Limbic system changes in Huntington's disease: diffusion tensor MRI data with ROI analysis

Full Text

Introduction

Huntington's disease (HD) is a severe progressive disorder with autosomal dominant inheritance, characterized by three core symptom domains: motor dysfunction (predominantly choreic hyperkinesias), cognitive decline, and psychiatric symptoms. Currently, diagnosis relies on genetic testing and clinical examination. Modern neuroimaging techniques may be of great value for differential diagnosis of HD from clinically similar phenotypes, as well as for monitoring the progression of neurodegeneration, particularly in light of emerging innovative disease-modifying therapies.

Traditionally, HD has been viewed as a neurodegenerative disease with primary and most severe involvement of the striatum. However, accumulating evidence over the past decade strongly indicates that HD is a whole-brain disease, with early involvement of both gray and white matter [1, 2]. Current models of HD pathogenesis are shifting from the concept of isolated basal ganglia damage toward an understanding of the disease as a systemic disruption of large-scale neural networks [3]. Within this framework, damage to white matter tracts is of particular interest, and the degree of damage to specific tracts may be related to their role in the organization of these networks (known as "hubs") [4].

Diffusion tensor MRI (DT-MRI) studies have consistently demonstrated widespread microstructural white matter changes both in manifest HD patients and at the premanifest stage, years before the onset of overt motor symptoms [5]. For a long time, research has focused primarily on tracts associated with motor deficits. However, meta-analyses and data from large consortia such as ENIGMA-HD have convincingly shown that the corpus callosum may be involved very early in the pathological process, and that changes observed in this tract serve as a highly sensitive disease marker [6].

It is well established that affective and behavioral symptoms (apathy, depression, irritability) are the earliest manifestations of HD and often precede motor symptoms by years, significantly impairing patients' quality of life. It is hypothesized that these symptoms arise from damage to limbic system structures [7]. However, in contrast to motor tracts, the pattern and temporal dynamics of microstructural changes in key limbic white matter pathways — such as the fornix, cingulum, and inferior longitudinal fasciculus (ILF) — have been studied to a much lesser extent. Available evidence points to early changes in the fornix, detectable both by MRI morphometry (volume reduction) and by pathological examination (demyelination, reduced oligodendrocyte gene expression), already at the prodromal stage of HD [8]. DT-MRI studies have also demonstrated reduced fractional anisotropy (FA) in the cingulum and ILF in manifest patients, with these changes correlating with the severity of cognitive impairment [9, 10]. A comprehensive analysis of damage patterns in these tracts using additional diffusion metrics (axial diffusivity — AD, and radial diffusivity — RD) to differentiate axonal degeneration from demyelination, as well as their association with the results of affective and cognitive impairment assessment using specific instruments at different disease stages, remains underexplored.

Thus, despite the recognized importance of non-motor symptoms and the role of the limbic system in HD pathogenesis, key questions regarding the specific patterns of limbic white matter tract damage at different disease stages remain unresolved. These gaps can be addressed using modern DT-MRI and ROI-based analysis, which can overcome limitations related to fiber crossing and allow assessment of the contribution of distinct damage patterns (axonopathy or demyelination) within each tract.

We hypothesized that microstructural changes in limbic tracts in HD are heterogeneous: in the fornix (as the main efferent pathway of the hippocampus), changes arise already at the premanifest stage, whereas the cingulum and ILF are affected predominantly at the manifest stage. We also expected that the patterns of AD and RD changes would differ between tracts, reflecting a predominance of axonal damage or demyelination, and that the observed diffusion changes would correlate with the severity of affective and cognitive impairment.

The aim of this study was to assess, using ROI-based analysis, the microstructural state of key limbic tracts (cingulum, fornix, and ILF) in patients at manifest and premanifest stages of HD, and to analyze correlations between diffusion metrics and neuropsychological test scores (including anxiety and depression scales) and genetic markers.

Materials and methods

Participants

The study enrolled 64 individuals divided into three groups matched for sex and age (Table 1). The first group consisted of 24 patients with a confirmed diagnosis of HD based on genetic testing, who had clear clinical manifestations of the disease (symptomatic HD). The second group comprised 20 patients who were carriers of the mutant gene according to genetic testing but had no obvious clinical manifestations of the disease (presymptomatic HD). The third group (control group) included 20 healthy volunteers without the mutant gene. The study was approved by the Local Ethics Committee of the Federal State Budgetary Scientific Institution "Russian Center of Neurology and Neurosciences" (Protocol No. 10-12/23 dated December 20, 2023). All patients provided written informed consent for DNA diagnostics, clinical examination, and brain MRI according to a specialized protocol.

 

Table 1. Demographic characteristics of patient groups

Characteristic	Group 1 (n = 24)	Group 2 (n = 20)	Control group (n = 20)	p
Sex:
male	14 (58.3%)	9 (45%)	10 (50%)	0.671
female	10 (41.7%)	11 (55%)	10 (50%)
Age at examination, years	48 [40; 61]	32 [27; 37]	48 [24; 60]	0.003

 

Non-inclusion criteria for all groups included: concomitant neurological diseases (other than HD), severe medical conditions in the decompensation stage, psychiatric disorders preventing compliance with instructions during MRI and neuropsychological testing, history of traumatic brain injury with loss of consciousness, cerebrovascular diseases (history of acute cerebrovascular accident), intracranial space-occupying lesions, and contraindications to MRI.

Criteria for manifest HD: confirmed HTT mutation (CAG repeat count ≥ 36 in the mutant allele) and presence of clinically significant motor manifestations (score ≥ 4 on the UHDRS motor subscale).

Criteria for premanifest HD: confirmed HTT mutation (CAG repeat count ≥ 36) and absence of clinically significant motor manifestations (score on the UHDRS motor subscale corresponding to a diagnostic confidence level of ≤ 3).

The first two groups underwent detailed neurological examination, assessment of general and family history, testing using scales for motor function (Unified Huntington's Disease Rating Scale — UHDRS), cognitive function (Montreal Cognitive Assessment — MoCA, Frontal Assessment Battery — FAB), affective domain (Hospital Anxiety and Depression Scale — HADS), attention and psychomotor speed (Schulte tables), as well as brain MRI on a Siemens Magnetom Prisma 3T scanner according to a specialized protocol including both standard (anatomical) sequences and diffusion tensor MRI with subsequent reconstruction of the white matter tracts of interest. The third group also underwent brain MRI using the same protocol.

In patients of Group 1, a comparative analysis of diffusion metrics (FA, AD, RD, and mean diffusivity, MD) was performed for two tracts: the cingulum and the ILF, separately in the right and left cerebral hemispheres. The fornix was excluded from the comparative analysis due to pronounced distortions of the physical tract model caused by hydrocephalus ex vacuo and the resulting significant alteration of diffusion metrics, which does not reflect actual structural changes and leads to obvious distortion of the final results. Diffusion metrics were compared between manifest patients, premanifest carriers, and the control group. Correlations were assessed between diffusion metrics and disease duration, UHDRS, MoCA, FAB, HADS (separately for anxiety and depression subscales), Schulte tables (separately for mean completion time and exhaustion index T4/T1), and CAG repeat count in the mutant HTT allele.

In patients of Group 2, a comparative analysis of diffusion metrics was performed in two regions of interest within each studied tract, separately in the right and left cerebral hemispheres. Correlations were assessed between diffusion metrics and HADS (separately for anxiety and depression subscales), Schulte tables (separately for mean completion time and exhaustion index T4/T1), CAG repeat count, CC/IT index (ratio of the width of the anterior horns to the intercaudate distance), FH/CC (bicaudate index, assessed on MRI and reflecting the degree of internal hydrocephalus), and the CAP score (CAG-Age Product), a calculated measure reflecting the estimated time to clinical onset of HD.

