Regional normal-appearing white matter metabolic profiles reflect information processing speed and verbal learning in early RRMS: A multi-voxel ¹H-MRS cross-sectional study

Regional normal-appearing white matter metabolic profiles reflect information processing speed and verbal learning in early RRMS: A multi-voxel ¹H-MRS cross-sectional study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Regional normal-appearing white matter metabolic profiles reflect information processing speed and verbal learning in early RRMS: A multi-voxel ¹H-MRS cross-sectional study Lorand Sakalaš¹, Vesna Suknjaja¹·², Jasmina Boban³·² This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9394188/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Cognitive impairment affects 40–70% of multiple sclerosis (MS) patients. Regional domain-selective associations between normal-appearing white matter (NAWM) metabolites and cognition remain underexplored. Methods: In this cross-sectional study, 52 early RRMS patients (EDSS 0–3; all on interferon-beta) and 44 healthy controls underwent multi-voxel 3D ¹H-MRS (3T, PRESS) across six supracallosal NAWM regions. Between-group metabolite differences were FDR-corrected. Spearman correlations and age-adjusted partial correlations assessed MRS–cognition associations. ROC analysis with DeLong testing evaluated discriminative ability. Results: Significant NAA/Cr differences were found in four of six regions. Left frontal NAA/Cr (R1) correlated with SDMT (ρ = 0.350, p = 0.012) and discriminated below-average processing speed (AUC = 0.724; 95% CI: 0.553–0.880; DeLong vs. age: ΔAUC = 0.208, p = 0.036). Left parietal NAA/Cr (R5) showed a trend association with RAVLT z-score (ρ = 0.261, p = 0.064). Age-controlled partial correlations were directionally consistent but non-significant (both p > 0.10). Neither fatigue nor mood correlated with MRS parameters. Conclusion: Multi-voxel ¹H-MRS reveals topographically heterogeneous NAWM damage with domain-selective cognitive associations in early RRMS. NAA/Cr R1 shows preliminary discriminative ability formally superior to age alone, but requires confirmation in larger cohorts. multiple sclerosis magnetic resonance spectroscopy normal-appearing white matter cognitive impairment Figures Figure 1 Figure 2 1. Introduction Cognitive impairment occurs in 40–70% of patients with multiple sclerosis (MS) across disease stages and directly affects vocational status and quality of life [ 1 ]. The most commonly affected cognitive domains are information processing speed — measured by the Symbol Digit Modalities Test (SDMT) — and verbal learning, measured by the Rey Auditory Verbal Learning Test (RAVLT) [ 2 ]. Of particular clinical importance is the fact that cognitive decline may precede measurable neurological disability on the Expanded Disability Status Scale (EDSS) by years, highlighting the need for sensitive neuroradiological biomarkers in early disease [ 3 ]. Conventional MRI parameters only moderately correlate with cognitive deficits (clinico-radiological paradox) [ 4 ], partly because diffuse microstructural NAWM damage remains invisible on standard sequences. ¹H-MRS quantifies neurochemical processes in vivo: NAA/Cr reflects axonal integrity, Cho/Cr marks demyelination, and mI/Cr indicates glial activation [ 5 , 6 ]. Previous MRS studies have consistently shown NAA/Cr reduction in NAWM of MS patients [ 5 , 6 ], but most studies have analyzed global or single-region changes. The disconnection syndrome hypothesis posits that damage to specific white matter pathways underlies domain-selective cognitive deficits: frontal damage affects processing speed, while parietal damage affects verbal learning [ 7 , 8 ]. Accordingly, we designed a cross-sectional study of 52 patients with early RRMS (EDSS 0–3; all on interferon-beta) in which we determined metabolite ratios across six supracallosal NAWM regions and examined their domain-selective associations with SDMT, RAVLT z-score, and MSFC (Multiple Sclerosis Functional Composite). Additionally, we tested specificity by examining whether MRS parameters correlate with fatigue and depressive symptoms, common non-cognitive MS symptoms that could act as confounders. 2. Subjects and Methods 2.1 Subjects This prospective cross-sectional study enrolled 52 adult patients with RRMS diagnosed according to the 2017 McDonald criteria [9] and 44 healthy controls (HC) matched for age and sex. Patient inclusion criteria were: EDSS 0–3.0, no clinical relapse or corticosteroid therapy within the preceding 30 days, and treatment with interferon-beta (IFN-β-1a or IFN-β-1b). One enrolled patient could not complete the MRI examination and was excluded from neuroimaging analysis (final MRS sample: n = 51). HC underwent MRI/MRS but not cognitive testing, serving exclusively as a metabolite reference group. All participants provided written informed consent. The study was approved by the institutional Ethics Committee and conducted in accordance with the Declaration of Helsinki. 2.2 Cognitive and clinical assessment Neurological disability was assessed using the EDSS. Cognitive testing was performed exclusively in RRMS patients. Information processing speed was assessed using the oral version of the SDMT (90-second administration). Below-average processing speed performance was defined as a raw score ≤ 56, corresponding to the mean of Serbian healthy controls in the BICAMS validation study (HC mean 56.3; Drulović et al. [10]) and conceptually aligned with the functional benchmark of occupational difficulties according to Benedict et al. [11]. This threshold is lower than the standard BICAMS impairment criterion (z ≤ −1.5, raw score ≤ 40) and detects a broader spectrum of below-average cognitive performance. Verbal learning and memory were assessed using the RAVLT (five successive learning trials; total score A1–A5, range 0–75). RAVLT z-scores were calculated using Serbian normative data by Janićijević et al. [12], stratified by sex and age group (18–29, 30–39, 40–49, 50–59, ≥60 years); impairment was defined as z ≤ −1.5. Global cognition was screened using the MoCA. Fatigue was quantified using the Fatigue Impact Scale (FIS, range 0–160). Mood was assessed using the Depression Anxiety Stress Scales-21 (DASS-21). Functional disability was quantified using the MSFC composite z-score (Timed 25-Foot Walk [T25FW], 9-Hole Peg Test [9HPT], and PASAT). 2.3 MRI and ¹H-MRS protocol All examinations were performed on a 3T MRI scanner (Trio Tim, Siemens, Erlangen, Germany) at a single center. Conventional imaging (T1W, T2W, FLAIR, 3D T1W MPR, diffusion) confirmed the absence of focal lesions within spectroscopic voxels. Proton 3D multi-voxel MRS of supracallosal white matter was performed using PRESS (TR/TE: 1700/135 ms long echo, 1700/30 ms short echo; FOV 160 × 160 × 160 mm; VOI 80 × 80 × 80 mm; resulting voxel size 10 × 10 × 10 mm [1.0 cm³]; acquisition time 8 min 17 s). Six outer-volume saturation regions and automatic volume-selective shimming were applied, yielding a 64-voxel (8 × 8) grid. Six anatomical regions were defined: R1 = left frontal subcortical, R2 = right frontal subcortical, R3 = left deep white matter (corona radiata), R4 = right deep white matter, R5 = left parietal subcortical, R6 = right parietal subcortical. NAA (2.0 ppm), Cho (3.2 ppm), and tCr (3.0 ppm) were quantified on long echo time (TE = 135 ms); mI (3.5 ppm) on short echo time sequence (TE = 30 ms). All ratios were expressed relative to creatine, the internal reference standard. Metabolite ratios were determined using the Siemens Syngo software. Spectra with CRLB > 20% for NAA or Cr were excluded; mean CRLB for NAA was 6.2% (range 2.1–18.4%). Low signal-to-noise spectra were repositioned until obtaining spectra of satisfactory quality. Correction for cerebrospinal fluid partial volume within voxels was not applied, representing a potential source of systematic noise, particularly in patients with more pronounced atrophy (see Limitations). All the spectra were positioned and analyzed by an experienced neuroradiologist (over 15 years of experience). One extreme mI/Cr value for R4 (40.43; remaining range 0.34–0.93) was identified as a data entry error and excluded from analysis for that region (n = 50 for mI/Cr R4; noted in Table 2). During the preparation of this manuscript, the authors used Claude (Anthropic) for language translation (Serbian to English), language editing, statistical code verification, and manuscript formatting. The authors reviewed, verified, and edited all AI-assisted output against the original data and take full responsibility for the content of the publication. 