Energy dysregulation is considered a key contributing factor to neuro-axonal loss in multiple sclerosis (MS). Diffusion-weighted spectroscopy (DWS), allowing to measure in-vivo the diffusion properties of endogenous intracellular metabolites¹, has the potential to provide novel insights into the early phase of neuronal structural damage and energetic mismatch in MS. This technique combines specific measures of neuro-axonal damage, such as the diffusivity of total N-acetylaspartate (tNAA), an intra-neuronal amino acid−derivative, with measures of cell energy metabolism, such as the diffusivity of total creatine (tCr). tCr represents the sum of creatine and phosphocreatine, both intermediaries in the ATP-generating, creatine-kinase reaction, and reflects the status of energy metabolism in neuronal and glial cells. We employed DWS to explore the neuro-axonal damage and the ongoing energy dyregulation in the brain of patients with MS, and to investigate the clinical relevance of these processes.Twenty-five patients with MS (15 women, mean age=45.8±13yrs) were clinically assessed on the Expanded Disability Status Scale (EDSS)² and on the Timed-walked test (TWT)³ and underwent conventional imaging on a 3T whole-body Siemens scanner equipped with a 32-channel receive coil along with a group of 20 age- and gender-matched healthy controls (HC). A single-VOI diffusion-weighted PRESS⁴ sequence was employed to measure diffusion of tNAA and tCr. A VOI of dimensions 25(AP)x13(RL)x20(FH)mm³ was located in the white matter (WM) of the corona radiate (Fig1), outside visible lesions in patients. A second VOI of 23(AP)x30(LR)x18(FH) mm³ was positioned in the gray matter (GM) of the bilateral thalami (Fig2). Sequence parameters were: TR=3 cardiac cycles, TE=120ms, spectral width=2kHz, and number of sample points=1024. The diffusion weighting was applied in one direction using a diffusion time Δ=60 ms, duration of bipolar gradients δ=30ms, and gradient strength g=33mT/m, resulting in a b value of 3100s/mm2. 52 spectra for each condition were acquired in the corona radiata and 64 in the thalami. Phase and frequency drift corrections were performed using the residual water peak. Eddy current corrections were performed using the phase information from the non-water-suppressed data. Quantification of spectral data was performed with LCModel⁵. Brain tissues were classified in WM, GM and cerebrospinal fluid, and normalized thalamic volume (NTV) was calculated. The apparent diffusion coefficient (ADC) of tNAA and tCr was obtained evaluating the signal decay induced by diffusion weighting. Metabolite concentrations were also calculated, using the non-suppressed water reference peak and accounting for metabolite and water T2 and T1 decays, as well as for voxel tissue composition (VTC). Linear regressions were employed to assess: (i) group differences in metabolite ADCs and concentrations, accounting for age, gender, VTC, and NTV for GM-derived metrics; (ii) correlations between DWS-derived measures and clinical scores.Examples of spectra measured with and without diffusion weighting in the WM and GM of a single patient with MS are reported in Fig.1 and 2. In the WM, tNAA concentration was found to be significantly reduced in patients compared with HC (p=0.001), while the decrease in ADC(tNAA) in patients did not reach statistical significance (Fig3a). In the GM, a significant reduction in both tNAA concentration and ADC(tNAA) was found in patients compared with HC (p=0.001 and p=0.006 respectively, Fig3b). While tCr concentration showed no group difference in the WM, ADC(tCr) was significantly decreased in patients (p=0.016, Fig3c). In the GM, both tCr concentration and ADC(tCr) were significantly reduced in patients compared with HC (p=0.009 and p=0.031 respectively, Fig3d). Patients with more severe disability, as measured by EDSS, showed lower ADC(tNAA) values in the thalami (p=0.01, beta-coeff=-0.54), while patients with slower walking speed, as measured by TWT, presented more restricted tNAA and tCr diffusivity values in the thalami (p=0.01, beta-coeff=-0.81 and p=0.04, beta-coeff=-0.56, respectively).A reduced tNAA and tCr diffusivity was found in the brain of patients with MS, which may reflect an ongoing neuro-axonal damage and a simultaneous energy dysregulation affecting neurons and/or glial cells in this disease. The reduction in ADC(tNAA) in the GM of patients with MS can be interpreted through a structural deformation of neurons⁶. Two hypotheses may be proposed to explain the decrease in tCr diffusivity in this disease: (i) an abnormal reduction of intercellular creatine transport; and/or (ii) an impaired function of the creatine-kinase B⁷, resulting in a pathological increase in the phospocreatine/creatine ratio. DWS allows to capture in-vivo the ongoing neuro-axonal pathology as well as the potentially reversible phase of energy dysregulation which could precede the onset of neuro-axonal degeneration. This technique may offer the opportunity to investigate the mismatch between energy demand and supply that could trigger neurodegeneration in conditions of increased energy demand in MS.