MRI protocol

DT-MRI was performed on a Siemens Magnetom Prisma 3T scanner using an echo-planar imaging sequence with multiple b-values of 1000 and 2500 s/mm². Echo time was 82 ms, repetition time — 5600 ms. The number of diffusion gradient directions for each b-value was 64. Voxel size was 2 × 2 × 2 mm. Data were adjusted using TOPUP and EDDY protocols, along with motion correction algorithms [11]. Data were normalized to standard MNI space using the ICBM152 atlas to ensure accurate application of tract templates across participants with varying brain anatomy.

Region of interest definition and tract reconstruction

Tract reconstruction was performed using a deterministic fiber tracking algorithm in DSI Studio software [12, 13]. Tractography parameters (Table 2) were set individually for each tract studied, followed by the creation of a template for group-level analysis.

 

Table 2. Main parameters for segmentation of cerebral white matter tracts

Parameter	Cingulum	Fornix	ILF
Tracking threshold	0.08	0.10	0.08
Angular threshold	55	70	45
Step size, mm	1.0	0.5	0.7
Minimum length, mm	30	30	30
Maximum length, mm	200	200	200
Termination conditions (tracts, seeds)	20,000 (tracts)	5,000 (tracts)	20,000 (seeds)

 

In accordance with the study objectives, three key limbic white matter pathways were selected: the cingulum, the fornix, and the ILF. The choice of these tracts is based on their central role in emotional responses, memory processes, attention, and visuo-affective integration — functions that are impaired already at early stages of HD and manifest clinically as irritability, apathy, anxiety, and reduced concentration [7].

The cingulum is the main association tract of the limbic lobe, connecting the cingulate cortex, prefrontal cortex, hippocampus, and other limbic structures. It plays a key role in emotional regulation, selective attention, and cognitive control. Damage to this tract in HD may be a direct cause of affective and behavioral symptoms [9, 10].

The fornix is the main efferent pathway of the hippocampus, transmitting information from the hippocampus to the mammillary bodies, septal nuclei, and hypothalamus. This pathway is crucial for declarative memory consolidation and emotional learning. Its damage explains the memory impairments observed in some HD patients [8].

The ILF is a large association tract connecting the occipital (visual) cortex with temporal structures, including the amygdala and hippocampus. It is involved in rapid transmission of visual information to limbic centers, which is necessary for recognizing emotionally salient stimuli and generating appropriate behavioral responses. Changes in this tract may correlate with emotional dysregulation and impaired visuo-affective processing [9, 10].

Templates of the above-described tracts were created based on data from healthy participants in the control group. These templates were then applied to reconstruct the tracts for each study participant (with identical tractography parameters). Subsequently, in each segmented tract, regions of interest (ROIs) were manually placed at predefined levels. All ROIs were placed manually by a single investigator to minimize inter-operator variability, using standard anatomical landmarks in DSI Studio software. ROI placement was performed in a plane perpendicular to the main fiber orientation of the tract, at sites of greatest fiber density and anatomical stability, avoiding gray matter regions and potential fiber crossings with other tracts.

In the cingulum, ROIs were placed in the coronal plane (perpendicular to the fiber orientation) at the level of the genu and body of the cingulum (Fig. 1).

 

Fig. 1. Overlay of anatomical data and MR tractography results (sagittal plane) in the cingulum.

A — left cingulum. Red and green spheres indicate ROI placement used for diffusion metric analysis. B — right cingulum. Yellow and blue spheres indicate ROI placement used for diffusion metric analysis.

 

In the fornix, ROIs were placed in the coronal plane (perpendicular to the fiber orientation) at the level of the body and crura of the fornix (Fig. 2).

 

Fig. 2. Overlay of anatomical data and MR tractography results (sagittal plane) in the fornix.

A — left fornix. Purple and pink spheres indicate ROI placement used for diffusion metric analysis. B — right fornix. Yellow and green spheres indicate ROI placement used for diffusion metric analysis.

 

In the ILF, ROIs were placed in the coronal plane (perpendicular to the fiber orientation) in the occipital and temporal lobe regions (Fig. 3).

 

Fig. 3. Overlay of anatomical data and MR tractography results (sagittal plane) in the ILF.

A — left ILF. Turquoise and pink spheres indicate ROI placement used for diffusion metric analysis. B — right ILF. Orange and purple spheres indicate ROI placement used for diffusion metric analysis.

 

Metric extraction

For each tract, mean values of diffusion metrics — FA, MD, AD, and RD — were extracted from the corresponding ROIs using the built-in tools of DSI Studio.

Statistical analysis

Statistical analysis was performed using SPSS Statistics v.27.0 (IBM). The null hypothesis was rejected at a significance level of p < 0.05 (for post-hoc pairwise comparisons, at an adjusted significance level of p_adj < 0.05). The distribution type of continuous variables was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests. For continuous variables with non-normal distribution, median and quartiles were used for description. For comparisons between two groups on continuous variables with non-normal distribution, the Mann–Whitney U test was used; for comparisons among three groups, the Kruskal–Wallis test was applied. In cases where significant overall differences between groups were found, post-hoc pairwise comparisons were performed (with Bonferroni correction for multiple comparisons). Spearman's rank correlation analysis was used to assess relationships between continuous variables. Given the large number of correlation tests and the relatively small sample size, correlation analysis between diffusion metrics and clinical-genetic parameters was exploratory in nature. No correction for multiple comparisons was applied for this family of tests, and the results should be considered hypothesis-generating for subsequent confirmatory studies.

Results

A comparative analysis of diffusion metrics (FA, MD, AD, and RD) was performed in the defined regions of interest of the cingulum, fornix, and ILF across the three groups: manifest patients, premanifest mutation carriers, and healthy controls (Table 3).

 

Table 3. Comparison of diffusion metrics in limbic system tracts (in each ROI) across the three patient groups, Me [Q₁; Q₃]