2.4 Statistical analysis Distribution normality was tested using the Shapiro–Wilk test. Group comparisons were performed using Student's t-test for normally distributed variables or the Mann–Whitney U test for non-normal distributions and ordinal data. Between-group MRS differences (18 comparisons: 3 metabolites × 6 regions) were corrected using the FDR procedure according to Benjamini–Hochberg [13]; effect sizes were expressed as Cohen's d. Associations between MRS ratios and cognitive/clinical outcomes were assessed using Spearman rank correlation. For primary hypothesized pairs (NAA/Cr R1 vs. SDMT; NAA/Cr R5 vs. RAVLT z-score), partial correlations controlling for age were calculated. The discriminative ability of NAA/Cr R1 for below-average SDMT performance was assessed using ROC curve analysis; the optimal cut-off was determined by the Youden index, and 95% CIs for AUC were obtained by bootstrap resampling (2,000 iterations, seed = 2024). The incremental value of NAA/Cr R1 beyond age was tested by comparing logistic regression models (NAA/Cr R1 alone, age alone, NAA/Cr R1 + age) via AUC comparison; the statistical significance of AUC differences between correlated models was assessed using the DeLong test [14]. Differences between cognitive subgroups were assessed using the Kruskal–Wallis test with subsequent Bonferroni-corrected pairwise tests. Statistical significance was defined at a two-sided α = 0.05. All analyses were performed in Python 3.12 (SciPy 1.11). 3. Results 3.1 Subject characteristics Fifty-two RRMS patients and 44 healthy controls were enrolled. Groups were matched for age (RRMS 39.7 ± 9.8 vs. HC 40.3 ± 9.5 years; p = 0.736) and sex (both groups ~81% female; p = 1.000). RRMS patients had significantly more years of formal education (median 16 [IQR 12–17] vs. 12 [12–16] years; p = 0.006). All patients were in early RRMS: EDSS 0.0–3.0 (median 2.0), disease duration 5.0 [IQR 3.8–10.2] years, all on IFN-β therapy. Detailed characteristics are presented in Online Resource Table S1. On cognitive testing, 38/52 (73.1%) had SDMT ≤ 56, and 14/52 (26.9%) had RAVLT impairment (z ≤ −1.5); 12 (23.1%) had dual impairment, 12 (23.1%) had no impairment, and 26 (50.0%) had isolated below-average SDMT performance. MoCA (26.9 ± 2.3) did not correlate with MRS parameters (all p > 0.05). 3.2 Regional NAWM metabolite changes Multi-regional MRS revealed topographically inhomogeneous metabolic changes (Table 2; Fig. 1). After FDR correction, significant NAA/Cr differences were found in four of six regions (R1, R3, R5: decreased in RRMS; R4: paradoxically elevated — likely a methodological artifact, see Discussion). Cho/Cr was elevated in R1, R3, R5, and R6 (all p_FDR < 0.01); mI/Cr showed a regionally heterogeneous pattern. [Table 2 about here] 3.3 Correlations between NAWM metabolites and cognitive outcomes Spearman analysis revealed a domain-selective pattern (Table 3; Fig. 2): NAA/Cr R1 significantly correlated with SDMT (ρ = 0.350, p = 0.012; ρ² ≈ 12.2%), while NAA/Cr R5 showed a trend association with RAVLT z-score (ρ = 0.261, p = 0.064; ρ² ≈ 6.8%). Both associations were regionally specific. Partial correlations controlling for age showed attenuated, directionally consistent, but statistically non-significant values: SDMT–NAA/Cr R1 (partial r = 0.205, p = 0.148) and RAVLT z–NAA/Cr R5 (partial r = 0.151, p = 0.290). These findings suggest that age partly mediates both associations and that the correlations are preliminary in nature; definitive confirmation requires larger samples and multivariate models. Neither FIS nor DASS-21 correlated with NAA/Cr (all p > 0.50). EDSS correlated with SDMT (ρ = −0.343, p = 0.013), and depressive symptoms with RAVLT z-score (ρ = −0.396, p = 0.004). [Table 3 about here] 3.4 MSFC and frontal NAWM integrity MSFC composite scores significantly correlated with NAA/Cr in three regions: R1 left frontal (ρ = 0.325, p = 0.020), R2 right frontal (ρ = 0.375, p = 0.007), and R6 right parietal (ρ = 0.325, p = 0.020). Bilateral frontal associations were the strongest MRS–clinical correlations in the dataset. We note that these correlations were not corrected for multiple testing beyond the primary hypotheses and should be considered exploratory findings requiring confirmation. 3.5 Discriminative ability and cognitive subgroup analysis NAA/Cr R1 discriminated below-average from intact patients: AUC = 0.724 (95% CI: 0.553–0.880; Fig. 2c), optimal threshold NAA/Cr ≤ 1.890 (sensitivity 64.9%, specificity 78.6%). Logistic regression with age as a covariate showed that age alone has negligible discriminative ability (AUC = 0.515); the DeLong test confirmed that NAA/Cr R1 had significantly greater discriminative ability than age alone (ΔAUC = 0.208; z = 2.094; p = 0.036). Adding age to the NAA/Cr R1 model did not improve discrimination (AUC = 0.730 vs. 0.724; DeLong p = 0.706). Supplementary analysis with the BICAMS threshold (SDMT ≤ 40): AUC = 0.635 (95% CI: 0.445–0.820; 11 impaired vs. 40 intact). The Kruskal–Wallis test confirmed a NAA/Cr R1 gradient across cognitive profiles (H = 7.79, p = 0.020; n = 49): no impairment 1.989 ± 0.127; SDMT-only 1.818 ± 0.242; dual impairment 1.777 ± 0.172 (Fig. 2d; post-hoc: no impairment vs. dual p_adj = 0.013). 4. Discussion 4.1 Summary of key findings This study demonstrates three findings: (1) regionally heterogeneous NAWM metabolic changes (FDR-significant in 4/6 regions); (2) domain-selective associations — frontal NAA/Cr correlates with SDMT, parietal with RAVLT z-score (trend); (3) absence of correlation with fatigue and depression. Partial correlations lose significance when controlling for age, and explained variance remains modest (ρ² ≈ 12% for SDMT), positioning the findings as preliminary. 4.2 Methodological advantages of multi-regional MRS over single-voxel studies Most previous MRS studies of cognition in MS used single-voxel spectroscopy (SVS), which assumes NAWM damage uniformity — our data refute this (Cohen's d ranged from 0.209 to 2.195), consistent with Sun et al. [15]. Additionally, large SVS voxels (8–25 cm³) risk lesional contamination, whereas our 1 cm³ voxels allow individual T2/FLAIR screening; and SVS cannot test spatial hypotheses [8, 16]. Our 3D PRESS protocol at 3T [17] overcomes these limitations with clinically acceptable acquisition time. 4.3 Regional specificity and the disconnection syndrome hypothesis Our multi-regional design reveals a directionally dissociative pattern: frontal NAA/Cr (R1) correlates with SDMT (ρ = 0.350) but not with RAVLT z-score; parietal NAA/Cr (R5) shows a trend association with RAVLT z-score (ρ = 0.261, p = 0.064) but not with SDMT. This dissociation is neuroanatomically coherent — SDMT depends on fronto-striatal circuits [18], while RAVLT involves parietal-hippocampal networks [19]. The stepwise NAA/Cr R1 gradient across cognitive profiles (H = 7.79, p = 0.020) suggests a dose-dependent relationship between axonal integrity and cognitive status, consistent with the disconnection syndrome hypothesis [8]. 4.4 Inferential limitations: partial correlations and effect size Both primary correlations lose significance when controlling for age: SDMT–NAA/Cr R1 (partial r = 0.205, p = 0.148) and RAVLT z–NAA/Cr R5 (r = 0.151, p = 0.290). NAA/Cr may reflect general aging rather than MS-specific pathology; alternatively, aging and MS neurodegeneration share common mechanisms (oxidative stress, mitochondrial dysfunction), and partial correlation may overcorrect by removing genuinely MS-relevant variance. Without longitudinal data, these interpretations cannot be distinguished. Explained variance is modest (ρ² ≈ 12% for SDMT, ~6.8% for RAVLT z), positioning NAWM metabolic changes as one of several pathophysiological mechanisms of cognitive impairment in MS; comprehensive prediction requires multivariate models combining MRS, DTI, lesion burden, and volumetry. 4.5 Asymmetric distribution and the R4 artifact Relative preservation of right frontal (R2) and right parietal (R6) NAWM may reflect asymmetric demyelination distribution known in MS; the R2–MSFC correlation (uncorrected for multiple testing) requires confirmation. The finding of higher NAA/Cr in RRMS at R4 (d = 1.602) likely reflects a methodological artifact due to corona radiata heterogeneity: hemispheric white matter asymmetry, differing shimming quality near the paranasal sinuses, or corpus callosum signal contamination. Without DTI co-registration, precise identification of the cause remains speculative. 4.6 Fatigue, depression, and neurobiological specificity The absence of correlation between NAA/Cr and FIS or DASS-21 (all p > 0.50) provides evidence of specificity of MRS markers for cognitive outcomes, consistent with the literature on different mechanisms of fatigue in MS [20, 21]. Depressive symptoms correlate with RAVLT z-score (ρ = −0.396, p = 0.004), but the partial correlation of NAA/Cr R5–RAVLT controlling for age and depression (r = 0.162, p = 0.257) is numerically higher than without controlling for depression (r = 0.151), suggesting that depression does not act as a confounder of this association [22]. 