Differential tractography identifies a distinct pattern of white matter alterations in essential tremor with or without resting tremor

Graphical abstract

Highlights

• •Differential tractography detected subtle microstructural differences in Essential Tremor (ET).
• •Mean diffusivity was increased in ET and ET with resting tremor (rET) suggesting WM integrity loss.
• •ET and rET patients showed WM tract alterations, prominent in the left hemisphere in ET and in the right one in rET.
• •Decussating and non-decussating DRTT tracts were involved in rET, while only the non-decussating DRTT was affected in ET.
• •Our findings support the development of targeted diagnostic and therapeutic strategies based on tremor subtype.

Abstract

Essential Tremor (ET) is characterized by action tremor often associated with resting tremor (rET). Although previous studies have identified widespread brain white matter (WM) alterations in ET patients, differences between ET and rET have been less explored. In this study we employed differential tractography to investigate WM microstructural alterations in these tremor disorders.

We conducted a Diffusion Tensor Imaging (DTI) study on age- and sex-matched cohorts: 25 healthy controls (HC), 30 ET, and 30 rET patients. Differential tractography using DSI Studio was employed to pairwise compare fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) among cohorts.

ET and rET patients compared to HC exhibited similar widespread MD increase especially in basal ganglia and brainstem projections. WM changes were more pronounced in the left cerebral hemisphere and cerebellum (crus I and II) in ET, while in rET patients WM alterations were prevalent in right cerebral hemisphere and cerebellum crus I. Small FA decrease was found in rET but not in ET patients. ET patients showed changes in the left non-decussating dentato-rubro-thalamic tract (ndDRTT), whereas rET patients showed changes in both left ndDRTT and right decussating DRTT. In conclusion, our findings confirmed the DRTT involvement in essential tremor and demonstrated that ET and rET exhibited similar microstructural WM changes in the brain, with different hemispheric involvement—greater on the left side in ET and on the right side in rET—suggesting that these tremor disorders may be distinct subtypes of the same disease.

1. Introduction

Essential Tremor (ET) is one of the most common movement disorders, characterized primarily by action tremor (Bhatia et al., 2018, Louis and Ferreira, 2010). A subset of ET patients (ranging from 2 to 88 % of cases) also exhibits resting tremor (rET), adding complexity to the clinical profile of ET (Louis and Faust, 2020). At the present time, controversy exists on the classification of essential tremor associated with resting tremor, recently termed essential tremor plus (Louis et al., 2020). Some studies (Erro et al., 2022) suggested the hypothesis that essential tremor with resting tremor (essential tremor plus) reflects a different entity from ET, other studies hypothesized that rest tremor may be a late manifestation in essential tremor, and finally other authors suggested that rest tremor in ET might indicate superimposed Parkinson’s disease (Arabia et al., 2018). Thus, understanding the underlying neurobiological differences between ET patients with and without resting tremor is crucial for accurate diagnosis and targeted treatment strategies (Bhatia et al., 2018).

Magnetic Resonance Imaging (MRI) and specifically Diffusion Tensor Imaging (DTI) have provided valuable insights into the microstructural changes of the brain white matter (WM) associated with various neurological conditions (Figley et al., 2021). In ET, previous studies have identified alterations in cerebellar and thalamic pathways (Benito-Leon et al., 2017, Caligiuri et al., 2017, Jia et al., 2011, Klein et al., 2011, Novellino et al., 2016, Saini et al., 2012, Shin et al., 2008, Tantik Pak et al., 2021), which are crucial for motor control. More in detail, Klein et al. (Klein et al., 2011) found decreased FA in the superior cerebellar peduncle, a critical tract connecting the cerebellum to the thalamus, in ET patients compared to healthy controls. Pietracupa et al. (Pietracupa et al., 2019) detected widespread WM alterations in ET patients, mostly in the corticospinal tract, cerebellar peduncles, and corpus callosum.

DTI research on ET patients with resting tremor (rET) has revealed more extensive and severe microstructural changes compared to ET patients without resting tremor (Caligiuri et al., 2017), especially in a pathway connecting globus pallidus, caudate, and supplementary motor area.

The specific differences in brain tractography between ET and rET patients, however, remain less explored. In the current study, we first employed differential tractography technique based on a deterministic diffusion fiber tracking algorithm (Yeh et al., 2013) to compare the WM microstructure between ET patients with and without resting tremor. Differential tractography is an advanced imaging technique used to assess and analyze differences in WM tracts between different groups of subjects or within the same group over time (Yeh et al., 2016). Compared to conventional tractography that maps all pathways and is insensitive to neuronal change, differential tractography integrates an innovative “tracking-the-difference” paradigm into the fiber tracking process. This results in a new tractography modality that tracks the exact portion of pathways exhibiting substantial differences in diffusion metrics (Maffei et al., 2022, Maier-Hein et al., 2017, Yeh et al., 2013).

Differential tractography has been increasingly applied in neurological research to study various conditions, such as Parkinson’s disease (Mojtahed Zadeh et al., 2018, Sanchez-Catasus et al., 2021), restless legs syndrome (Park et al., 2023), traumatic brain injury (Huang et al., 2023, Li et al., 2022), brain tumors (Yeh et al., 2021), multiple sclerosis, Huntington’s disease, amyotrophic lateral sclerosis, and epilepsy (Yeh et al., 2019a). However, in the context of ET , differential tractography has been never applied, and the present study aims at investigating whether this novel approach can help distinguish the WM microstructural neuroanatomical differences between ET patients with and without resting tremor, compared to healthy controls.