Parameter			Group 1 (n = 24)	Group 2 (n = 20)	Control group (n = 20)	p
Right cingulum	genu	FA	0.296 [0.246; 0.324]	0.337 [0.300; 0.369]	0.331 [0.292; 0.370]	0.023*
		MD	0.869 [0.827; 0.919]	0.824 [0.813; 0.839]	0.821 [0.786; 0.851]	0.021*
		AD	1.140 [1.113; 1.186]	1.129 [1.090; 1.172]	1.147 [1.071; 1.173]	0.661
		RD	0.752 [0.674; 0.782]	0.672 [0.662; 0.691]	0.682 [0.626; 0.718]	0.006*
	body	FA	0.345 [0.277; 0.392]	0.413 [0.369; 0.448]	0.447 [0.375; 0.480]	< 0.001*
		MD	0.850 [0.809; 0.886]	0.831 [0.785; 0.860]	0.789 [0.767; 0.823]	0.019*
		AD	1.184 [1.143; 1.202]	1.210 [1.177; 1.271]	1.235 [1.124; 1.302]	0.125
		RD	0.699 [0.631; 0.740]	0.602 [0.573; 0.660]	0.589 [0.546; 0.621]	< 0.001*
Left cingulum	genu	FA	0.300 [0.277; 0.332]	0.368 [0.328; 0.405]	0.383 [0.355; 0.410]	< 0.001*
		MD	0.860 [0.833; 0.900]	0.832 [0.816; 0.856]	0.797 [0.773; 0.835]	< 0.001*
		AD	1.158 [1.111; 1.231]	1.200 [1.137; 1.223]	1.171 [1.135; 1.201]	0.289
		RD	0.712 [0.690; 0.754]	0.654 [0.632; 0.685]	0.623 [0.590; 0.650]	< 0.001*
	body	FA	0.353 [0.314; 0.382]	0.420 [0.389; 0.455]	0.434 [0.408; 0.460]	< 0.001*
		MD	0.833 [0.801; 0.887]	0.820 [0.790; 0.855]	0.800 [0.763; 0.823]	0.024*
		AD	1.216 [1.097; 1.240]	1.246 [1.176; 1.295]	1.237 [1.182; 1.285]	0.092
		RD	0.669 [0.621; 0.731]	0.603 [0.571; 0.624]	0.580 [0.543; 0.609]	< 0.001*
Right fornix	body	FA	–	0.295 [0.283; 0.322]	0.305 [0.277; 0.331]	0.849
		MD	–	1.023 [0.941; 1.130]	1.000 [0.947; 1.194]	0.918
		AD	–	1.381 [1.264; 1.513]	1.316 [1.263; 1.663]	0.965
		RD	–	0.848 [0.788; 0.940]	0.846 [0.787; 0.966]	0.942
	crura	FA	–	0.330 [0.308; 0.396]	0.376 [0.358; 0.410]	0.023*
		MD	–	0.980 [0.920; 1.055]	1.014 [0.934; 1.053]	0.549
		AD	–	1.428 [1.251; 1.471]	1.466 [1.351; 1.579]	0.193
		RD	–	0.771 [0.709; 0.847]	0.788 [0.726; 0.819]	0.737
Left fornix	body	FA	–	0.322 [0.280; 0.343]	0.319 [0.288; 0.351]	0.849
		MD	–	1.115 [0.840; 1.614]	0.974 [0.859; 1.436]	0.804
		AD	–	1.489 [1.173; 2.088]	1.320 [1.182; 1.910]	0.650
		RD	–	0.900 [0.674; 1.377]	0.798 [0.707; 1.217]	0.759
	crura	FA	–	0.323 [0.312; 0.341]	0.357 [0.331; 0.388]	0.013*
		MD	–	1.131 [0.979; 1.366]	1.129 [1.031; 1.183]	0.826
		AD	–	1.575 [1.376; 1.841]	1.599 [1.464; 1.720]	0.715
		RD	–	0.908 [0.784; 1.128]	0.902 [0.826; 0.930]	0.474
Right ILF	posterior	FA	0.523 [0.480; 0.537]	0.570 [0.527; 0.602]	0.607 [0.570; 0.627]	< 0.001*
		MD	0.968 [0.943; 1.097]	0.877 [0.823; 0.983]	0.834 [0.806; 0.857]	< 0.001*
		AD	1.619 [1.535; 1.723]	1.514 [1.454; 1.668]	1.492 [1.435; 1.569]	0.006*
		RD	0.669 [0.613; 0.755]	0.568 [0.508; 0.651]	0.501 [0.475; 0.540]	< 0.001*
	anterior	FA	0.368 [0.326; 0.394]	0.390 [0.369; 0.411]	0.406 [0.385; 0.430]	0.006*
		MD	0.886 [0.844; 0.914]	0.841 [0.818; 0.876]	0.803 [0.773; 0.822]	< 0.001*
		AD	1.234 [1.172; 1.315]	1.198 [1.163; 1.239]	1.165 [1.128; 1.201]	0.020*
		RD	0.704 [0.667; 0.743]	0.658 [0.644; 0.685]	0.622 [0.600; 0.641]	< 0.001*
Left ILF	posterior	FA	0.508 [0.485; 0.533]	0.538 [0.512; 0.573]	0.570 [0.556; 0.607]	< 0.001*
		MD	0.937 [0.892; 1.048]	0.884 [0.843; 0.995]	0.821 [0.786; 0.855]	< 0.001*
		AD	1.534 [1.484; 1.648]	1.516 [1.438; 1.586]	1.456 [1.364; 1.531]	0.102
		RD	0.643 [0.600; 0.737]	0.603 [0.527; 0.698]	0.512 [0.465; 0.547]	< 0.001*
	anterior	FA	0.476 [0.437; 0.498]	0.507 [0.485; 0.527]	0.527 [0.507; 0.547]	< 0.001*
		MD	0.885 [0.836; 0.906]	0.799 [0.787; 0.830]	0.792 [0.780; 0.802]	< 0.001*
		AD	1.362 [1.292; 1.430]	1.298 [1.230; 1.357]	1.289 [1.234; 1.321]	0.028*
		RD	0.642 [0.593; 0.686]	0.560 [0.548; 0.592]	0.548 [0.523; 0.564]	< 0.001*

Note. Here and in the Table 4, * indicates significant correlations.

 

In patients of Groups 1 and 2, correlations between diffusion metrics and clinical scales (UHDRS, MoCA, FAB, HADS, Schulte tables) and genetic markers (CAG repeat count, CAP score, FH/CC and CC/IT indices) were also assessed. Due to the technical limitations previously described, Group 1 was excluded from the fornix analysis; therefore, for this tract, only Group 2 and the control group were compared.

Comparison of diffusion metrics between groups

Cingulum. Comparison across the three groups revealed statistically significant differences for most diffusion metrics in both hemispheres and in both ROIs (genu and body). In Group 1 patients compared to controls, FA was decreased, while MD and RD were increased; changes in AD did not reach statistical significance (Table 4). Post-hoc analysis showed that the differences were primarily driven by Group 1, whereas premanifest carriers did not differ significantly from controls. For example, in the right cingulum genu, Group 1 patients showed increased RD (p = 0.011); the decrease in FA did not reach statistical significance (p = 0.062). In the body of the right cingulum, Group 1 patients showed decreased FA (p < 0.001), increased RD (p < 0.001), and increased MD (p = 0.015). In the left hemisphere, similar changes were observed: in the genu — decreased FA (p < 0.001) and increased RD (p < 0.001); in the body — decreased FA (p < 0.001), increased MD (p = 0.024), and increased RD (p < 0.001). Premanifest carriers did not differ significantly from controls for any metric.

 

Table 4. Post-hoc pairwise comparisons (p_adj)

Parameter			Group 1 vs. Group 2	Group 1 vs. Control	Group 2 vs. Control
Right cingulum	genu	FA	0.044*	0.062	1.000
		MD	0.097	0.027*	1.000
		AD	Post-hoc comparisons not performed
		RD	0.024*	0.011*	1.000
	body	FA	0.020*	< 0.001*	0.751
		MD	0.650	0.015*	0.346
		AD	Post-hoc comparisons not performed
		RD	0.032*	< 0.001*	0.490
Left cingulum	genu	FA	0.003*	< 0.001*	1.000
		MD	0.235	< 0.001*	0.107
		AD	Post-hoc comparisons not performed
		RD	0.028*	< 0.001*	0.117
	body	FA	0.003*	< 0.001*	1.000
		MD	0.925	0.021*	0.277
		AD	Post-hoc comparisons not performed
		RD	0.007*	< 0.001*	0.605
Right ILF	posterior	FA	0.011*	< 0.001*	0.196
		MD	0.007*	< 0.001*	0.072
		AD	0.163	0.005*	0.734
		RD	0.007*	< 0.001*	0.039*
	anterior	FA	0.460	0.004*	0.278
		MD	0.081	< 0.001*	0.021*
		AD	0.680	0.016*	0.392
		RD	0.032*	< 0.001*	0.017*
Left ILF	posterior	FA	0.123	< 0.001*	0.055
		MD	0.398	< 0.001*	0.067
		AD	Post-hoc comparisons not performed
		RD	0.358	< 0.001*	0.066
	anterior	FA	0.030*	< 0.001*	0.515
		MD	0.001*	< 0.001*	0.714
		AD	0.090	0.053	1.000
		RD	0.002*	< 0.001*	0.461

 

Fornix. Comparisons were performed between Group 2 patients and the control group. Statistically significant differences were found only in the fornix crura: patients in Group 2 showed decreased FA bilaterally, both in the right (p = 0.023) and left (p = 0.013) crura. No significant differences were found for other metrics (MD, AD, RD) in either the body or the crura of the fornix. In the fornix body, all metrics were comparable between Group 2 and controls.