4.7 Clinical implications and perspectives The discriminative ability of NAA/Cr R1 (AUC = 0.724) is formally superior to age alone (AUC = 0.515; DeLong ΔAUC = 0.208, p = 0.036), while adding age to the NAA/Cr R1 model does not improve discrimination (DeLong p = 0.706). This indicates that discriminative power derives from the metabolic marker rather than from an age effect. Replication in larger cohorts (n ≥ 150) with multivariate correction remains necessary. The fact that ~73% of patients on IFN-β still show below-average SDMT performance is consistent with the smouldering MS concept [23]; the difference from the prevalence in Drulović et al. [10] (52.7%) reflects the difference in applied thresholds (functional ≤ 56 vs. BICAMS z ≤ −1.5). Cognitive reserve may mitigate the rate of decline even in the presence of existing metabolic changes [24]. 4.8 Study limitations Limitations include: (1) cross-sectional design; (2) small sample size (power ≈ 48% for ρ ≈ 0.27; ≈ 72% for ρ = 0.35); (3) healthy controls without cognitive testing; (4) residual confounding by education; (5) RAVLT norms [12] based on a preprint study (n = 200); (6) MSFC correlations without multiple testing correction; (7) exclusive use of IFN-β limits generalizability; (8) no CSF partial volume correction — the creatine ratio partially compensates but incompletely, as CSF differentially affects metabolites; future studies should include absolute quantification. 5. Conclusion Multi-voxel ¹H-MRS reveals topographically selective NAWM metabolic changes in early RRMS with domain-consistent associations: frontal NAA/Cr correlates with processing speed, while parietal NAA/Cr shows a trend association with verbal learning. Specificity for cognitive — rather than affective — outcomes supports the biological basis of the findings. Partial correlations lose significance when controlling for age, indicating synergistic aging and neurodegenerative processes whose contributions cannot be disentangled in the present sample. NAA/Cr R1 shows preliminary discriminative ability (AUC 0.724) formally superior to age alone (DeLong p = 0.036), but definitive confirmation requires validation in larger cohorts with multivariate control for age, lesion burden, and atrophy. Statements and Declarations Competing Interests: The authors declare that they have no competing interests. Ethics approval: This study was approved by the institutional Ethics Committee. Informed consent was obtained from all individual participants included in the study. Data availability: De-identified data are available from the corresponding author upon reasonable request. Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References Chiaravalloti ND, DeLuca J (2008) Cognitive impairment in multiple sclerosis. Lancet Neurol 7:1139–1151. https://doi.org/10.1016/S1474-4422(08)70259-X Sumowski JF, Benedict R, Enzinger C et al (2018) Cognition in multiple sclerosis: state of the field and priorities for the future. Neurology 90:278–288. https://doi.org/10.1212/WNL.0000000000004977 Kappos L, Wolinsky JS, Giovannoni G et al (2020) Contribution of relapse-independent progression vs relapse-associated worsening to overall confirmed disability accumulation in typical relapsing multiple sclerosis. JAMA Neurol 77:1132–1140. https://doi.org/10.1001/jamaneurol.2020.1568 Barkhof F (2002) The clinico-radiological paradox in multiple sclerosis revisited. Curr Opin Neurol 15:239–245. https://doi.org/10.1097/00019052-200206000-00003 De Stefano N, Matthews PM, Fu L et al (1998) Axonal damage correlates with disability in patients with relapsing-remitting multiple sclerosis. Brain 121:1469–1477. https://doi.org/10.1093/brain/121.8.1469 De Stefano N, Narayanan S, Francis GS et al (2001) Evidence of axonal damage in the early stages of multiple sclerosis and its relevance to disability. Arch Neurol 58:65–70. https://doi.org/10.1001/archneur.58.1.65 Schoonheim MM, Meijer KA, Geurts JJG (2015) Network collapse and cognitive impairment in multiple sclerosis. Front Neurol 6:82. https://doi.org/10.3389/fneur.2015.00082 Dineen RA, Vilisaar J, Hlinka J et al (2009) Disconnection as a mechanism for cognitive dysfunction in multiple sclerosis. Brain 132:239–249. https://doi.org/10.1093/brain/awn275 Thompson AJ, Banwell BL, Barkhof F et al (2018) Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol 17:162–173. https://doi.org/10.1016/S1474-4422(17)30470-2 Drulović J, Tončev G, Nadj Č et al (2022) Validation of the Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS) in a large cohort of relapsing-remitting MS patients. Acta Clin Croat 61:62–68. https://doi.org/10.20471/acc.2022.61.01.08 Benedict RH, Drake AS, Irwin LN et al (2016) Benchmarks of meaningful impairment on the MSFC and BICAMS. Mult Scler 22:1874–1882. https://doi.org/10.1177/1352458516633517 Janićijević N, Purić D (2025) Reliability and convergent validity of the Rey-Auditory Verbal Learning Test (RAVLT) against working memory tasks in a Serbian community sample. Preprint. https://osf.io/4ywpr/ Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 57:289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x DeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44:837–845. https://doi.org/10.2307/2531595 Sun J, Song H, Yang Y et al (2017) Metabolic changes in normal appearing white matter in multiple sclerosis patients using multivoxel magnetic resonance spectroscopy imaging. Medicine 96:e6534. https://doi.org/10.1097/MD.0000000000006534 Heckova E, Dal-Bianco A, Strasser B et al (2022) Extensive brain pathologic alterations detected with 7.0-T MR spectroscopic imaging associated with disability in multiple sclerosis. Radiology 303:199–209. https://doi.org/10.1148/radiol.210614 Inglese M, Li BS, Rusinek H et al (2003) Diffusely elevated cerebral choline and creatine in relapsing-remitting multiple sclerosis. Magn Reson Med 50:190–195. https://doi.org/10.1002/mrm.10481 Riccitelli GC, Pagani E, Rodegher M et al (2019) Imaging patterns of gray and white matter abnormalities associated with PASAT and SDMT performance in relapsing-remitting multiple sclerosis. Mult Scler 25:204–216. https://doi.org/10.1177/1352458517743091 Wagner AD, Shannon BJ, Kahn I, Buckner RL (2005) Parietal lobe contributions to episodic memory retrieval. Trends Cogn Sci 9:445–453. https://doi.org/10.1016/j.tics.2005.07.001 Induruwa I, Constantinescu CS, Gran B (2012) Fatigue in multiple sclerosis — a brief review. J Neurol Sci 323:9–15. https://doi.org/10.1016/j.jns.2012.08.007 Tartaglia MC, Narayanan S, Francis SJ et al (2004) The relationship between diffuse axonal damage and fatigue in multiple sclerosis. Arch Neurol 61:201–207. https://doi.org/10.1001/archneur.61.2.201 Arnett PA, Barwick FH, Beeney JE (2008) Depression in multiple sclerosis: review and theoretical proposal. J Int Neuropsychol Soc 14:691–724. https://doi.org/10.1017/S1355617708081174 Giovannoni G, Popescu V, Wuerfel J et al (2022) Smouldering multiple sclerosis: the real MS. Ther Adv Neurol Disord 15:17562864211066751. https://doi.org/10.1177/17562864211066751 Sumowski JF, Rocca MA, Leavitt VM et al (2014) Brain reserve and cognitive reserve protect against cognitive decline over 4.5 years in multiple sclerosis. Neurology 82:1776–1783. https://doi.org/10.1212/WNL.0000000000000433 Tables Table 2 NAWM MRS ratios in RRMS vs. healthy controls (FDR-corrected) NAWM region RRMS (n=51) Mean ± SD HC (n=44) Mean ± SD p (FDR) Cohen's d Direction NAA/Cr (N-acetylaspartate/creatine) R1: Frontal subcortical (L) 1.843 ± 0.247 2.016 ± 0.258 0.009 0.709 ↓ MS R2: Frontal subcortical (R) 1.832 ± 0.239 1.770 ± 0.250 0.106 0.251 n.s. R3: Deep white matter (L) 1.969 ± 0.278 2.686 ± 0.372 <0.001 2.195 ↓ MS R4: Deep white matter (R) 2.004 ± 0.275 1.597 ± 0.226 <0.001 1.602 ↑ MS* R5: Parietal subcortical (L) 2.077 ± 0.404 2.230 ± 0.315 0.002 0.417 ↓ MS R6: Parietal subcortical (R) 2.085 ± 0.443 2.004 ± 0.313 0.581 0.209 n.s. Cho/Cr (choline/creatine) R1: Frontal subcortical (L) 1.187 ± 0.184 0.939 ± 0.173 <0.001 1.381 ↑ MS R2: Frontal subcortical (R) 1.174 ± 0.180 1.153 ± 0.142 0.562 0.129 n.s. R3: Deep white matter (L) 1.189 ± 0.173 0.792 ± 0.179 <0.001 2.245 ↑ MS R4: Deep white matter (R) 1.180 ± 0.186 1.113 ± 0.124 0.063 0.417 n.s. R5: Parietal subcortical (L) 1.127 ± 0.203 0.984 ± 0.169 0.001 0.758 ↑ MS R6: Parietal subcortical (R) 1.154 ± 0.209 0.892 ± 0.093 <0.001 1.583 ↑ MS mI/Cr (myo-inositol/creatine) R1: Frontal subcortical (L) 0.527 ± 0.130 0.436 ± 0.135 0.002 0.672 ↑ MS R2: Frontal subcortical (R) 0.548 ± 0.136 0.575 ± 0.139 0.157 0.198 n.s. R3: Deep white matter (L) 0.515 ± 0.148 0.405 ± 0.150 <0.001 0.731 ↑ MS R4: Deep white matter (R)‡ 0.537 ± 0.131 0.542 ± 0.146 0.462 0.035 n.s. R5: Parietal subcortical (L) 0.497 ± 0.119 0.612 ± 0.171 0.001 0.782 ↓ MS R6: Parietal subcortical (R) 0.520 ± 0.189 0.402 ± 0.110 <0.001 0.744 ↑ MS Data: mean ± SD. *Value higher in MS; likely methodological artifact (see text). ‡One extreme value excluded (n = 50 for mI/Cr R4; see Methods). Bold p-values: FDR-corrected p < 0.05. n.s. not significant Table 3 Spearman correlation between NAWM NAA/Cr and clinico-cognitive outcomes Outcome MRS ratio NAWM region Spearman ρ p-value Part. r (adj. age) SDMT (processing speed) NAA/Cr R1: Front. subcort. (L) +0.350 0.012* 0.205 (p=0.148, n.s.) R2–R6 all regions n.s. — — RAVLT z-score (verbal learning) NAA/Cr R5: Pariet. subcort. (L) +0.261 0.064 0.151 (p=0.290, n.s.) R1: Front. subcort. (L) +0.187 0.190 — MSFC composite NAA/Cr R1: Front. subcort. (L) +0.325 0.020* — R2: Front. subcort. (R) +0.375 0.007** — R6: Pariet. subcort. (R) +0.325 0.020* — FIS (fatigue) All All 6 regions n.s. — — DASS-21 (mood) All All 6 regions n.s. — — ρ Spearman coefficient, ρ² explained variance. Partial r adjusted for age (Pearson, residual method). *p < 0.05; **p < 0.01. n.s. not significant. RAVLT z-scores calculated using Serbian norms (Janićijević et al. [12]), stratified by sex and age group. FIS Fatigue Impact Scale, DASS-21 Depression Anxiety Stress Scales-21, MSFC Multiple Sclerosis Functional Composite Additional Declarations No competing interests reported. Supplementary Files ANBOnlineResourceTableS1.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9394188","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":630880936,"identity":"2aa92dff-4f19-4517-a24c-0ffb60cfa1e4","order_by":0,"name":"Lorand Sakalaš¹","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIie3OMQrCMBiG4b8E0iXaNVDUK0QKLhZ7lYpQVyfnipBJcS14Cb1BJGCX2q5CF90dFBfBDlZx6BR1c8gLCWR4yAeg0/1h9deNy0PAEAAu+0hwhUBJgt+J/IKY8+N1NAbPmsmDuBWZY0U+vtxUhMSOHaXQj3acbeY879C9j5ZERWgAdo2DDxkwUQtzlxGBkHIYDdC9JF4rM8+bokhfxFAOowF+/mKsdjMmCRYdZoYIlMPIFndJSvvrJBnJBh84NJ5MkYpYJkc5GbteMxmuj6ei115MkVQOe0erDyP8DHQ6nU6n7AH1dUUIoFeiLQAAAABJRU5ErkJggg==","orcid":"","institution":"Clinical Center of Vojvodina","correspondingAuthor":true,"prefix":"","firstName":"Lorand","middleName":"","lastName":"Sakalaš¹","suffix":""},{"id":630880937,"identity":"12f1a16e-ebd1-403b-8486-f722850d1b32","order_by":1,"name":"Vesna Suknjaja¹·²","email":"","orcid":"","institution":"University of Novi Sad","correspondingAuthor":false,"prefix":"","firstName":"Vesna","middleName":"","lastName":"Suknjaja¹·²","suffix":""},{"id":630880938,"identity":"8e8b42d3-4d5c-4716-9723-cb483b60a1cd","order_by":2,"name":"Jasmina Boban³·²","email":"","orcid":"","institution":"University of Novi Sad","correspondingAuthor":false,"prefix":"","firstName":"Jasmina","middleName":"","lastName":"Boban³·²","suffix":""}],"badges":[],"createdAt":"2026-04-12 12:38:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9394188/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9394188/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108097838,"identity":"d0f26781-a1bd-400d-bd84-26d737dbc78d","added_by":"auto","created_at":"2026-04-29 10:17:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":4555833,"visible":true,"origin":"","legend":"\u003cp\u003eRegional NAWM metabolite differences between RRMS and healthy controls. Bar plots (mean ± SEM) of NAA/Cr, Cho/Cr, and mI/Cr across six NAWM regions. Asterisks indicate FDR-corrected significance (p \u0026lt; 0.05). HC healthy controls, R1–R6 as defined in Methods\u003c/p\u003e","description":"","filename":"ANBFig1.png","url":"https://assets-eu.researchsquare.com/files/rs-9394188/v1/eda7cc370bdc63198e04ac3c.png"},{"id":108097840,"identity":"511009b3-73ac-455e-b77c-d5943ed69294","added_by":"auto","created_at":"2026-04-29 10:17:11","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":8115849,"visible":true,"origin":"","legend":"\u003cp\u003eDomain-selective MRS–cognition correlations and discriminative ability. a SDMT vs. NAA/Cr R1 (ρ = 0.350, p = 0.012). b RAVLT z-score vs. NAA/Cr R5 (ρ = 0.261, p = 0.064). c ROC curve for NAA/Cr R1 as a discriminator of below-average SDMT (AUC = 0.724; 95% CI 0.553–0.880). d NAA/Cr R1 by cognitive profile (KW p = 0.020; post-hoc: no impairment vs. dual p_adj = 0.013)\u003c/p\u003e","description":"","filename":"ANBFig2.png","url":"https://assets-eu.researchsquare.com/files/rs-9394188/v1/50d9bab294d4f43623cba481.png"},{"id":108181828,"identity":"d5639b36-365a-422a-a068-2d7b707f637b","added_by":"auto","created_at":"2026-04-30 08:58:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":7114407,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9394188/v1/6e10056c-936c-471a-bfc0-749060c7fa32.pdf"},{"id":108097837,"identity":"5fe06178-3566-4889-9d6b-cfdc30db53fd","added_by":"auto","created_at":"2026-04-29 10:17:11","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":9741,"visible":true,"origin":"","legend":"","description":"","filename":"ANBOnlineResourceTableS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9394188/v1/b3e7c1c20c70ce6a98d11ba4.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Regional normal-appearing white matter metabolic profiles reflect information processing speed and verbal learning in early RRMS: A multi-voxel ¹H-MRS cross-sectional study","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eCognitive impairment occurs in 40\u0026ndash;70% of patients with multiple sclerosis (MS) across disease stages and directly affects vocational status and quality of life [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. The most commonly affected cognitive domains are information processing speed \u0026mdash; measured by the Symbol Digit Modalities Test (SDMT) \u0026mdash; and verbal learning, measured by the Rey Auditory Verbal Learning Test (RAVLT) [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Of particular clinical importance is the fact that cognitive decline may precede measurable neurological disability on the Expanded Disability Status Scale (EDSS) by years, highlighting the need for sensitive neuroradiological biomarkers in early disease [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eConventional MRI parameters only moderately correlate with cognitive deficits (clinico-radiological paradox) [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], partly because diffuse microstructural NAWM damage remains invisible on standard sequences. \u0026sup1;H-MRS quantifies neurochemical processes in vivo: NAA/Cr reflects axonal integrity, Cho/Cr marks demyelination, and mI/Cr indicates glial activation [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003ePrevious MRS studies have consistently shown NAA/Cr reduction in NAWM of MS patients [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], but most studies have analyzed global or single-region changes. The disconnection syndrome hypothesis posits that damage to specific white matter pathways underlies domain-selective cognitive deficits: frontal damage affects processing speed, while parietal damage affects verbal learning [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e Accordingly, we designed a cross-sectional study of 52 patients with early RRMS (EDSS 0\u0026ndash;3; all on interferon-beta) in which we determined metabolite ratios across six supracallosal NAWM regions and examined their domain-selective associations with SDMT, RAVLT z-score, and MSFC (Multiple Sclerosis Functional Composite). Additionally, we tested specificity by examining whether MRS parameters correlate with fatigue and depressive symptoms, common non-cognitive MS symptoms that could act as confounders.