2. Materials and methods

2.1. Participants

Sixty ET patients, comprising 30 with (rET) and 30 without resting tremor, along with 25 healthy controls (HC), were included in this study. Participants were recruited at the Institute of Neurology and the Neuroscience Research Center of Magna Graecia University, Catanzaro, Italy, between 2018 and 2022. All patients underwent comprehensive neurological examinations conducted by movement disorder specialists, and diagnoses of ET or rET were made based on the latest diagnostic criteria established by the Movement Disorder Society task force (Bhatia et al., 2018). All patients underwent a 3 T brain MRI and a single photon emission computed tomography with 123I-ioflupane (DaTscan). Moreover, in all rET patients, the presence of rest tremor was confirmed by surface electromyographic assessment (for accurate differentiation from classic ET) with the patient seated, the arm flexed, and the forearm fully supported on the chair armrest and the tremulous hand hanging down from the chair armrest. Exclusion criteria for both ET and rET patients were potential dysmetabolic causes of tremor, abnormal DaTscan, diffuse brain vascular lesions, or basal ganglia/brainstem lesions on MRI.

Participants were also excluded if they had current or past use of medications known to exacerbate or cause tremor, such as amiodarone, amphetamines, beta-adrenergic agonists, antipsychotics, prednisone, lithium, and valproate. None of the control subjects had any history of neurological, psychiatric, or major medical illnesses.

All study procedures and ethical aspects were approved by the institutional review board of Magna Graecia University, Catanzaro, Italy. Written informed consent was obtained from all participants.

2.2. Statistical analysis

Differences among the three groups in sex distribution was assessed by Fisher’s exact test, while age, disease duration, Mini Mental State Examination (MMSE) and education level in years were compared using a one-way analysis of variance (ANOVA) with Tukey’s post-hoc for pairwise comparisons. Statistical analysis was conducted with R language v.4.1.2.

2.3. MRI acquisition

All study participants underwent a brain MR with a 3 T-MR750 GE MRI scanner (8-channel head coil), with a recently described protocol (Calomino et al., 2024), including: a 3-dimensional T1-weighted volumetric spoiled gradient echo (GE) (sagittal section; repetition time/echo time 9.2/3.7 ms; slice thickness 1.0 mm; frequency and phase encoding matrix 256 x 256; flip angle 12°; field of view 25.6 mm), a T2-weighted fast spin echo (axial section; repetition time/echo time 5462/85 ms; slice thickness 4.0 mm; frequency and phase encoding matrix 512 x 256; field of view 24 mm), a T2-weighted fluid attenuated inversion recovery FLAIR (axial section; repetition time/echo time/inversion time 9500/100/2250 ms; slice thickness 4.0 mm (voxel size 2.0 × 2.0 × 4.0 mm³); frequency and phase-encoding matrix 512 x 256) sequences, and a diffusion-weighted volumes, acquired by using spin-echo echo-planar imaging (TE/TR = 87/10,000 ms, bandwidth 250KHz, matrix size 128 × 128, 80 axial slices, voxel size 2.0 × 2.0 × 2.0 mm³) with 27 equally distributed orientations for the diffusion-sensitizing gradients at a b-value of 1.000 s/mm². Head movements were minimized using foam pads around the participants’ heads.

2.4. DTI preprocessing

The DTI images preprocessing and further tractography were applied using the software DSI Studio (https://dsi-studio.labsolver.org) by Yeh et al. (Yeh et al., 2013). DICOM files were converted to SRC files adopting the batch procedure provided by DSI Studio. Images were corrected using DSI Studio option that calls FSL’s eddy to handle eddy current distortion and motion artefacts. The accuracy of b-table orientation was examined by comparing fiber orientations with those of a population-averaged template (Yeh, 2022). The restricted diffusion was quantified using restricted diffusion imaging (Yeh et al., 2017) and the diffusion data were reconstructed using generalized q-sampling imaging (Yeh et al., 2010) with a diffusion sampling length ratio of 1.25. The tensor metrics were calculated using DWI with b-value lower than 1750 s/mm².

The reconstruction step processes the SRC files to generate the FIB file, which stores the vector field (fiber orientations) and anisotropy information (the magnitude) that can be used by DSI Studio to conduct fiber tracking. We employed GQI with the default setting as reconstruction method, a deterministic fiber tracking algorithm (Yeh et al., 2013) with augmented tracking strategies (Yeh, 2020) to improve reproducibility. The anisotropy threshold was randomly selected. The angular threshold was randomly selected from 45 degrees to 90 degrees. The step size was set to voxel spacing. Tracks with length shorter than 35 or longer than 150 mm were discarded. Topology-informed pruning (Yeh et al., 2019b) was applied to the tractography with 32 iterations to remove false connections.