Inferior longitudinal fasciculus. Pronounced intergroup differences were observed in all regions and both hemispheres. In Group 1 patients compared to controls, FA was decreased, while MD, AD, and RD were increased (for most comparisons, p < 0.05). Post-hoc analysis showed that significant changes were predominantly driven by Group 1; however, in several instances, premanifest carriers also differed from controls. In the right ILF posterior portion, Group 1 patients showed decreased FA (p < 0.001) and increased MD (p < 0.001), AD (p = 0.005), and RD (p < 0.001); in the anterior portion — decreased FA (p = 0.004) and increased MD (p < 0.001), AD (p = 0.016), and RD (p < 0.001). In premanifest carriers, the right ILF anterior portion showed increased MD (p = 0.021) and RD (p = 0.017) compared to controls, while the posterior portion showed a trend toward increased RD (p = 0.039). In the left hemisphere, manifest patients showed significant changes in FA, MD, and RD (p < 0.05), while AD reached borderline significance (p = 0.053). Premanifest carriers did not show statistically significant differences from controls (posterior portion: p = 0.066 for RD and p = 0.055 for FA).

Correlation analysis

In the manifest patient group and premanifest carriers, correlation analysis was performed between diffusion metrics and clinical-genetic parameters (Tables 5, 6). The main clinically significant associations are summarized below.

 

Table 5. Correlation of diffusion metrics with clinical scales, disease duration, and CAG repeat count in Group 1 patients (n = 24)

Parameter			Disease duration	UHDRS	MoCA	FAB	HADS A	HADS D	Schulte (time)	Schulte (exhaustion)	CAG repeats
Right cingulum	genu	FA	r = –0.259	r = –0.231	r = 0.212	r = 0.282	r = –0.294	r = 0.033	r = –0.165	r = 0.040	r = 0.004
			p = 0.270	p = 0.328	p = 0.369	p = 0.229	p = 0.208	p = 0.892	p = 0.486	p = 0.867	p = 0.987
		MD	r = 0.086	r = 0.169	r = –0.137	r = –0.139	r = 0.058	r = –0.067	r = 0.239	r = 0.008	r = 0.038
			p = 0.718	p = 0.477	p = 0.564	p = 0.558	p = 0.807	p = 0.780	p = 0.310	p = 0.972	p = 0.874
		AD	r = –0.034	r = –0.063	r = 0.120	r = 0.159	r = –0.194	r = 0.025	r = 0.111	r = 0.024	r = –0.109
			p = 0.886	p = 0.793	p = 0.614	p = 0.503	p = 0.413	p = 0.917	p = 0.640	p = 0.920	p = 0.648
		RD	r = 0.212	r = 0.216	r = –0.196	r = –0.216	r = 0.174	r = –0.002	r = 0.171	r = 0.014	r = 0.080
			p = 0.369	p = 0.360	p = 0.407	p = 0.361	p = 0.463	p = 0.992	p = 0.470	p = 0.955	p = 0.738
	body	FA	r = –0.488	r = 0.155	r = 0.343	r = 0.142	r = –0.328	r = –0.358	r = –0.188	r = 0.366	r = 0.239
			p = 0.029	p = 0.513	p = 0.139	p = 0.549	p = 0.158	p = 0.121	p = 0.427	p = 0.112	p = 0.311
		MD	r = 0.420	r = –0.095	r = –0.315	r = –0.070	r = 0.270	r = 0.337	r = 0.143	r = –0.351	r = 0.003
			p = 0.065	p = 0.690	p = 0.175	p = 0.768	p = 0.249	p = 0.146	p = 0.548	p = 0.129	p = 0.990
		AD	r = –0.091	r = 0.083	r = 0.036	r = 0.013	r = 0.093	r = 0.038	r = 0.089	r = –0.105	r = 0.226
			p = 0.702	p = 0.728	p = 0.879	p = 0.957	p = 0.696	p = 0.874	p = 0.710	p = 0.661	p = 0.337
		RD	r = 0.476	r = –0.060	r = –0.411	r = –0.140	r = 0.340	r = 0.408	r = 0.188	r = –0.357	r = –0.084
			p = 0.034	p = 0.801	p = 0.072	p = 0.556	p = 0.142	p = 0.074	p = 0.427	p = 0.123	p = 0.724
Left cingulum	genu	FA	r = –0.123	r = –0.349	r = 0.250	r = 0.002	r = 0.082	r = –0.087	r = –0.229	r = –0.077	r = –0.016
			p = 0.607	p = 0.131	p = 0.288	p = 0.992	p = 0.732	p = 0.715	p = 0.332	p = 0.745	p = 0.947
		MD	r = –0.155	r = 0.029	r = –0.243	r = –0.294	r = –0.018	r = 0.126	r = 0.274	r = –0.196	r = –0.156
			p = 0.513	p = 0.902	p = 0.302	p = 0.208	p = 0.939	p = 0.597	p = 0.243	p = 0.407	p = 0.512
		AD	r = –0.229	r = –0.403	r = 0.138	r = 0.005	r = –0.044	r = –0.038	r = 0.045	r = –0.233	r = –0.150
			p = 0.331	p = 0.078	p = 0.561	p = 0.982	p = 0.854	p = 0.874	p = 0.850	p = 0.322	p = 0.527
		RD	r = 0.017	r = 0.275	r = –0.395	r = –0.295	r = –0.004	r = 0.189	r = 0.373	r = –0.033	r = –0.103
			p = 0.944	p = 0.240	p = 0.084	p = 0.206	p = 0.987	p = 0.426	p = 0.105	p = 0.890	p = 0.665
	body	FA	r = –0.165	r = 0.023	r = 0.236	r = 0.326	r = –0.421	r = –0.052	r = –0.177	r = 0.331	r = –0.094
			p = 0.486	p = 0.922	p = 0.316	p = 0.161	p = 0.064	p = 0.827	p = 0.454	p = 0.154	p = 0.693
		MD	r = 0.138	r = 0.170	r = –0.508	r = –0.239	r = 0.227	r = 0.191	r = 0.531	r = –0.036	r = 0.045
			p = 0.562	p = 0.473	p = 0.022	p = 0.310	p = 0.335	p = 0.420	p = 0.016	p = 0.880	p = 0.851
		AD	r = 0.065	r = 0.031	r = –0.222	r = 0.073	r = –0.019	r = 0.133	r = 0.323	r = 0.086	r = –0.119
			p = 0.784	p = 0.897	p = 0.347	p = 0.758	p = 0.937	p = 0.577	p = 0.164	p = 0.717	p = 0.619
		RD	r = 0.206	r = 0.181	r = –0.555	r = –0.433	r = 0.428	r = 0.200	r = 0.460	r = –0.243	r = 0.068
			p = 0.383	p = 0.445	p = 0.011	p = 0.057	p = 0.060	p = 0.398	p = 0.041	p = 0.302	p = 0.775
Right ILF	posterior	FA	r = 0.240	r = –0.410	r = 0.230	r = 0.513	r = –0.068	r = –0.092	r = –0.180	r = 0.331	r = –0.079
			p = 0.258	p = 0.047	p = 0.281	p = 0.010	p = 0.751	p = 0.671	p = 0.400	p = 0.115	p = 0.715
		MD	r = –0.046	r = 0.253	r = –0.031	r = –0.249	r = 0.121	r = –0.046	r = 0.083	r = –0.185	r = 0.156
			p = 0.832	p = 0.234	p = 0.887	p = 0.241	p = 0.572	p = 0.831	p = 0.701	p = 0.387	p = 0.466
		AD	r = 0.123	r = 0.010	r = 0.175	r = 0.071	r = 0.126	r = –0.135	r = –0.103	r = –0.017	r = 0.246
			p = 0.566	p = 0.965	p = 0.413	p = 0.742	p = 0.558	p = 0.530	p = 0.632	p = 0.936	p = 0.247
		RD	r = –0.158	r = 0.249	r = –0.094	r = –0.336	r = 0.018	r = –0.012	r = 0.128	r = –0.243	r = 0.080
			p = 0.461	p = 0.241	p = 0.663	p = 0.109	p = 0.935	p = 0.955	p = 0.550	p = 0.252	p = 0.710
	anterior	FA	r = 0.081	r = –0.085	r = –0.040	r = 0.143	r = –0.462	r = –0.157	r = 0.048	r = 0.341	r = –0.068
			p = 0.707	p = 0.695	p = 0.852	p = 0.504	p = 0.023	p = 0.463	p = 0.823	p = 0.103	p = 0.752
		MD	r = 0.202	r = 0.348	r = –0.243	r = –0.140	r = 0.098	r = 0.233	r = 0.127	r = –0.157	r = 0.001
			p = 0.344	p = 0.096	p = 0.253	p = 0.515	p = 0.649	p = 0.273	p = 0.556	p = 0.465	p = 0.995
		AD	r = 0.206	r = 0.216	r = –0.327	r = –0.063	r = –0.136	r = 0.168	r = 0.287	r = 0.182	r = –0.100
			p = 0.334	p = 0.312	p = 0.118	p = 0.769	p = 0.528	p = 0.432	p = 0.174	p = 0.395	p = 0.642
		RD	r = 0.010	r = 0.232	r = –0.031	r = –0.082	r = 0.368	r = 0.292	r = –0.087	r = –0.367	r = 0.142
			p = 0.963	p = 0.275	p = 0.887	p = 0.702	p = 0.077	p = 0.165	p = 0.686	p = 0.078	p = 0.507
Left ILF	posterior	FA	r = –0.137	r = –0.494	r = 0.520	r = 0.619	r = –0.267	r = –0.169	r = –0.291	r = 0.281	r = 0.191
			p = 0.522	p = 0.014	p = 0.009	p = 0.001	p = 0.207	p = 0.431	p = 0.168	p = 0.184	p = 0.372
		MD	r = –0.051	r = 0.409	r = –0.269	r = –0.421	r = 0.241	r = 0.147	r = 0.135	r = –0.334	r = –0.010
			p = 0.812	p = 0.047	p = 0.203	p = 0.041	p = 0.257	p = 0.493	p = 0.529	p = 0.111	p = 0.964
		AD	r = –0.064	r = 0.350	r = –0.185	r = –0.304	r = 0.224	r = 0.146	r = 0.110	r = –0.251	r = 0.019
			p = 0.768	p = 0.094	p = 0.386	p = 0.149	p = 0.292	p = 0.495	p = 0.607	p = 0.237	p = 0.929
		RD	r = –0.034	r = 0.433	r = –0.359	r = –0.510	r = 0.218	r = 0.149	r = 0.190	r = –0.369	r = –0.053
			p = 0.874	p = 0.035*	p = 0.085	p = 0.011*	p = 0.307	p = 0.486	p = 0.373	p = 0.076	p = 0.807
	anterior	FA	r = 0.138	r = –0.180	r = 0.000	r = 0.115	r = –0.108	r = –0.144	r = –0.038	r = 0.486	r = –0.248
			p = 0.519	p = 0.399	p = 1.000	p = 0.593	p = 0.616	p = 0.501	p = 0.859	p = 0.016*	p = 0.242
		MD	r = 0.069	r = 0.134	r = –0.216	r = –0.219	r = 0.264	r = 0.280	r = –0.063	r = –0.435	r = 0.165
			p = 0.749	p = 0.532	p = 0.312	p = 0.304	p = 0.212	p = 0.185	p = 0.768	p = 0.034*	p = 0.441
		AD	r = 0.005	r = –0.197	r = 0.045	r = 0.087	r = –0.003	r = 0.041	r = –0.286	r = 0.087	r = –0.101
			p = 0.981	p = 0.355	p = 0.835	p = 0.685	p = 0.990	p = 0.850	p = 0.176	p = 0.686	p = 0.637
		RD	r = –0.090	r = 0.159	r = –0.167	r = –0.185	r = 0.176	r = 0.252	r = –0.007	r = –0.478	r = 0.293
			p = 0.677	p = 0.457	p = 0.436	p = 0.386	p = 0.411	p = 0.234	p = 0.974	p = 0.018*	p = 0.165