\u003c/p\u003e"},{"header":"2. Subjects and Methods","content":"\u003cp\u003e2.1 Subjects\u003c/p\u003e\n\u003cp\u003eThis prospective cross-sectional study enrolled 52 adult patients with RRMS diagnosed according to the 2017 McDonald criteria [9] and 44 healthy controls (HC) matched for age and sex. Patient inclusion criteria were: EDSS 0–3.0, no clinical relapse or corticosteroid therapy within the preceding 30 days, and treatment with interferon-beta (IFN-β-1a or IFN-β-1b). One enrolled patient could not complete the MRI examination and was excluded from neuroimaging analysis (final MRS sample: n = 51). HC underwent MRI/MRS but not cognitive testing, serving exclusively as a metabolite reference group. All participants provided written informed consent. The study was approved by the institutional Ethics Committee and conducted in accordance with the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e2.2 Cognitive and clinical assessment\u003c/p\u003e\n\u003cp\u003eNeurological disability was assessed using the EDSS. Cognitive testing was performed exclusively in RRMS patients. Information processing speed was assessed using the oral version of the SDMT (90-second administration). Below-average processing speed performance was defined as a raw score ≤ 56, corresponding to the mean of Serbian healthy controls in the BICAMS validation study (HC mean 56.3; Drulović et al. [10]) and conceptually aligned with the functional benchmark of occupational difficulties according to Benedict et al. [11]. This threshold is lower than the standard BICAMS impairment criterion (z ≤ −1.5, raw score ≤ 40) and detects a broader spectrum of below-average cognitive performance. Verbal learning and memory were assessed using the RAVLT (five successive learning trials; total score A1–A5, range 0–75). RAVLT z-scores were calculated using Serbian normative data by Janićijević et al. [12], stratified by sex and age group (18–29, 30–39, 40–49, 50–59, ≥60 years); impairment was defined as z ≤ −1.5. Global cognition was screened using the MoCA. Fatigue was quantified using the Fatigue Impact Scale (FIS, range 0–160). Mood was assessed using the Depression Anxiety Stress Scales-21 (DASS-21). Functional disability was quantified using the MSFC composite z-score (Timed 25-Foot Walk [T25FW], 9-Hole Peg Test [9HPT], and PASAT).\u003c/p\u003e\n\u003cp\u003e2.3 MRI and ¹H-MRS protocol\u003c/p\u003e\n\u003cp\u003eAll examinations were performed on a 3T MRI scanner (Trio Tim, Siemens, Erlangen, Germany) at a single center. Conventional imaging (T1W, T2W, FLAIR, 3D T1W MPR, diffusion) confirmed the absence of focal lesions within spectroscopic voxels. Proton 3D multi-voxel MRS of supracallosal white matter was performed using PRESS (TR/TE: 1700/135 ms long echo, 1700/30 ms short echo; FOV 160 × 160 × 160 mm; VOI 80 × 80 × 80 mm; resulting voxel size 10 × 10 × 10 mm [1.0 cm³]; acquisition time 8 min 17 s). Six outer-volume saturation regions and automatic volume-selective shimming were applied, yielding a 64-voxel (8 × 8) grid. Six anatomical regions were defined: R1 = left frontal subcortical, R2 = right frontal subcortical, R3 = left deep white matter (corona radiata), R4 = right deep white matter, R5 = left parietal subcortical, R6 = right parietal subcortical. NAA (2.0 ppm), Cho (3.2 ppm), and tCr (3.0 ppm) were quantified on long echo time (TE = 135 ms); mI (3.5 ppm) on short echo time sequence (TE = 30 ms). All ratios were expressed relative to creatine, the internal reference standard. Metabolite ratios were determined using the Siemens Syngo software. Spectra with CRLB \u0026gt; 20% for NAA or Cr were excluded; mean CRLB for NAA was 6.2% (range 2.1–18.4%). Low signal-to-noise spectra were repositioned until obtaining spectra of satisfactory quality. Correction for cerebrospinal fluid partial volume within voxels was not applied, representing a potential source of systematic noise, particularly in patients with more pronounced atrophy (see Limitations). All the spectra were positioned and analyzed by an experienced neuroradiologist (over 15 years of experience). One extreme mI/Cr value for R4 (40.43; remaining range 0.34–0.93) was identified as a data entry error and excluded from analysis for that region (n = 50 for mI/Cr R4; noted in Table 2).\u003c/p\u003e\n\u003cp\u003eDuring the preparation of this manuscript, the authors used Claude (Anthropic) for language translation (Serbian to English), language editing, statistical code verification, and manuscript formatting. The authors reviewed, verified, and edited all AI-assisted output against the original data and take full responsibility for the content of the publication.\u003c/p\u003e\n\u003cp\u003e2.4 Statistical analysis\u003c/p\u003e\n\u003cp\u003eDistribution normality was tested using the Shapiro–Wilk test. Group comparisons were performed using Student's t-test for normally distributed variables or the Mann–Whitney U test for non-normal distributions and ordinal data. Between-group MRS differences (18 comparisons: 3 metabolites × 6 regions) were corrected using the FDR procedure according to Benjamini–Hochberg [13]; effect sizes were expressed as Cohen's d. Associations between MRS ratios and cognitive/clinical outcomes were assessed using Spearman rank correlation. For primary hypothesized pairs (NAA/Cr R1 vs. SDMT; NAA/Cr R5 vs. RAVLT z-score), partial correlations controlling for age were calculated. The discriminative ability of NAA/Cr R1 for below-average SDMT performance was assessed using ROC curve analysis; the optimal cut-off was determined by the Youden index, and 95% CIs for AUC were obtained by bootstrap resampling (2,000 iterations, seed = 2024). The incremental value of NAA/Cr R1 beyond age was tested by comparing logistic regression models (NAA/Cr R1 alone, age alone, NAA/Cr R1 + age) via AUC comparison; the statistical significance of AUC differences between correlated models was assessed using the DeLong test [14]. Differences between cognitive subgroups were assessed using the Kruskal–Wallis test with subsequent Bonferroni-corrected pairwise tests. Statistical significance was defined at a two-sided α = 0.05. All analyses were performed in Python 3.12 (SciPy 1.11).\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003e3.1 Subject characteristics\u003c/p\u003e\n\u003cp\u003eFifty-two RRMS patients and 44 healthy controls were enrolled. Groups were matched for age (RRMS 39.7 ± 9.8 vs. HC 40.3 ± 9.5 years; p = 0.736) and sex (both groups ~81% female; p = 1.000). RRMS patients had significantly more years of formal education (median 16 [IQR 12–17] vs. 12 [12–16] years; p = 0.006). All patients were in early RRMS: EDSS 0.0–3.0 (median 2.0), disease duration 5.0 [IQR 3.8–10.2] years, all on IFN-β therapy. Detailed characteristics are presented in Online Resource Table S1.\u003c/p\u003e\n\u003cp\u003eOn cognitive testing, 38/52 (73.1%) had SDMT ≤ 56, and 14/52 (26.9%) had RAVLT impairment (z ≤ −1.5); 12 (23.1%) had dual impairment, 12 (23.1%) had no impairment, and 26 (50.0%) had isolated below-average SDMT performance. MoCA (26.9 ± 2.3) did not correlate with MRS parameters (all p \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003e3.2 Regional NAWM metabolite changes\u003c/p\u003e\n\u003cp\u003eMulti-regional MRS revealed topographically inhomogeneous metabolic changes (Table 2; Fig. 1). After FDR correction, significant NAA/Cr differences were found in four of six regions (R1, R3, R5: decreased in RRMS; R4: paradoxically elevated — likely a methodological artifact, see Discussion). Cho/Cr was elevated in R1, R3, R5, and R6 (all p_FDR \u0026lt; 0.01); mI/Cr showed a regionally heterogeneous pattern.\u003c/p\u003e\n\u003cp\u003e[Table 2 about here]\u003c/p\u003e\n\u003cp\u003e3.3 Correlations between NAWM metabolites and cognitive outcomes\u003c/p\u003e\n\u003cp\u003eSpearman analysis revealed a domain-selective pattern (Table 3; Fig. 2): NAA/Cr R1 significantly correlated with SDMT (ρ = 0.350, p = 0.012; ρ² ≈ 12.2%), while NAA/Cr R5 showed a trend association with RAVLT z-score (ρ = 0.261, p = 0.064; ρ² ≈ 6.8%). Both associations were regionally specific.\u003c/p\u003e\n\u003cp\u003ePartial correlations controlling for age showed attenuated, directionally consistent, but statistically non-significant values: SDMT–NAA/Cr R1 (partial r = 0.205, p = 0.148) and RAVLT z–NAA/Cr R5 (partial r = 0.151, p = 0.290). These findings suggest that age partly mediates both associations and that the correlations are preliminary in nature; definitive confirmation requires larger samples and multivariate models.\u003c/p\u003e\n\u003cp\u003eNeither FIS nor DASS-21 correlated with NAA/Cr (all p \u0026gt; 0.50). EDSS correlated with SDMT (ρ = −0.343, p = 0.013), and depressive symptoms with RAVLT z-score (ρ = −0.396, p = 0.004).\u003c/p\u003e\n\u003cp\u003e[Table 3 about here]\u003c/p\u003e\n\u003cp\u003e3.4 MSFC and frontal NAWM integrity\u003c/p\u003e\n\u003cp\u003eMSFC composite scores significantly correlated with NAA/Cr in three regions: R1 left frontal (ρ = 0.325, p = 0.020), R2 right frontal (ρ = 0.375, p = 0.007), and R6 right parietal (ρ = 0.325, p = 0.020). Bilateral frontal associations were the strongest MRS–clinical correlations in the dataset. We note that these correlations were not corrected for multiple testing beyond the primary hypotheses and should be considered exploratory findings requiring confirmation.\u003c/p\u003e\n\u003cp\u003e3.5 Discriminative ability and cognitive subgroup analysis\u003c/p\u003e\n\u003cp\u003eNAA/Cr R1 discriminated below-average from intact patients: AUC = 0.724 (95% CI: 0.553–0.880; Fig. 2c), optimal threshold NAA/Cr ≤ 1.890 (sensitivity 64.9%, specificity 78.6%). Logistic regression with age as a covariate showed that age alone has negligible discriminative ability (AUC = 0.515); the DeLong test confirmed that NAA/Cr R1 had significantly greater discriminative ability than age alone (ΔAUC = 0.208; z = 2.094; p = 0.036). Adding age to the NAA/Cr R1 model did not improve discrimination (AUC = 0.730 vs. 0.724; DeLong p = 0.706). Supplementary analysis with the BICAMS threshold (SDMT ≤ 40): AUC = 0.635 (95% CI: 0.445–0.820; 11 impaired vs. 40 intact).