2.5. Differential tractography

Differential tractography reconstructs DTI data using advanced algorithms (Yeh, 2022), such as generalized q-sampling imaging (GQI) (Yeh et al., 2010), which improves the accuracy of fiber orientation estimation compared to traditional DTI methods (Maffei et al., 2022, Maier-Hein et al., 2017). The WM tracts are mapped out to investigate differences between groups (e.g., patients vs. controls) with tract-specific measures, such as FA, axial diffusivity (AD), radial diffusivity (RD), and mean diffusivity (MD), which are then statistically analyzed to identify significant changes (Yeh et al., 2016). Differential tractography enhances sensitivity by specifically targeting and comparing individual tracts, allowing for the detection of subtle microstructural differences that might be overlooked by conventional methods (Maffei et al., 2022, Maier-Hein et al., 2017). By incorporating topology-informed pruning (Yeh et al., 2019a, Yeh et al., 2019b, Yeh et al., 2019a), it reduces false positives and focuses on anatomically and functionally relevant tracts (Maffei et al., 2022, Maier-Hein et al., 2017), offering detailed connectivity analysis and enabling a more nuanced understanding of connectivity changes associated with different conditions or diseases (Maffei et al., 2022, Maier-Hein et al., 2017).

In this study, for each group comparison (HC-ET, HC-rET and ET-rET), we created four connectometry databases (one for each metric FA, AD, RD, MD). A connectometry database (Yeh et al., 2016) is necessary to aggregate multiple FIB files into a common template space that allows group-level analysis (ICBM-152 space, voxel size 1 × 1 × 1 mm³). Subjects demographic file was created and used to include group variables in the regression model: in the comparisons HC-ET and HC-rET, age and sex were used as covariates; in the comparison ET-rET, disease duration was also added as covariate together with age and sex. Each pairwise regression model provides two result maps, one of fiber bundles where the diffusion metric evaluated is higher in one cohort compared with the second cohort, and the other where the diffusion metric evaluated is lower in one cohort compared with the second cohort (e.g. ET > HC and ET < HC). The False Discovery Rate (FDR) threshold was set to 0.05 as highly confirmatory value to consider as output only tracks with a highly significant p-value. To estimate and correct the FDR for the type-I error inflation due to multiple comparisons, a total of 4,000 randomized permutations was applied to the group label to obtain the null distribution of the track length and to minimize the false positives.

2.6. Fiber clustering through network topology

To obtain the labelling of significantly different tracts among groups, we applied the automatic tract cluster option provided by DSI Studio, which employes the population-averaged atlas by Yeh et al. (Yeh, 2022), called HCP-842, to categorize WM fibers based on projection of basal ganglia and of brainstem, association, commissure and cerebellum pathways. This innovative population-level description of the structural connectome provides a detailed characterization of the normative 3D trajectories of white matter fascicles (Yeh, 2022). It also delineates how gray matter regions in the cerebrum, cerebellum, and brainstem are interconnected by nearly all macroscopic white matter pathways. The full atlas, including the track trajectories and connectograms, is publicly available at https://brain.labsolver.org.

Finally, to quantify the asymmetry between the hemispheres of the fiber clusters statistically significant different among groups, in term of number of tracts and volume, we applied the following formula:

$$\text{AI}=\text{Asymmetry index}=(\text{Left}-\text{Right}/\text{Left}+\text{Right})\ast \text{1}00,$$

where negative values mean a rightward lateralization, while positive ones mean leftward asymmetry for increasing metrics (MD). As further analysis, we performed a Spearman’s correlation between significantly different clusters in the three comparisons and the MMSE scores.

3. Results

3.1. Participants characteristics

The demographic and clinical characteristics of the three groups (rET, ET and HC) are summarized in Table 1.

Table 1.

Demographic and clinical characteristics of the three subject cohorts, healthy controls (HC), Essential Tremor (ET) and Essential tremor patients with resting tremor (rET).

	HC (n = 25)	ET (n = 30)	rET (n = 30)	p-value	Post-hoc
Age	68.76 ± 8.12	63.70 ± 9.97	63.83 ± 11.02	0.07	N.S.
Sex M/F	18/7	18/12	16/14	0.36	N.A
Disease Duration	N.A.	13.13 ± 13	18.93 ± 13.46	0.09	N.A.
MMSE	28 ± 1.89	25.93 ± 2.56	25.6 ± 2.38	0.01	HC > ET, rET
Rest tremor side R/L	N.A.	N.A.	20/10	N.A.	N.A.

Abbreviations: ET = essential tremor; rET = essential tremor with resting tremor; HC = healthy controls; R = Right; L = Left; N.S. = not significant at p < 0.05; N.A. = not applicable.

The analysis revealed no significant differences in age (p = 0.07) or sex distribution among the three groups. The duration of disease was longer in rET patients compared to ET patients, though this difference was not statistically significant (p = 0.09). All patients had bilateral action tremor; resting tremor was prominent on the right side in 20 rET patient and prominent on the left side in the remaining 10 rET patients (Table 1). All patients had tremor in both upper limbs, while no patient had tremor in the lower limbs. None of rET patients had overt parkinsonian symptoms (bradykinesia or rigidity), consistent with the normal DaTscan result in all patients. Regarding the Mini Mental State Examination (MMSE), both ET and rET patients had lower scores compared to HC, without differences between the two patient groups.