Note. Here and in the Table 5, correlations are presented without correction for multiple comparisons (exploratory analysis). *p < 0.05.

 

Table 6. Correlation of diffusion metrics with bicaudate index, CAP score, CAG repeat count, and clinical scales in Group 2 patients (n = 20)

Parameter			CAG repeats	FH/CC index	CC/IT index	CAP score	HADS A	HADS D	Schulte (time)	Schulte (exhaustion)
Right cingulum	genu	FA	r = 0.111	r = –0.097	r = –0.061	r = 0.166	r = –0.023	r = 0.041	r = 0.238	r = –0.099
			p = 0.642	p = 0.684	p = 0.799	p = 0.483	p = 0.924	p = 0.863	p = 0.313	p = 0.677
		MD	r = –0.378	r = –0.122	r = 0.106	r = –0.369	r = 0.196	r = 0.075	r = –0.304	r = 0.148
			p = 0.101	p = 0.609	p = 0.656	p = 0.109	p = 0.406	p = 0.753	p = 0.193	p = 0.533
		AD	r = –0.058	r = –0.249	r = 0.123	r = –0.090	r = 0.241	r = 0.101	r = 0.125	r = –0.077
			p = 0.809	p = 0.289	p = 0.606	p = 0.705	p = 0.306	p = 0.672	p = 0.600	p = 0.748
		RD	r = –0.202	r = 0.012	r = 0.154	r = –0.214	r = 0.074	r = –0.109	r = –0.298	r = 0.178
			p = 0.393	p = 0.960	p = 0.516	p = 0.366	p = 0.756	p = 0.649	p = 0.202	p = 0.454
	body	FA	r = –0.037	r = 0.025	r = –0.037	r = –0.099	r = –0.212	r = –0.072	r = –0.047	r = –0.102
			p = 0.876	p = 0.917	p = 0.877	p = 0.679	p = 0.371	p = 0.763	p = 0.845	p = 0.668
		MD	r = –0.093	r = –0.068	r = 0.092	r = 0.251	r = –0.063	r = –0.032	r = 0.104	r = 0.315
			p = 0.695	p = 0.774	p = 0.699	p = 0.287	p = 0.793	p = 0.893	p = 0.663	p = 0.176
		AD	r = –0.072	r = 0.090	r = –0.108	r = 0.068	r = –0.227	r = 0.030	r = 0.191	r = 0.259
			p = 0.762	p = 0.705	p = 0.652	p = 0.777	p = 0.335	p = 0.901	p = 0.420	p = 0.269
		RD	r = 0.005	r = –0.036	r = 0.059	r = 0.271	r = 0.008	r = 0.002	r = 0.122	r = 0.239
			p = 0.982	p = 0.880	p = 0.804	p = 0.248	p = 0.975	p = 0.992	p = 0.609	p = 0.310
Left cingulum	genu	FA	r = 0.009	r = 0.019	r = –0.086	r = 0.037	r = –0.093	r = –0.032	r = –0.078	r = –0.198
			p = 0.970	p = 0.937	p = 0.718	p = 0.877	p = 0.697	p = 0.893	p = 0.743	p = 0.403
		MD	r = –0.340	r = –0.048	r = –0.025	r = –0.223	r = 0.112	r = 0.046	r = –0.132	r = 0.366
			p = 0.143	p = 0.840	p = 0.915	p = 0.345	p = 0.639	p = 0.848	p = 0.578	p = 0.112
		AD	r = –0.118	r = 0.182	r = –0.277	r = –0.038	r = 0.069	r = –0.188	r = –0.286	r = 0.094
			p = 0.621	p = 0.442	p = 0.236	p = 0.872	p = 0.773	p = 0.427	p = 0.222	p = 0.693
		RD	r = –0.339	r = –0.088	r = 0.123	r = –0.239	r = 0.044	r = 0.207	r = –0.027	r = 0.537
			p = 0.144	p = 0.712	p = 0.606	p = 0.311	p = 0.854	p = 0.382	p = 0.910	p = 0.015*
Left cingulum	body	FA	r = –0.211	r = 0.126	r = –0.241	r = –0.166	r = –0.064	r = –0.080	r = –0.224	r = 0.137
			p = 0.371	p = 0.597	p = 0.305	p = 0.486	p = 0.788	p = 0.739	p = 0.342	p = 0.565
		MD	r = 0.035	r = –0.039	r = 0.207	r = –0.186	r = –0.097	r = –0.298	r = –0.147	r = 0.065
			p = 0.884	p = 0.870	p = 0.382	p = 0.433	p = 0.683	p = 0.201	p = 0.535	p = 0.786
		AD	r = –0.027	r = 0.184	r = –0.154	r = –0.248	r = 0.096	r = –0.134	r = –0.170	r = 0.108
			p = 0.911	p = 0.438	p = 0.518	p = 0.291	p = 0.687	p = 0.574	p = 0.474	p = 0.649
		RD	r = 0.267	r = –0.053	r = 0.255	r = 0.072	r = –0.066	r = –0.164	r = –0.006	r = –0.163
			p = 0.254	p = 0.823	p = 0.278	p = 0.762	p = 0.783	p = 0.490	p = 0.980	p = 0.492
Right fornix	genu	FA	r = 0.464	r = –0.356	r = 0.300	r = 0.309	r = 0.082	r = 0.076	r = –0.302	r = –0.312
			p = 0.052	p = 0.147	p = 0.227	p = 0.212	p = 0.746	p = 0.763	p = 0.223	p = 0.208
		MD	r = –0.483	r = 0.019	r = 0.111	r = –0.059	r = –0.536	r = –0.257	r = 0.311	r = 0.424
			p = 0.042*	p = 0.942	p = 0.662	p = 0.816	p = 0.022*	p = 0.303	p = 0.210	p = 0.079
		AD	r = –0.441	r = 0.046	r = 0.106	r = 0.013	r = –0.554	r = –0.243	r = 0.290	r = 0.475
			p = 0.067	p = 0.855	p = 0.675	p = 0.958	p = 0.017*	p = 0.332	p = 0.243	p = 0.046*
		RD	r = –0.499	r = 0.088	r = 0.049	r = –0.118	r = –0.520	r = –0.288	r = 0.329	r = 0.420
			p = 0.035*	p = 0.729	p = 0.847	p = 0.642	p = 0.027*	p = 0.247	p = 0.182	p = 0.082
	body	FA	r = 0.104	r = 0.195	r = –0.187	r = –0.028	r = 0.139	r = 0.032	r = –0.370	r = –0.381
			p = 0.683	p = 0.438	p = 0.457	p = 0.913	p = 0.582	p = 0.898	p = 0.130	p = 0.119
		MD	r = –0.563	r = –0.334	r = 0.578	r = –0.084	r = –0.158	r = 0.088	r = –0.197	r = 0.133
			p = 0.015*	p = 0.176	p = 0.012*	p = 0.741	p = 0.532	p = 0.729	p = 0.433	p = 0.598
		AD	r = –0.523	r = –0.285	r = 0.498	r = –0.069	r = 0.002	r = 0.267	r = –0.404	r = –0.068
			p = 0.026*	p = 0.251	p = 0.035*	p = 0.785	p = 0.993	p = 0.285	p = 0.097	p = 0.788
