The DTI-ALPS index covaries with regional thalamic tissue susceptibility rather than clinical severity in Parkinson’s disease: a cross-sectional multimodal MRI study

Abstract Background The diffusion tensor image analysis along the perivascular space (DTI-ALPS) index is widely used as a non-invasive proxy for glymphatic function, but its biological specificity is increasingly questioned. Its region of interest lies in white matter sharing thalamocortical architecture with susceptibility-rich deep nuclei, so tissue composition may contribute to the index. We asked what it covaries with in Parkinson’s disease. Methods Seventy-six patients with idiopathic Parkinson’s disease underwent 3.0 T multimodal MRI and 16 clinical scales in this single-centre study. Diffusion imaging was single-shell (b = 2000 s mm − 2 , 60 directions). We derived the DTI-ALPS index, its four directional diffusivity components, bi-tensor free water, and total and source-separated susceptibility in 84 deep grey-matter and thalamic regions. Age-adjusted Spearman partial correlations were used, with Benjamini–Hochberg false-discovery-rate control within each imaging marker, not across the battery. Basal-ganglia free water was a pre-specified positive control. Results In an exploratory 84-region screen, the index covaried with total susceptibility of the left centralis lateralis (intralaminar thalamus; ρ = −0.42, q = 0.007; family-wise permutation p= 0.010). It was regional, not global (global ρ = 0.01, unchanged with candidate regions excluded; ρ = −0.48 adjusted for it), and robust to atrophy, white-matter free water and perivascular spaces. Age and intralaminar susceptibility, not motor severity, predicted ALPS (R 2 = 0.40). The ratio-level effect was left-lateralised (right ρ = −0.25), but component analysis showed it to be denominator-driven, carried predominantly by Dyy in the projection region (left ρ = +0.38, q = 0.004; bilateral mean ρ = +0.46) rather than by the numerator, consistent across hemispheres and accounting for about 90% of between-subject variance. No ALPS–clinical association survived correction, whereas the four free-water compartments yielded nine. The study could detect moderate (|ρ| ≥0.32), not small, associations. Conclusions The DTI-ALPS index covaried with regional thalamic tissue susceptibility and age rather than clinical status, chiefly through the denominator of the ratio rather than its perivascular numerator. These cross-sectional associations do not establish causation; the localisation is exploratory. Where deep-nuclei composition is altered, a tissue measure should accompany DTI-ALPS; interpreting it as a specific clearance measure warrants caution.