3.2. Differential tractography

3.2.1. Fractional anisotropy (FA)

ET versus HC. The ET group did not show any changes in FA compared to healthy controls (Fig. 1A).

Fig. 1.

Pairwise differences in fractional anisotropy (FA) among healthy controls (HC), Essential Tremor (ET) and Essential Tremor patients with resting tremor (rET). (A) HC vs. ET: Differences in FA between healthy controls and ET patients. (B) HC vs. rET: Differences in FA between healthy controls and rET patients. (C) ET vs. rET: Differences in FA between ET and rET patients. All results are statistically significant at FDR p < 0.05 and corrected for age and sex for ET-HC and rET-HC, and for age, sex and disease duration for rET-ET. The brain images are displayed in sagittal (left and right hemispheres), coronal and axial views, with the left (L) and right (R) sides labeled accordingly . Abbreviations: Post Thal Rad = Posterior Thalamic Radiation, Post CSTR = Posterior corticostriatal tract, Occ CP = Occipital Corticopontine tract.

rET versus HC. ET patients with resting tremor exhibited very few bundles where FA was reduced compared to healthy controls (Fig.1B in blue and Table S2), with an involvement of the projection pathways of the basal ganglia (left posterior thalamic radiation, left posterior corticostriatal fibers and right fornix), of the projection pathways of the brainstem (left occipital corticopontine tract), of the association pathways (left middle longitudinal fasciculus), and of the commissural pathways (forceps major of corpus callosum). Geometric characteristics (number of tracts and volume) of brain fibers where FA was significantly lower in rET patients compared to healthy controls are reported in Table 2. A focal increase of FA in rET patients compared to HC was found in the crus I of the right cerebellum (Table S3 and in red in Fig. 1B).

Table 2.

Geometric characteristics (number of tract and volume) of the fiber clusters where the Fractional Anisotropy (FA) was significantly reduced in rET patients compared with healthy controls (HC). Significantly different bundles between groups are clustered following the topology by Yeh at al. (Yeh, 2022).

FA contrast	rET < HC
	Tracts number	Total volume [mm3]
Association	11	237
Commissure	49	625
Projection Basal Ganglia	209	1,770
Projection Brainstem	9	255

All results were statistically significant at FDR p < 0.05 and corrected for age and sex.

ET versus rET. The comparison between ET and rET patients confirmed the increment of FA in rET patients in the right cerebellum, at the level of the crus I (Table S3 and in red in Fig. 1C).

3.2.2. Mean diffusivity (MD)

ET versus HC. ET patients exhibited higher MD values compared to healthy controls (in red in Fig. 2A and Table S1), showing widespread alterations in the brain more evident in the left hemisphere, including the projection pathways of basal ganglia and brainstem, the association pathways, the commissural pathways, and the cerebellum.

Fig. 2.

Pairwise differences in mean diffusivity (MD) among healthy controls (HC), Essential Tremor (ET) and Essential Tremor patients with resting tremor (rET). (A) HC vs. ET: Differences in MD between healthy controls and ET patients. (B) HC vs. rET: Differences in MD between healthy controls and rET patients. (C) ET vs. rET: Differences in MD between ET and rET patients. All results are statistically significant at FDR p < 0.05 and corrected for age and sex for ET-HC and rET-HC, and for age, sex and disease duration for rET-ET. The brain images are displayed in sagittal (left and right hemispheres), coronal and axial views, with the left (L) and right (R) sides labeled accordingly . Abbreviations: Post Thal Rad = Posterior Thalamic Radiation, Sup Thal Rad = Superior Thalamic Radiation, Parietal CP = parietal corticopontine, Post CSTR = Posterior corticostriatal tract, Hipp Alveus = Hippocampus Alveus

More in detail, Fig. 3 depicts the fibers related to the basal ganglia projection where MD was significantly higher in ET compared to HC (Fig. 3 top row in magenta). The altered WM tracts in the left cerebral hemispheres were the anterior, superior and posterior thalamic radiations, the optic radiation, posterior and superior corticostriatal tracts while fornix and superior thalamic radiation showed increase in both hemispheres. The changes were more extensive in the left cerebral hemisphere in comparison with the right side, as also shown by the geometric characteristics (number of tracts and volume) and the positive asymmetry index reported in Table 3.

Fig. 3.

Differences in mean diffusivity (MD) in the basal ganglia between healthy controls (HC) and Essential Tremor (ET) or Essential Tremor patients with resting tremor (rET). Top Row: Differences in MD between ET patients and healthy controls. Bottom Row: Differences in MD between rET patients and healthy controls. The brain images are similarly displayed in sagittal (left hemisphere) coronal and axial views, with the anterior (A) and posterior (P) sides labeled accordingly. All results are statistically significant at FDR p < 0.05 and corrected for age and sex. Abbreviations: Ant Thal rad = Anterior Thalamic Radiation, Sup Thal Rad = Superior Thalamic Radiation, Post Thal Rad = Posterior Thalamic Radiation, Opt Rad = Optic Radiation, Post CSTR = Posterior corticostriatal tract, Ant CSTR = Anterior corticostriatal tract

Table 3.

Geometric characteristics (number of tract and volume) of the fiber clusters where the Mean Diffusivity (MD) is significantly higher in ET and rET patients compared with healthy controls (HC). Significantly different bundles between groups are clustered following the topology by Yeh at al. (Yeh, 2022).