		RD	r = –0.484	r = –0.338	r = 0.538	r = –0.060	r = –0.148	r = 0.106	r = 0.022	r = 0.221
			p = 0.042*	p = 0.170	p = 0.021*	p = 0.813	p = 0.557	p = 0.677	p = 0.932	p = 0.378
Left fornix	body	FA	r = 0.453	r = 0.193	r = –0.445	r = 0.211	r = 0.062	r = –0.021	r = –0.106	r = –0.028
			p = 0.059	p = 0.442	p = 0.064	p = 0.401	p = 0.806	p = 0.934	p = 0.675	p = 0.913
		MD	r = –0.388	r = –0.272	r = 0.581	r = 0.029	r = –0.320	r = –0.100	r = 0.133	r = –0.045
			p = 0.112	p = 0.275	p = 0.011*	p = 0.909	p = 0.196	p = 0.692	p = 0.598	p = 0.858
		AD	r = –0.300	r = –0.325	r = 0.573	r = 0.116	r = –0.369	r = –0.196	r = 0.135	r = –0.073
			p = 0.226	p = 0.188	p = 0.013*	p = 0.648	p = 0.131	p = 0.437	p = 0.593	p = 0.773
		RD	r = –0.448	r = –0.248	r = 0.583	r = –0.002	r = –0.277	r = –0.009	r = 0.158	r = –0.014
			p = 0.063	p = 0.321	p = 0.011*	p = 0.994	p = 0.266	p = 0.970	p = 0.531	p = 0.955
	crura	FA	r = 0.251	r = 0.093	r = –0.179	r = –0.021	r = –0.072	r = –0.179	r = –0.300	r = –0.126
			p = 0.315	p = 0.714	p = 0.478	p = 0.935	p = 0.778	p = 0.478	p = 0.226	p = 0.618
		MD	r = –0.052	r = 0.020	r = 0.283	r = 0.008	r = –0.387	r = –0.347	r = –0.265	r = –0.241
			p = 0.837	p = 0.938	p = 0.256	p = 0.974	p = 0.113	p = 0.158	p = 0.287	p = 0.336
		AD	r = 0.007	r = 0.051	r = 0.274	r = 0.060	r = –0.418	r = –0.398	r = –0.375	r = –0.348
			p = 0.977	p = 0.842	p = 0.271	p = 0.813	p = 0.084	p = 0.102	p = 0.126	p = 0.157
		RD	r = –0.098	r = –0.030	r = 0.325	r = 0.004	r = –0.381	r = –0.288	r = –0.236	r = –0.218
			p = 0.698	p = 0.906	p = 0.188	p = 0.987	p = 0.119	p = 0.247	p = 0.345	p = 0.385
Right ILF	posterior	FA	r = 0.106	r = –0.219	r = 0.069	r = 0.077	r = 0.395	r = 0.396	r = –0.059	r = 0.181
			p = 0.655	p = 0.353	p = 0.772	p = 0.748	p = 0.085	p = 0.084	p = 0.806	p = 0.444
		MD	r = 0.080	r = 0.160	r = –0.211	r = –0.101	r = –0.475	r = –0.438	r = –0.176	r = –0.340
			p = 0.738	p = 0.500	p = 0.371	p = 0.672	p = 0.035*	p = 0.053	p = 0.458	p = 0.142
		AD	r = 0.198	r = 0.091	r = –0.154	r = –0.029	r = –0.218	r = –0.240	r = –0.313	r = –0.375
			p = 0.404	p = 0.703	p = 0.516	p = 0.905	p = 0.355	p = 0.308	p = 0.179	p = 0.103
		RD	r = –0.024	r = 0.246	r = –0.179	r = –0.152	r = –0.484	r = –0.455	r = –0.110	r = –0.294
			p = 0.922	p = 0.296	p = 0.450	p = 0.522	p = 0.031*	p = 0.044*	p = 0.645	p = 0.208
Right ILF	anterior	FA	r = 0.463	r = 0.102	r = –0.214	r = 0.063	r = 0.236	r = –0.275	r = –0.126	r = –0.587
			p = 0.040*	p = 0.670	p = 0.364	p = 0.791	p = 0.315	p = 0.241	p = 0.596	p = 0.007*
		MD	r = 0.249	r = –0.382	r = 0.300	r = 0.627	r = –0.103	r = 0.026	r = 0.353	r = 0.093
			p = 0.289	p = 0.097	p = 0.198	p = 0.003*	p = 0.666	p = 0.913	p = 0.126	p = 0.696
		AD	r = 0.295	r = –0.251	r = 0.100	r = 0.379	r = 0.175	r = –0.197	r = –0.006	r = –0.363
			p = 0.206	p = 0.285	p = 0.675	p = 0.099	p = 0.460	p = 0.404	p = 0.980	p = 0.116
		RD	r = 0.068	r = –0.369	r = 0.307	r = 0.484	r = –0.213	r = 0.145	r = 0.383	r = 0.374
			p = 0.775	p = 0.110	p = 0.187	p = 0.031*	p = 0.367	p = 0.541	p = 0.095	p = 0.104
Left ILF	posterior	FA	r = 0.065	r = 0.080	r = –0.182	r = –0.087	r = 0.449	r = 0.357	r = 0.138	r = 0.077
			p = 0.787	p = 0.738	p = 0.442	p = 0.714	p = 0.047*	p = 0.122	p = 0.561	p = 0.745
		MD	r = –0.209	r = –0.004	r = 0.186	r = –0.047	r = –0.530	r = –0.326	r = –0.492	r = –0.248
			p = 0.377	p = 0.987	p = 0.432	p = 0.843	p = 0.016*	p = 0.161	p = 0.028*	p = 0.291
		AD	r = –0.374	r = 0.114	r = 0.037	r = –0.358	r = –0.207	r = –0.008	r = –0.483	r = –0.149
			p = 0.104	p = 0.631	p = 0.877	p = 0.121	p = 0.381	p = 0.974	p = 0.031*	p = 0.531
		RD	r = –0.207	r = –0.034	r = 0.216	r = –0.053	r = –0.542	r = –0.490	r = –0.412	r = –0.212
			p = 0.380	p = 0.887	p = 0.360	p = 0.825	p = 0.014*	p = 0.028*	p = 0.071	p = 0.369
	anterior	FA	r = 0.211	r = 0.213	r = –0.299	r = 0.078	r = 0.269	r = 0.239	r = 0.012	r = –0.053
			p = 0.371	p = 0.367	p = 0.200	p = 0.745	p = 0.251	p = 0.309	p = 0.960	p = 0.826
		MD	r = –0.355	r = 0.058	r = 0.019	r = 0.193	r = –0.153	r = 0.073	r = 0.111	r = 0.455
			p = 0.125	p = 0.808	p = 0.936	p = 0.414	p = 0.519	p = 0.761	p = 0.640	p = 0.044*
		AD	r = 0.122	r = 0.118	r = –0.161	r = 0.312	r = 0.267	r = 0.376	r = 0.152	r = 0.151
			p = 0.610	p = 0.620	p = 0.499	p = 0.181	p = 0.256	p = 0.103	p = 0.523	p = 0.525
		RD	r = –0.417	r = –0.185	r = 0.324	r = 0.099	r = –0.224	r = –0.084	r = 0.048	r = 0.217
			p = 0.067	p = 0.434	p = 0.164	p = 0.679	p = 0.343	p = 0.724	p = 0.840	p = 0.359