\u003c/p\u003e\n\u003cp\u003eThe Kruskal–Wallis test confirmed a NAA/Cr R1 gradient across cognitive profiles (H = 7.79, p = 0.020; n = 49): no impairment 1.989 ± 0.127; SDMT-only 1.818 ± 0.242; dual impairment 1.777 ± 0.172 (Fig. 2d; post-hoc: no impairment vs. dual p_adj = 0.013).\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003e4.1 Summary of key findings\u003c/p\u003e\n\u003cp\u003eThis study demonstrates three findings: (1) regionally heterogeneous NAWM metabolic changes (FDR-significant in 4/6 regions); (2) domain-selective associations — frontal NAA/Cr correlates with SDMT, parietal with RAVLT z-score (trend); (3) absence of correlation with fatigue and depression. Partial correlations lose significance when controlling for age, and explained variance remains modest (ρ² ≈ 12% for SDMT), positioning the findings as preliminary.\u003c/p\u003e\n\u003cp\u003e4.2 Methodological advantages of multi-regional MRS over single-voxel studies\u003c/p\u003e\n\u003cp\u003eMost previous MRS studies of cognition in MS used single-voxel spectroscopy (SVS), which assumes NAWM damage uniformity — our data refute this (Cohen's d ranged from 0.209 to 2.195), consistent with Sun et al. [15]. Additionally, large SVS voxels (8–25 cm³) risk lesional contamination, whereas our 1 cm³ voxels allow individual T2/FLAIR screening; and SVS cannot test spatial hypotheses [8, 16]. Our 3D PRESS protocol at 3T [17] overcomes these limitations with clinically acceptable acquisition time.\u003c/p\u003e\n\u003cp\u003e4.3 Regional specificity and the disconnection syndrome hypothesis\u003c/p\u003e\n\u003cp\u003eOur multi-regional design reveals a directionally dissociative pattern: frontal NAA/Cr (R1) correlates with SDMT (ρ = 0.350) but not with RAVLT z-score; parietal NAA/Cr (R5) shows a trend association with RAVLT z-score (ρ = 0.261, p = 0.064) but not with SDMT. This dissociation is neuroanatomically coherent — SDMT depends on fronto-striatal circuits [18], while RAVLT involves parietal-hippocampal networks [19]. The stepwise NAA/Cr R1 gradient across cognitive profiles (H = 7.79, p = 0.020) suggests a dose-dependent relationship between axonal integrity and cognitive status, consistent with the disconnection syndrome hypothesis [8].\u003c/p\u003e\n\u003cp\u003e4.4 Inferential limitations: partial correlations and effect size\u003c/p\u003e\n\u003cp\u003eBoth primary correlations lose significance when controlling for age: SDMT–NAA/Cr R1 (partial r = 0.205, p = 0.148) and RAVLT z–NAA/Cr R5 (r = 0.151, p = 0.290). NAA/Cr may reflect general aging rather than MS-specific pathology; alternatively, aging and MS neurodegeneration share common mechanisms (oxidative stress, mitochondrial dysfunction), and partial correlation may overcorrect by removing genuinely MS-relevant variance. Without longitudinal data, these interpretations cannot be distinguished.\u003c/p\u003e\n\u003cp\u003eExplained variance is modest (ρ² ≈ 12% for SDMT, ~6.8% for RAVLT z), positioning NAWM metabolic changes as one of several pathophysiological mechanisms of cognitive impairment in MS; comprehensive prediction requires multivariate models combining MRS, DTI, lesion burden, and volumetry.\u003c/p\u003e\n\u003cp\u003e4.5 Asymmetric distribution and the R4 artifact\u003c/p\u003e\n\u003cp\u003eRelative preservation of right frontal (R2) and right parietal (R6) NAWM may reflect asymmetric demyelination distribution known in MS; the R2–MSFC correlation (uncorrected for multiple testing) requires confirmation. The finding of higher NAA/Cr in RRMS at R4 (d = 1.602) likely reflects a methodological artifact due to corona radiata heterogeneity: hemispheric white matter asymmetry, differing shimming quality near the paranasal sinuses, or corpus callosum signal contamination. Without DTI co-registration, precise identification of the cause remains speculative.\u003c/p\u003e\n\u003cp\u003e4.6 Fatigue, depression, and neurobiological specificity\u003c/p\u003e\n\u003cp\u003eThe absence of correlation between NAA/Cr and FIS or DASS-21 (all p \u0026gt; 0.50) provides evidence of specificity of MRS markers for cognitive outcomes, consistent with the literature on different mechanisms of fatigue in MS [20, 21]. Depressive symptoms correlate with RAVLT z-score (ρ = −0.396, p = 0.004), but the partial correlation of NAA/Cr R5–RAVLT controlling for age and depression (r = 0.162, p = 0.257) is numerically higher than without controlling for depression (r = 0.151), suggesting that depression does not act as a confounder of this association [22].\u003c/p\u003e\n\u003cp\u003e4.7 Clinical implications and perspectives\u003c/p\u003e\n\u003cp\u003eThe discriminative ability of NAA/Cr R1 (AUC = 0.724) is formally superior to age alone (AUC = 0.515; DeLong ΔAUC = 0.208, p = 0.036), while adding age to the NAA/Cr R1 model does not improve discrimination (DeLong p = 0.706). This indicates that discriminative power derives from the metabolic marker rather than from an age effect. Replication in larger cohorts (n ≥ 150) with multivariate correction remains necessary. The fact that ~73% of patients on IFN-β still show below-average SDMT performance is consistent with the smouldering MS concept [23]; the difference from the prevalence in Drulović et al. [10] (52.7%) reflects the difference in applied thresholds (functional ≤ 56 vs. BICAMS z ≤ −1.5). Cognitive reserve may mitigate the rate of decline even in the presence of existing metabolic changes [24].\u003c/p\u003e\n\u003cp\u003e4.8 Study limitations\u003c/p\u003e\n\u003cp\u003eLimitations include: (1) cross-sectional design; (2) small sample size (power ≈ 48% for ρ ≈ 0.27; ≈ 72% for ρ = 0.35); (3) healthy controls without cognitive testing; (4) residual confounding by education; (5) RAVLT norms [12] based on a preprint study (n = 200); (6) MSFC correlations without multiple testing correction; (7) exclusive use of IFN-β limits generalizability; (8) no CSF partial volume correction — the creatine ratio partially compensates but incompletely, as CSF differentially affects metabolites; future studies should include absolute quantification.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eMulti-voxel \u0026sup1;H-MRS reveals topographically selective NAWM metabolic changes in early RRMS with domain-consistent associations: frontal NAA/Cr correlates with processing speed, while parietal NAA/Cr shows a trend association with verbal learning. Specificity for cognitive \u0026mdash; rather than affective \u0026mdash; outcomes supports the biological basis of the findings. Partial correlations lose significance when controlling for age, indicating synergistic aging and neurodegenerative processes whose contributions cannot be disentangled in the present sample. NAA/Cr R1 shows preliminary discriminative ability (AUC 0.724) formally superior to age alone (DeLong p\u0026thinsp;=\u0026thinsp;0.036), but definitive confirmation requires validation in larger cohorts with multivariate control for age, lesion burden, and atrophy.\u003c/p\u003e"},{"header":"Statements and Declarations","content":"\u003cp\u003eCompeting Interests: The authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003eEthics approval: This study was approved by the institutional Ethics Committee. Informed consent was obtained from all individual participants included in the study.