MD contrast	ET > HC		rET > HC
	Tracts number	Total volume [mm3]	Tracts number	Total volume [mm3]
Association	388	4,021	262	2,809
Cerebellum	635	5,035	1,582	6,482
Commissure	2,718	7,641	2268	7,053
Projection Basal Ganglia	3,666	16,034	5,715	13,958
Projection Brainstem	203	3,894	156	3,400
Whole brain changes	7,610	36,625	9,983	33,702
Projection Basal Ganglia L	2,875	13,012	2,449	6,247
Projection Basal Ganglia R	791	3,022	3,266	7,711
Projection Brainstem L	203	3,894	104	2,284
Projection Brainstem R	0	0	46	942
Cerebellum L	299	2,184	1,050	3,471
Cerebellum R	336	2,851	531	2,974
Asymmetry Index %
AI Projection Basal Ganglia	56.85	62.31	−14.30	−10.49
AI Projection Brainstem	100.00	100.00	38.67	41.60
AI Cerebellum	−5.83	−13.25	32.83	7.71

All results are statistically significant at FDR p < 0.05 and corrected for age and sex. Abbreviations: AI = Asymmetry index, where negative values mean rightward asymmetry and positive ones mean leftward asymmetry.

The fibers of the brainstem projection where MD was significantly higher in ET compared to HC are depicted in the top row of Fig. 4 (in magenta). All the alterations of WM tracts were exclusively located the left hemisphere (asymmetry index, Table 3), and were occipital and parietal corticopontine, corticospinal and the medial lemniscus bundles and the nondecussating dentato-rubro-thalamic tract (ndDRTT).

Fig. 4.

Differences in mean diffusivity (MD) in the projection brainstem between healthy controls (HC) and Essential Tremor (ET) or Essential Tremor patients with resting tremor (rET). Top Row: Differences in MD between ET patients and healthy controls. Bottom Row: Differences in MD between rET patients and healthy controls. The brain images are similarly displayed in sagittal (left hemisphere), coronal and axial views, with the anterior (A) and posterior (P) sides labeled accordingly. All results are statistically significant at FDR p < 0.05 and corrected for age and sex. Abbreviations: Med Lemniscus = Medial Lemniscus, CST = Corticospinal tract, Occ CP = Occipital Corticopontine tract, Par CP = Parietal Corticopontine tract, ndDRTT = non-decussating Dentato-rubro-thalamic tract, dDRTT RL = decussating Dentato-rubro-thalamic tract right-left

Fig. S3 in supplementary material depicts the fibers belonging to the association pathways (Fig. S3A) and to the commissural pathways (Fig. S3B), where MD was significantly higher in ET compared to HC (Fig. S3 top row in magenta). The WM tracts showing higher MD values in ET than in HC were the inferior frontal occipital and inferior longitudinal fasciculus, the cingulum, the arcuate fasciculus and the hippocampus alveus, more evident in the left cerebral hemisphere. The involved commissural pathways were the body, tapetum and forceps major of the corpus callosum and the anterior commissure occipital.

Finally, Fig. 5 depicts the fibers belonging to the cerebellum where MD was significantly higher in ET compared to HC (Fig. 5 top row in magenta). The altered parts of cerebellum included the bilateral crus I and the right crus II, with a small involvement of the right lobule VI.

Fig. 5.

Differences in mean diffusivity (MD) in cerebellum pathways between healthy controls (HC) and Essential Tremor (ET) or Essential Tremor patients with resting tremor (rET). Top Row: Differences in MD between ET patients and healthy controls. Bottom Row: Differences in MD between rET patients and healthy controls. All results are statistically significant at FDR p < 0.05 and corrected for age and sex. The brain images are similarly displayed in sagittal (left hemisphere), coronal and axial views, with the anterior (A) and posterior (P) sides labeled accordingly

rET versus HC. The rET group had significantly higher MD compared to the healthy controls in the brain especially in the projection pathways of basal ganglia and brainstem. Of note, some fiber bundles (posterior and superior thalamic radiation, posterior cortico-striatal tract, optic radiation were involved bilaterally, predominantly in the right side of the brain; occipital corticopontine tracts and medial lemniscus were involved bilaterally with prevalent changes in the left cerebral hemisphere. On the contrary, the MD of the anterior thalamic radiation and the dDRTT tract was increased only in the right side of the brain, while the ndDRTT, the corticospinal tract and the parietal corticopontine bundle were altered exclusively in the left hemisphere. (Fig. 3, Fig. 4; Table 3).

In the MD contrast rET > HC (Fig. S3A bottom row in green), the WM tracts in the cluster of association pathways were the right inferior longitudinal fasciculus, the left cingulum, the left arcuate fasciculus and the bilateral hippocampus alveus. The changes of commissural pathways were localized the body, tapetum and forceps major of the corpus callosum and the anterior commissure occipital (Fig. S3B bottom row in green).

Regarding the cerebellum, rET patients had higher MD than HC in the bilateral crus I (Fig. 5 bottom row in green) with prevalent fiber change in the left side (Table 3).

