture the three-dimensional structure of biological tissue in a more natural way than previous studies [8,11]. We demonstrate the method using a biological phantom (asparagus) which is a useful model with similar microstructure to white matter [12].…”