 

Group 1 patients. The most pronounced correlations were observed for the cingulum and the ILF, predominantly in the left hemisphere. In the left cingulum (body), increased MD and RD were associated with worse cognitive performance (MoCA) and longer task completion time on the Schulte tables. A trend toward a positive correlation between RD and the HADS anxiety subscale was also observed. In the left ILF (posterior portion), FA correlated positively with MoCA and FAB and negatively with UHDRS, whereas MD and RD showed inverse relationships. These findings indicate an association between microstructural changes in the ILF and the severity of motor and cognitive impairment.

Group 2 patients. In premanifest carriers, the most significant correlations were found for the fornix and the ILF. In the fornix (particularly on the right), MD and RD correlated negatively with the HADS anxiety subscale and CAG repeat count, and positively with the CC/IT index. In the right ILF, FA correlated positively with CAG repeat count and negatively with Schulte exhaustion index; MD correlated positively with the CAP score. In the left ILF, MD and RD correlated negatively with HADS (anxiety and depression) and with task completion time on the Schulte tables.

Discussion

The biomarker issue is one of the central challenges in the current strategy against HD and other neurodegenerative diseases [14, 15], and modern neuroimaging methods play a pivotal role in this regard. The results obtained in this study suggest a heterogeneous pattern of white matter damage in HD, involving key limbic tracts: the cingulum, fornix, and ILF. Using ROI-based analysis and comprehensive assessment of diffusion metrics (FA, MD, AD, RD), we have shown that white matter changes in HD exhibit tract-specific patterns that differ both in their stage-dependent progression and in the nature of microstructural damage (axonal degeneration vs. demyelination). These findings are consistent with the current view of HD as a systemic disorder of large-scale neural networks [2, 4] and expand our understanding of the role of the limbic system in the development of early non-motor manifestations [16].

Cingulum involvement. In the cingulum of Group 1 patients, we observed decreased FA and increased MD and RD, with AD preserved. This pattern indicates predominant myelin damage (demyelination), which is consistent with the findings of R. Della Nave et al., who used TBSS to demonstrate decreased FA and increased RD in the cingulum of HD patients, interpreting these changes as indicative of demyelination [10]. Moreover, in a longitudinal study by C. Estevez-Fraga et al., changes in FA in association tracts, including the cingulum, were shown to progress in parallel with clinical deterioration [17], supporting our findings of changes characteristic of manifest-stage patients. In the present study, cingulum changes were observed only in Group 1; Group 2 patients did not differ significantly from controls. This suggests that the cingulum is involved relatively late compared to other tracts, and its damage may be associated with the progression of cognitive and affective symptoms as the disease advances. These observations are in line with the work of O. Phillips et al., who showed that deep white matter changes, including the cingulum, become more pronounced at later disease stages [18].

The most pronounced correlations in manifest patients were observed in the left cingulum: increased MD and RD were associated with worse cognitive performance (MoCA) and longer task completion time on the Schulte tables, reflecting attention and psychomotor speed. In addition, a trend toward a correlation between RD in the body of the left cingulum and the HADS anxiety subscale was observed, which may point to the role of this tract in emotional regulation. These findings are consistent with the work of P. Azevedo et al., who demonstrated an association between decreased FA in the cingulum and cognitive impairment in HD [9], as well as with the study by S. Gregory et al., which showed that microstructural white matter changes correlate with clinical manifestations of the disease [19].

Fornix involvement. In contrast to the cingulum, changes in the fornix were already detected at the premanifest stage — specifically, decreased FA in the fornix crura, with no significant changes in MD, AD, or RD. This pattern may reflect the earliest microstructural disturbances, preceding the development of demyelination or axonal degeneration. Our findings are consistent with the landmark study by Gabery et al., who used a multimodal approach (MRI morphometry, histology, RNA sequencing) to demonstrate early fornix pathology in HD — volume reduction, demyelination, and decreased oligodendrocyte gene expression already at the prodromal stage [8].