Overall, as shown in Table 3, when each patient group was compared with controls, fiber alterations in brainstem and basal ganglia projections was prevalent in the cerebral left hemisphere in ET (fiber change left: 3,078; right: 791) and in the right cerebral hemisphere in rET (fiber change left: 2,553; right: 3,312).

Moreover, our findings showed that total WM changes was slightly higher in ET (volume 36,625 mm³) than in rET (volume 33,702 mm³) and that ET patients had more pronounced changes on the left-side, while rET had more on the right-side (Table 3). Indeed, in the left cerebral hemisphere the WM volume of fiber alterations in brainstem and basal ganglia projections was 16,906 mm³ in ET and 8,531 mm³ in rET while in the right cerebral hemisphere WM volume of fiber changes was 3,022 mm³ in ET, and 8,653 mm³ in rET (Table 3). The different distribution of ET and rET between cerebral hemispheres was confirmed by asymmetry values (Table 3).

ET versus rET. The comparison between ET and rET patients revealed very few bundles where MD was higher in the rET cohort compared to ET group (Fig. 2C). In detail, small alterations were found in the projection pathways of basal ganglia (superior and posterior thalamic radiation, fornix and posterior corticostriatal tract) and of brainstem (parietal corticopontine tract), in the association pathways (left cingulum) (Table S2 and in red in Fig. 2C). On the contrary, higher MD was found in ET patients compared to rET patients in a small part of the tapetum of corpus callosum and of the left hippocampus alveus.

The most important differences between these two patient groups were detected in the cerebellum, where ET showed higher MD than rET in the bilateral rostral crus I and in the right crus II (cerebellum left volume changes 216 mm³, cerebellum right volume changes 1,251 mm³) (Table S2 and in blue in Fig. 2B). rET patients had higher MD than ET patients in bilateral caudal crus I (cerebellum left volume changes 969 mm³, cerebellum right volume changes 1,791 mm³) (Table S2 and in red in Fig. 2B).

3.2.3. Axial (AD) and radial diffusivity (RD)

Results about AD and RD in the three contrasts are reported in supplementary material (Fig. S1-S2). The mean of significantly different AD values in the three contrasts are reported in tables S4-S6. Changes of bundles had a very similar distribution to that observed with MD for all three contrasts among patient groups. In detail, WM bundles where AD and RD were higher in ET patients than HC were mainly localized in the left hemisphere, while in rET patients they were in right hemisphere (Fig. S1-S2).

No correlations were found between the diffusion metrics of significant fiber clusters in each group and the MMSE scores, suggesting that our findings were not driven by the slightly lower MMSE scores in ET patients.

4. Discussion

This study investigated the microstructural differences in white matter (WM) tracts in ET patients with and without resting tremor (rET) compared to healthy controls (HC) using advanced differential tractography. The results provide novel insights into the distinct neuroanatomical alterations associated with ET and rET, highlighting the different distribution of fiber changes in the two cerebral hemispheres between the two forms of essential tremor.

Our analysis indicated no significant changes in FA when comparing ET patients to healthy controls. On the contrary, in rET patients, there were reductions in FA in specific WM tracts compared to controls, including projection pathways of the basal ganglia, brainstem, and several association pathways. This reduction in FA suggests compromised WM integrity in these regions, aligning with more severe microstructural changes observed in rET (Caligiuri et al., 2017). Interestingly, rET patients exhibited an increase in FA in the right cerebellar crus I compared to both HC and ET patients. These findings imply that while overall WM integrity is compromised in rET, certain regions may undergo compensatory changes or reflect different pathological processes. The increase in FA in the right cerebellar crus I in rET patients may indicate localized microstructural reorganization (Andrews et al., 2024), but the interpretation of this result is not yet clear and needs further investigations.

Because the majority of cerebral white matter voxels contain multiple fiber populations and complex fiber geometries, FA, AD and RD measures become difficult to interpret while MD offers a more reliable and robust approach for drawing inferences about the underlying tissue characteristics (Figley et al., 2021, Maffei et al., 2022, Maier-Hein et al., 2017). Indeed, FA is useful for measuring the directional dependence of water diffusion, which can provide insights into fiber tract integrity and orientation, but its sensitivity to the relative volume fractions of crossing fibers and its dependency on the coherence of fiber orientation can lead to confounding results, particularly in regions with complex fiber geometries (Figley et al., 2021, Maffei et al., 2022, Maier-Hein et al., 2017). Due to these limitations, it is now generally suggested to avoid using FA, AD or RD to draw strong reverse inferences about WM microstructure (Figley et al., 2021, Genc et al., 2017). In contrast, MD measures the average rate of water diffusion within a voxel, thus it is theoretically more robust to the presence of multiple fiber populations and complex fiber geometries, making it a superior indicator of how tissues constrain diffusion within a voxel (Figley et al., 2021). This is further supported by empirical evidence indicating that MD has a stronger correlation with neurite density compared to FA, AD, or RD measures (Genc et al., 2017).