\u003c/p\u003e\n\u003cp\u003eData availability: De-identified data are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003eFunding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eChiaravalloti ND, DeLuca J (2008) Cognitive impairment in multiple sclerosis. Lancet Neurol 7:1139\u0026ndash;1151. https://doi.org/10.1016/S1474-4422(08)70259-X\u003c/li\u003e\n\u003cli\u003eSumowski JF, Benedict R, Enzinger C et al (2018) Cognition in multiple sclerosis: state of the field and priorities for the future. Neurology 90:278\u0026ndash;288. https://doi.org/10.1212/WNL.0000000000004977\u003c/li\u003e\n\u003cli\u003eKappos L, Wolinsky JS, Giovannoni G et al (2020) Contribution of relapse-independent progression vs relapse-associated worsening to overall confirmed disability accumulation in typical relapsing multiple sclerosis. JAMA Neurol 77:1132\u0026ndash;1140. https://doi.org/10.1001/jamaneurol.2020.1568\u003c/li\u003e\n\u003cli\u003eBarkhof F (2002) The clinico-radiological paradox in multiple sclerosis revisited. Curr Opin Neurol 15:239\u0026ndash;245. https://doi.org/10.1097/00019052-200206000-00003\u003c/li\u003e\n\u003cli\u003eDe Stefano N, Matthews PM, Fu L et al (1998) Axonal damage correlates with disability in patients with relapsing-remitting multiple sclerosis. Brain 121:1469\u0026ndash;1477. https://doi.org/10.1093/brain/121.8.1469\u003c/li\u003e\n\u003cli\u003eDe Stefano N, Narayanan S, Francis GS et al (2001) Evidence of axonal damage in the early stages of multiple sclerosis and its relevance to disability. Arch Neurol 58:65\u0026ndash;70. https://doi.org/10.1001/archneur.58.1.65\u003c/li\u003e\n\u003cli\u003eSchoonheim MM, Meijer KA, Geurts JJG (2015) Network collapse and cognitive impairment in multiple sclerosis. Front Neurol 6:82. https://doi.org/10.3389/fneur.2015.00082\u003c/li\u003e\n\u003cli\u003eDineen RA, Vilisaar J, Hlinka J et al (2009) Disconnection as a mechanism for cognitive dysfunction in multiple sclerosis. Brain 132:239\u0026ndash;249. https://doi.org/10.1093/brain/awn275\u003c/li\u003e\n\u003cli\u003eThompson AJ, Banwell BL, Barkhof F et al (2018) Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol 17:162\u0026ndash;173. https://doi.org/10.1016/S1474-4422(17)30470-2\u003c/li\u003e\n\u003cli\u003eDrulović J, Tončev G, Nadj Č et al (2022) Validation of the Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS) in a large cohort of relapsing-remitting MS patients. Acta Clin Croat 61:62\u0026ndash;68. https://doi.org/10.20471/acc.2022.61.01.08\u003c/li\u003e\n\u003cli\u003eBenedict RH, Drake AS, Irwin LN et al (2016) Benchmarks of meaningful impairment on the MSFC and BICAMS. Mult Scler 22:1874\u0026ndash;1882. https://doi.org/10.1177/1352458516633517\u003c/li\u003e\n\u003cli\u003eJanićijević N, Purić D (2025) Reliability and convergent validity of the Rey-Auditory Verbal Learning Test (RAVLT) against working memory tasks in a Serbian community sample. Preprint. https://osf.io/4ywpr/\u003c/li\u003e\n\u003cli\u003eBenjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 57:289\u0026ndash;300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x\u003c/li\u003e\n\u003cli\u003eDeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44:837\u0026ndash;845. https://doi.org/10.2307/2531595\u003c/li\u003e\n\u003cli\u003eSun J, Song H, Yang Y et al (2017) Metabolic changes in normal appearing white matter in multiple sclerosis patients using multivoxel magnetic resonance spectroscopy imaging. Medicine 96:e6534. https://doi.org/10.1097/MD.0000000000006534\u003c/li\u003e\n\u003cli\u003eHeckova E, Dal-Bianco A, Strasser B et al (2022) Extensive brain pathologic alterations detected with 7.0-T MR spectroscopic imaging associated with disability in multiple sclerosis. Radiology 303:199\u0026ndash;209. https://doi.org/10.1148/radiol.210614\u003c/li\u003e\n\u003cli\u003eInglese M, Li BS, Rusinek H et al (2003) Diffusely elevated cerebral choline and creatine in relapsing-remitting multiple sclerosis. Magn Reson Med 50:190\u0026ndash;195. https://doi.org/10.1002/mrm.10481\u003c/li\u003e\n\u003cli\u003eRiccitelli GC, Pagani E, Rodegher M et al (2019) Imaging patterns of gray and white matter abnormalities associated with PASAT and SDMT performance in relapsing-remitting multiple sclerosis. Mult Scler 25:204\u0026ndash;216. https://doi.org/10.1177/1352458517743091\u003c/li\u003e\n\u003cli\u003eWagner AD, Shannon BJ, Kahn I, Buckner RL (2005) Parietal lobe contributions to episodic memory retrieval. Trends Cogn Sci 9:445\u0026ndash;453. https://doi.org/10.1016/j.tics.2005.07.001\u003c/li\u003e\n\u003cli\u003eInduruwa I, Constantinescu CS, Gran B (2012) Fatigue in multiple sclerosis \u0026mdash; a brief review. J Neurol Sci 323:9\u0026ndash;15. https://doi.org/10.1016/j.jns.2012.08.007\u003c/li\u003e\n\u003cli\u003eTartaglia MC, Narayanan S, Francis SJ et al (2004) The relationship between diffuse axonal damage and fatigue in multiple sclerosis. Arch Neurol 61:201\u0026ndash;207. https://doi.org/10.1001/archneur.61.2.201\u003c/li\u003e\n\u003cli\u003eArnett PA, Barwick FH, Beeney JE (2008) Depression in multiple sclerosis: review and theoretical proposal. J Int Neuropsychol Soc 14:691\u0026ndash;724. https://doi.org/10.1017/S1355617708081174\u003c/li\u003e\n\u003cli\u003eGiovannoni G, Popescu V, Wuerfel J et al (2022) Smouldering multiple sclerosis: the real MS. Ther Adv Neurol Disord 15:17562864211066751. https://doi.org/10.1177/17562864211066751\u003c/li\u003e\n\u003cli\u003eSumowski JF, Rocca MA, Leavitt VM et al (2014) Brain reserve and cognitive reserve protect against cognitive decline over 4.5 years in multiple sclerosis. Neurology 82:1776\u0026ndash;1783. https://doi.org/10.1212/WNL.0000000000000433\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 2\u0026nbsp;\u003c/strong\u003eNAWM MRS ratios in RRMS vs. healthy controls (FDR-corrected)\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 213px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNAWM region\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRRMS (n=51) Mean \u0026plusmn; SD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHC (n=44) Mean \u0026plusmn; SD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep (FDR)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCohen\u0026apos;s d\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDirection\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eNAA/Cr (N-acetylaspartate/creatine)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR1: Frontal subcortical (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.843 \u0026plusmn; 0.247\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.016 \u0026plusmn; 0.258\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.009\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.709\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026darr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR2: Frontal subcortical (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.832 \u0026plusmn; 0.239\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.770 \u0026plusmn; 0.250\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.106\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.251\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR3: Deep white matter (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.969 \u0026plusmn; 0.278\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.686 \u0026plusmn; 0.372\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.195\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026darr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR4: Deep white matter (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.004 \u0026plusmn; 0.275\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.597 \u0026plusmn; 0.226\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.602\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR5: Parietal subcortical (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.077 \u0026plusmn; 0.404\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.230 \u0026plusmn; 0.315\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.417\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026darr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR6: Parietal subcortical (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.085 \u0026plusmn; 0.443\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.004 \u0026plusmn; 0.313\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.581\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.209\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eCho/Cr (choline/creatine)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR1: Frontal subcortical (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.187 \u0026plusmn; 0.184\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.939 \u0026plusmn; 0.173\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.381\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR2: Frontal subcortical (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.174 \u0026plusmn; 0.180\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.153 \u0026plusmn; 0.142\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.562\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.129\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR3: Deep white matter (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.189 \u0026plusmn; 0.173\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.792 \u0026plusmn; 0.179\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.245\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR4: Deep white matter (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.180 \u0026plusmn; 0.186\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.113 \u0026plusmn; 0.124\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.063\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.417\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR5: Parietal subcortical (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.127 \u0026plusmn; 0.203\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.984 \u0026plusmn; 0.169\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.758\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR6: Parietal subcortical (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.154 \u0026plusmn; 0.209\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.892 \u0026plusmn; 0.093\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.583\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003emI/Cr (myo-inositol/creatine)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR1: Frontal subcortical (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.527 \u0026plusmn; 0.130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.436 \u0026plusmn; 0.135\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.672\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR2: Frontal subcortical (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.548 \u0026plusmn; 0.136\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.575 \u0026plusmn; 0.139\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.157\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.198\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR3: Deep white matter (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.515 \u0026plusmn; 0.148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.405 \u0026plusmn; 0.150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.731\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR4: Deep white matter (R)\u0026Dagger;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.537 \u0026plusmn; 0.131\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.542 \u0026plusmn; 0.146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.462\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR5: Parietal subcortical (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.497 \u0026plusmn; 0.119\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.612 \u0026plusmn; 0.171\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.782\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026darr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR6: Parietal subcortical (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.520 \u0026plusmn; 0.189\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.402 \u0026plusmn; 0.110\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.744\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026uarr; MS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eData: mean \u0026plusmn; SD. *Value higher in MS; likely methodological artifact (see text). \u0026Dagger;One extreme value excluded (n = 50 for mI/Cr R4; see Methods). Bold p-values: FDR-corrected p \u0026lt; 0.05. n.s. not significant\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3\u0026nbsp;\u003c/strong\u003eSpearman correlation between NAWM NAA/Cr and clinico-cognitive outcomes\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOutcome\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMRS ratio\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNAWM region\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 80px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpearman \u0026rho;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 80px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 147px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePart. r (adj. age)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSDMT (processing speed)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNAA/Cr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR1: Front. subcort. (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e+0.350\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.012*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.205 (p=0.148, n.s.)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR2\u0026ndash;R6 all regions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRAVLT z-score (verbal learning)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNAA/Cr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR5: Pariet. subcort. (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e+0.261\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.064\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.151 (p=0.290, n.s.)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR1: Front. subcort. (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e+0.187\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.190\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMSFC composite\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNAA/Cr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR1: Front. subcort. (L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e+0.325\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.020*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR2: Front. subcort. (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e+0.375\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.007**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR6: Pariet. subcort. (R)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e+0.325\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.020*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFIS (fatigue)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAll\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAll 6 regions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDASS-21 (mood)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAll\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAll 6 regions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003en.s.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026rho; Spearman coefficient, \u0026rho;\u0026sup2; explained variance. Partial r adjusted for age (Pearson, residual method). *p \u0026lt; 0.05; **p \u0026lt; 0.01. n.s. not significant. RAVLT z-scores calculated using Serbian norms (Janićijević et al. [12]), stratified by sex and age group. FIS Fatigue Impact Scale, DASS-21 Depression Anxiety Stress Scales-21, MSFC Multiple Sclerosis Functional Composite\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"multiple sclerosis, magnetic resonance spectroscopy, normal-appearing white matter, cognitive impairment","lastPublishedDoi":"10.21203/rs.3.rs-9394188/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9394188/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBackground: Cognitive impairment affects 40–70% of multiple sclerosis (MS) patients. Regional domain-selective associations between normal-appearing white matter (NAWM) metabolites and cognition remain underexplored.\u003c/p\u003e\n\u003cp\u003eMethods: In this cross-sectional study, 52 early RRMS patients (EDSS 0–3; all on interferon-beta) and 44 healthy controls underwent multi-voxel 3D ¹H-MRS (3T, PRESS) across six supracallosal NAWM regions. Between-group metabolite differences were FDR-corrected. Spearman correlations and age-adjusted partial correlations assessed MRS–cognition associations. ROC analysis with DeLong testing evaluated discriminative ability.\u003c/p\u003e\n\u003cp\u003eResults: Significant NAA/Cr differences were found in four of six regions. Left frontal NAA/Cr (R1) correlated with SDMT (ρ = 0.350, p = 0.012) and discriminated below-average processing speed (AUC = 0.724; 95% CI: 0.553–0.880; DeLong vs. age: ΔAUC = 0.208, p = 0.036). Left parietal NAA/Cr (R5) showed a trend association with RAVLT z-score (ρ = 0.261, p = 0.064). Age-controlled partial correlations were directionally consistent but non-significant (both p \u0026gt; 0.10). Neither fatigue nor mood correlated with MRS parameters.\u003c/p\u003e\n\u003cp\u003eConclusion: Multi-voxel ¹H-MRS reveals topographically heterogeneous NAWM damage with domain-selective cognitive associations in early RRMS. NAA/Cr R1 shows preliminary discriminative ability formally superior to age alone, but requires confirmation in larger cohorts.\u003c/p\u003e","manuscriptTitle":"Regional normal-appearing white matter metabolic profiles reflect information processing speed and verbal learning in early RRMS: A multi-voxel ¹H-MRS cross-sectional study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-29 10:17:07","doi":"10.21203/rs.3.rs-9394188/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a2955abe-b3f7-4dd9-8a3e-ad8c28d629e3","owner":[],"postedDate":"April 29th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-04-30T09:48:25+00:00","index":43,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-29T10:17:07+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-29 10:17:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9394188","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9394188","identity":"rs-9394188","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}