In Group 2 patients — carriers of the mutant HTT gene — negative correlations were found between MD and RD in the fornix, predominantly on the right, and the HADS anxiety subscale, as well as correlations with CAG repeat count and the CC/IT index. The negative correlation (greater anxiety corresponding to lower MD and RD values) diverges from the expectation that more pronounced affective symptoms should be accompanied by greater tract damage. This direction of association may be related to the limited size of Group 2 (n = 20) and the relatively narrow variability of anxiety scores at this stage, which makes the direction of the correlation unstable. Furthermore, the correlation analysis was exploratory in nature, without correction for multiple comparisons. Given these limitations, this result should be considered preliminary and requires validation in larger cohorts with longitudinal designs.

Notably, the positive correlations with the CC/IT index point to a relationship between white matter damage and striatal atrophy — a key structure involved in HD pathogenesis. This relationship is supported by studies showing correlations between microstructural white matter changes and volumetric gray matter changes [2, 18, 19].

ILF involvement. The ILF demonstrated the most pronounced changes among the limbic tracts studied. In manifest patients, we observed decreased FA and increased values across all three diffusion metrics (MD, AD, and RD), which overall suggests combined axonal and myelin damage. This may reflect the significant role of this pathway in highly connected networks. According to O. Phillips et al., several tracts, including the ILF, show early and pronounced changes in HD [18].

Of note, even in Group 2 patients, changes in the ILF were observed: in the right hemisphere, the anterior portion of the tract showed increased MD and RD, indirectly indicating myelin sheath damage. Changes in the left hemisphere were less pronounced, which may reflect interhemispheric asymmetry of degeneration. Such asymmetry has been described in studies of deep white matter in HD [18, 19].

Correlation analysis revealed multiple associations between the ILF and clinical parameters. In Group 1 patients, FA in the left ILF correlated positively with cognitive scales (MoCA, FAB) and negatively with the severity of motor impairment (UHDRS), while MD and RD showed inverse correlations. This confirms that the ILF is an integrative tract involved in a wide range of functions, from cognitive to motor. In Group 2 patients, the most noteworthy correlations were between the ILF and the HADS anxiety and depression subscales, as well as with the Schulte exhaustion index. This suggests that ILF changes may underlie early affective and cognitive symptoms that precede motor manifestation of the disease.

Clinico-radiological correlations and biomarker potential. Correlations between changes in the cingulum and ILF and cognitive scales (MoCA, FAB) confirm the role of these tracts in maintaining cognitive function. Correlations with the HADS subscales (anxiety, depression) in Group 2 patients, particularly in the fornix and ILF, provide direct neuroimaging evidence supporting the hypothesis that early affective symptoms in HD have an organic basis and are linked to limbic system damage. These findings are consistent with the conclusions of a systematic review by M.L. Clark et al., who noted a high prevalence of depression and apathy in mutation carriers [16], as well as with the study by S. Gregory et al., which demonstrated an association between microstructural white matter changes and clinical manifestations [19].

Of particular interest are the correlations with genetic markers in premanifest mutation carriers. The association between changes in the fornix and ILF (predominantly MD and RD) and CAG repeat count, CAP score, and CC/IT index indicates that the diffusion metrics used are sensitive to the earliest stages of the disease. These results are consistent with evidence that the degree of CAG repeat expansion in the mutant HTT allele determines the severity and rate of progression of the neurodegeneration [1, 2]. These findings also have important implications for the differential diagnosis of cognitive impairment, particularly in its early forms [20].

The potential of DT-MRI as a biomarker for preclinical diagnosis and therapy monitoring is supported by the work of C. Estevez-Fraga et al., who showed that longitudinal changes in diffusion metrics can serve as a sensitive indicator of disease progression [17]. Experimental data from animal models [21] further confirm that white matter changes detected by MRI reflect actual axonal and myelin damage.

Conclusions

The analysis and interpretation of the study results demonstrate that HD is characterized by distinct patterns of limbic system damage that begin at the premanifest stage and progress by the time of disease manifestation. In the cingulum, demyelination predominates; in the fornix, early changes are detected in premanifest carriers (decreased FA in the fornix crura with no significant changes in MD, AD, or RD); and the ILF shows more severe changes involving both axonal and myelin damage. The identified correlations with affective (HADS) and cognitive (MoCA, FAB, Schulte tables) scales, as well as with genetic markers (CAG repeat count, CAP score), confirm the clinical significance of changes in limbic tracts.

Combined assessment of AD and RD metrics provides unique information on the nature of white matter fiber damage that is not available when using FA and MD alone. These findings point to the potential value of DT-MRI with diffusion metric analysis as an adjunctive tool for characterizing microstructural changes in limbic tracts at different stages of HD. Further studies, including longitudinal designs, larger cohorts, and diagnostic accuracy analyses, are needed to evaluate the potential of these measures for preclinical diagnosis, progression monitoring, and, in the longer term, assessment of disease-modifying therapy efficacy.

Study limitations. This study has several limitations that should be considered when interpreting the results. The cross-sectional design precludes direct conclusions about causal relationships or the temporal dynamics of the observed changes. Further longitudinal studies are needed to confirm the stage-dependent involvement of different tracts. In addition, the relatively small sample size, particularly in subgroups (e.g., for correlation analyses), may have limited statistical power to detect weaker effects, especially after correction for multiple comparisons. One of the most important limitations is the exclusion of the manifest group from the fornix analysis due to technical constraints (hydrocephalus ex vacuo distorting tractography results), which prevents assessment of the changes in this important limbic tract at later disease stages. Finally, region-of-interest placement was performed manually by a single investigator, which may have introduced some subjectivity. Blinding of the operator to group assignment was not implemented due to the technical specifics of the study design. To minimize bias, ROIs were placed strictly according to anatomical landmarks based on a single template, regardless of group assignment.

About the authors

Georgy A. Anikin

Russian Center of Neurology and Neurosciences

Author for correspondence.
Email: dante3red@mail.ru
ORCID iD: 0009-0005-2447-1418

Postgraduate Student, 5th Neurology Department, Institute of Clinical and Preventive Neurology

Russian Federation, Moscow

Alexey S. Filatov

Russian Center of Neurology and Neurosciences

Email: dante3red@mail.ru
ORCID iD: 0000-0002-5706-6997

Cand. Sci. (Med.), Radiologist, Radiology Department

Russian Federation, Moscow

Alina A. Liaskovik

Russian Center of Neurology and Neurosciences

Email: dante3red@mail.ru
ORCID iD: 0000-0001-8062-0784

Radiologist, Radiology Department

Russian Federation, Moscow

Sergey A. Klyushnikov

Russian Center of Neurology and Neurosciences

Email: dante3red@mail.ru
ORCID iD: 0000-0002-8752-7045

Cand. Sci. (Med.), Leading Researcher, 5th Neurology Department, Institute of Clinical and Preventive Neurology

Russian Federation, Moscow

Nataliya Y. Abramycheva

Russian Center of Neurology and Neurosciences

Email: dante3red@mail.ru
ORCID iD: 0000-0001-9419-1159

Cand. Sci. (Biol.), Senior Researcher, Head, Molecular Genetics Laboratory, 5th Neurology Department, Institute of Clinical and Preventive Neurology

Russian Federation, Moscow

Sergey N. Illarioshkin

Russian Center of Neurology and Neurosciences; The Russian University of Medicine

Email: annaly-nevrologii@neurology.ru
ORCID iD: 0000-0002-2704-6282

Dr. Sci. (Med.), Professor, Full Member of the RAS, Director, Brain Institute, Deputy Director for Science; Head, Neurology Department

Russian Federation, Moscow; Moscow

References

Supplementary files