In our study, differently from FA results, significant and widespread differences in MD were found across several brain regions, especially when each patient group was compared with HC. ET patients exhibited increased MD in various tracts, including the basal ganglia, brainstem projections, and cerebellum, compared to HC. These alterations, which were prevalent in the left side of the brain, suggested a potential loss of WM integrity possibly due to demyelination or axonal degeneration (Figley et al., 2021). Similar but more pronounced bilaterally changes of WM fiber bundles with prevalence on the right side of the brain were observed in rET patients compared to controls. The different cerebral lateralization of WM changes in our ET and rET patients, especially in the basal ganglia and brainstem tracts, underscores the potential involvement of asymmetric pathophysiological mechanisms in these tremor disorders. Small differences in the MD of white matter (Table S3) however, was observed when ET and rET patient groups were compared to each other. This finding agrees with previous studies (Novellino et al., 2016, Prasad et al., 2018) and supports the hypothesis that both patient groups with similar extent and pattern of brain WM lesions have the same disease.

We also demonstrated significant changes in the cerebellar pathways in both ET and rET cohorts, with ET patients showing alterations predominantly in the cerebellum (crus I and II) while rET patients exhibited changes only in crus I. This finding aligns with previous studies in ET demonstrating higher MD values in cerebellum WM than in controls (Prasad, 2018), suggesting the cerebellum's crucial role in ET pathology (Andrews et al., 2024, Novellino et al., 2016, Saini et al., 2012, Sarica et al., 2022), particularly in motor control and tremor generation (Prodoehl et al., 2013). The involvement of Crus II in ET confirms our previous data demonstrating the alterations of this cerebellar region in essential tremor (Sarica et al., 2022), a finding which has been confirmed in a very recent work that explored the WM microstructure of essential tremor following DBS (Andrews et al., 2024). Moreover, our findings agree with Prodoehl et al. (Prodoehl et al., 2013) who highlighted the importance of diffusion changes of basal ganglia and cerebellar pathways in distinguishing subjects with Parkinson's disease, atypical parkinsonism, and essential tremor from control subjects.

Previous DTI studies in ET demonstrated an involvement of the cerebello-thalamo-cortical network (CTC) but also of areas outside the tremor network (Klein et al., 2011, Saini et al., 2012).

It is noteworthy that our study demonstrates the involvement of the dentato-rubro-thalamic tract (DRTT), a bundle connecting the cerebellum to the thalamus, which fibers were altered in both ET and rET patients. DRTT has been shown to mediate tremor-reducing DBS effects in essential tremor (Andrews et al., 2024, Dembek et al., 2020, Deuter et al., 2024, Yang et al., 2020) and in other tremors irrespective of underlying pathology (Deuter et al., 2024, Deuter et al., 2023). The DRTT is classically described as a decussating pathway, but the existence of a nondecussating ipsilateral DRTT (ndDRTT) has been demonstrated (Meola et al., 2016). Some studies demonstrated that the decussating pathway ascends to the contralateral thalamus and connects to posterior and medial nuclei where it synapses with neurons ascending to the cortex (Deuter et al., 2023, Petersen et al., 2018) while nondecussating bundle reaches ipsilateral thalamus preferentially targeting anterior and lateral thalamic nuclei (Deuter et al., 2023, Petersen et al., 2018). Recent evidence suggests that DBS may exerts effect on action tremor reduction in ET acting on both fiber bundles with the crossing part showing stronger correlation with good clinical response (Deuter et al., 2023, Duanmu et al., 2023). Very recently some authors (Deuter et al., 2024) demonstrated that dDRTT may play a major role in the mechanism of alleviation of parkinsonian tremor and could serve as a possible DBS target for tremor dominant PD. On the other hand, recent studies investigating the effectiveness of MR-guided focused ultrasound (MRgFUS) in alleviating tremor reported that clinical effect was related to lesion overlap with the posterior DRTT (Kapadia et al., 2020, Zur et al., 2020). This evidence seems somewhat conflicting, with DBS studies suggesting that DRTT stimulation may alleviate tremor and MRgFUS studies suggesting that damaging this bundle may have the same effect. Overall, these previous finding together with our results suggest that the DRTT is a key region involved in tremor generation in ET patients, which shows DTI alterations and may benefit either from stimulation or focused damage possibly interrupting a dysfunctioning circuit. However, further research is needed to better understand the role of DRTT fibers in tremor generation. In the current study, we found MD changes in both DRTT tracts in rET while only left nd-DRTT was involved in ET. Of note, in our rET patients, the alterations of the right-to-left dDRTT bundle were consistent with the resting tremor which was prevalent in the right side of body in most cases (Table 1). Thus, these findings suggest that the involvement of both DRTT bundles may have a pivotal role in ET, and that the right crossing DRTT might be involved in resting tremor associated with rET, suggesting this bundle as a possible DBS target for treatment of rET patients.

The use of advanced differential tractography (Maffei et al., 2022, Maier-Hein et al., 2017) in our study was pivotal in elucidating the microstructural changes in WM tracts of ET patients, allowing for the detection of subtle differences that might be overlooked by conventional methods (Yeh et al., 2016, Yeh et al., 2019b, Yeh et al., 2013). Indeed, conventional tractography analysis tracks connections between different brain regions, which may render tract-specific analysis inconclusive (Huang et al., 2023, Li et al., 2022). The innovation of differential tractography lies in its adoption of the “tracking-the-difference” paradigm, in contrast to the “tracking-the-existence” approach used in traditional methods. Moreover, compared to traditional tractography, differential tractography more accurately reflects the structure and density of white matter tracts, considering crossing fibers and partial volume effects (Maffei et al., 2022, Maier-Hein et al., 2017). Moreover, by incorporating topology-informed pruning (Yeh et al., 2019b), it reduces false positives and focuses on anatomically and functionally relevant tracts, offering detailed connectivity analysis and enabling a more nuanced understanding of connectivity changes associated with essential tremor.

