FULL PAPER Insignificance of Active Flow for Neural Diffusion Weighted Imaging: A Negative Result Matan Mussel, 1 Lilah Inzelberg, 1,2 and Uri Nevo 1,2 * Purpose: To provide a biophysical basis to estimate the effect of cytoplasmic flow in neurons, and assess their contribution to the drop in the Apparent Diffusion Coefficient (ADC) in a nerve tissue following extreme conditions, such as brain injury and epileptic seizures. Methods: Three mechanisms are treated using the relevant physics of hydrodynamics and electrostatics: cargo induced streaming, electroosmosis, and membrane swelling. Results: We begin by discussing the lack of experimental evi- dence on the necessary velocities required to influence the Mag- netic Resonance (MR) experiments. This is followed by demonstrating that cargo induced streaming, a widely known phe- nomenon in plant cells, has a minor effect on the ADC in neurons. Subsequently, we suggest and analyze two additional mechanisms that may induce fluid displacement in neurons, and are related to the electrical activity: electroosmosis and membrane swelling. Conclusion: Although these mechanisms may induce interest- ing fluid displacements, these cannot explain the significant drop in the ADC. We conclude by outlining the criteria that any future mechanism should meet to have an influence on stan- dard diffusion-MR measurements. Magn Reson Med 000:000–000, 2016. V C 2016 International Society for Mag- netic Resonance in Medicine Key words: diffusion-MR; DWI; ADC; neurons; cytoplasmic streaming INTRODUCTION Diffusion weighted imaging (DWI) probes the displace- ment of water molecules by applying pulses of magnetic field gradients (1). The applied gradients encode molecu- lar positions by phases, such that the molecular dis- placement results in dephasing and in signal loss. It was demonstrated that a decrease in the measured apparent diffusion coefficient (ADC) of water in brain tissue occurs within minutes after the onset of ischemic stroke and nerve injury (2–4), drug induced epileptic seizures (5–7), cortical spreading depression (8,9), and after sen- sory or motor stimulation (10,11). The ADC coefficient reduces by 40–70%, 14–20%, 10–35%, and 1-4% respec- tively, in these conditions. Several hypotheses were suggested to explain the phe- nomenon of diffusion drop. These can be categorized into four classes: (1) morphological, structural or geomet- rical changes in the tissue microstructure (hereafter mor- phological changes); (2) changes in material properties; (3) neural functional changes; and (4) external factors unrelated to the neural component. Morphological mod- els, aimed to describe the change in ADC, are based on the fact that although diffusion in short time scales is similar in the intra- and extra-cellular compartments (12), in long diffusion times, the ADC is attenuated in the intracellular compartment by the membrane and by other intracellular barriers (13–15). Therefore, a drop in diffusion could be related to a sudden decrease in extra- cellular volume, a distortion of the extracellular space, or changes in intracellular compartments (14–18). A par- ticular example worth highlighting is axonal “beading”— localized dilations and constrictions of the axonal mem- brane (19–23). Such changes can potentially be detected by different modifications of the diffusion-MR technique (24–26). The conjectures that describe changes in materi- al properties suggest that the diffusion drop occurs by an alteration of either the polymeric structure within the cells (27,28), by changes in the membrane permeability (29) or by an increase of bound water characterized by a lower diffusivity (28). Functional changes are expected in cells after extreme events. Some of these ATP con- suming functions may induce microstreamings within the cells, and these intracellular changes could alter the DWI signal (5,12,13,30,31). Finally, it is possible that the drop in diffusion stems not from the neural component, but from external elements that attenuate the signal mea- surement. These include intravoxel incoherent motion of blood in the capillary network (IVIM) (32,33) and changes in blood-oxygen-level dependent (BOLD) con- trast (34,35). The variety of the above listed hypotheses highlights the complexity of the problem. The source of the dynam- ical changes in the ADC does not have to be confined to one mechanism, but rather comprise a combination of multiple components. A better understanding of the rela- tive contribution of each component may lead to an improved understanding of neural tissue dynamics, and possibly to optimization of DWI as a diagnostic tool (36). The above hypotheses on the source of dynamics in the ADC should be tested against particular experimental observations. A drop in the ADC was demonstrated in neural tissue in the absence of a significant hemodynam- ic response, an indication that the phenomenon is most likely to originate from the neural cells themselves (neu- rons and glia) (7,37). It was also shown that extracellular and intracellular fluids exhibit a simultaneous drop in their diffusion coefficient. This observation must be imposed as a restriction on any morphological model 1 The Iby and Aladar Fleischman Faculty of Engineering, Department of Bio- medical Engineering, Tel Aviv University, Tel Aviv, Israel. 2 Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel. Grant sponsor: Israeli Science Foundation to UN (1156/12). *Correspondence to: U. Nevo; Department of Biomedical Engineering, Tel Aviv University, Tel Aviv 69978, Israel. E-mail: nevouri@eng.tau.ac.il Received 25 May 2016; revised 28 June 2016; accepted 19 July 2016 DOI 10.1002/mrm.26375 Published online 00 Month 2016 in Wiley Online Library (wileyonlinelibrary. com). Magnetic Resonance in Medicine 00:00–00 (2016) V C 2016 International Society for Magnetic Resonance in Medicine 1