Taken together, our findings demonstrates that these two forms of tremor had a similar extent and pattern of WM brain changes and differed from each other especially in the cerebral lateralization of fiber changes, suggesting that ET and rET may be different imaging subtypes of the same disease rather than two different diseases.

There are several considerations to support this hypothesis. First, the strong similarity of brain WM fiber pathways changes between ET and rET found in the current study argues against the hypothesis that these forms of tremor are two different diseases. Both forms of tremor showed a similar pattern of altered pathways both at cerebral and cerebellar level, with the most important difference being in the lateralization of WM change in the brain which was prevalent in the left cerebral hemisphere in ET and in the right hemisphere in rET. Second, pathological studies also supported this hypothesis showing similar alterations in both ET and rET patients (Louis et al., 2011, Rajput et al., 2004). A study (Gionco et al., 2021) performed in 50 ET cases (24 ET and 26 ET-plus) confirmed that there were no pathological differences in cerebellar cortex between ET and ET-plus cases, supporting the notion that ET and ET-plus represent a single entity with heterogeneous clinical signs.

Another intriguing hypothesis is that patients with ET accumulate additional clinical features during their illness thus developing rest tremor in the late stage of the disease. In keeping with this hypothesis, patients with rET usually have longer disease duration and greater tremor severity than ET patient without rest tremor. Against the hypothesis that in rET patients rest tremor occurs as a late manifestation of the disease, are the current findings demonstrating that the total volume of the WM brain changes detected by differential tractography was very similar in both ET and rET (36.625 vs 33.702 mm³, respectively), despite rET patients had a longer disease duration. Moreover, although the WM changes involved very similar fiber tracts in both tremors, the cerebral distribution of lesions was different, with ET having a higher extent of alterations on the left side of the brain while rET had higher volume and fiber change bilaterally with a mild prevalence in the right cerebral hemisphere, where there was a change of the crossing DRTT, a bundle potentially involved in the development of rest tremor. Overall, all this evidence supports the hypothesis that these two forms of tremor are two distinct imaging subtypes of essential tremor rather than different stages of the same disease. Further studies to better elucidate this point are warranted.

Finally, the hypothesis that rET may represent a comorbidity with PD has been suggested by several authors although several works did not find DaTscan abnormalities nor clinical signs suggestive of PD such as hyposmia (Louis and Jurewicz, 2003) and micrographia (Martinez‐Hernandez and Louis, 2014). In addition, some studies demonstrated that rET patients showed a rest tremor with electrophysiologic characteristics (synchronous tremor) different from those detected in PD (Nisticò et al., 2011, Vescio et al., 2021), and that this tremor might raise from the involvement of cerebral circuits different from those involved in PD (Caligiuri et al., 2017, Li et al., 2021, Nicoletti et al., 2015). Furthermore, the most important pathological studies in rET patients (Louis et al., 2011, Rajput et al., 2004) failed to demonstrate Lewy body pathology thus excluding comorbidity with PD.

This study has several limitations that need to be acknowledged. Firstly, the small sample size, though sufficient for detecting significant differences, may limit the generalizability of the findings. Future studies with larger cohorts are needed to validate these results and to ensure the robustness of the observed microstructural differences in WM tracts in ET and rET patients. Second, standardized, detailed quantitative clinical data on rest, postural and kinetic tremor severity were not available in this patient cohort and thus correlation analyses between imaging and clinical data were not performed. Future studies to investigate this point are needed. Third, the cross-sectional design of the study does not allow for the assessment of the progression of WM changes over time. Longitudinal studies are essential to elucidate the temporal dynamics of WM alterations and their relationship with tremor severity and other clinical features. Additionally, while advanced differential tractography was utilized to enhance the sensitivity of detecting microstructural differences, this technique is not without its limitations. The potential for false positives, despite topology-informed pruning, cannot be completely ruled out. Furthermore, the study did not incorporate functional MRI or other advanced imaging techniques that could provide complementary insights into the functional implications of the observed WM structural changes. Future research should consider integrating multiple imaging modalities to provide a more comprehensive understanding of the neurobiological underpinnings of ET and rET.

5. Conclusion

In conclusion, this study revealed similar WM pathway alterations in ET and rET, highlighting the different involvement of the cerebral hemispheres in these two forms of essential tremor, suggesting that these two tremor disorders may be distinct imaging subtypes of the same disease rather than distinct entities. Advanced differential tractography identified significant MD increases, suggesting WM integrity loss due to demyelination or axonal degeneration.

These findings enhance our understanding of the neurobiological differences between ET and rET, providing insights that could inform targeted diagnostic and therapeutic strategies. Specifically, the distinct patterns of white matter alterations in the dentato-rubro-thalamic tracts (DRTT) and other pathways suggest that these imaging subtypes might benefit from differential therapeutic approaches. Indeed, our results might influence the design of interventions such as focused ultrasound thalamotomy or neuromodulation techniques, where the choice of hemisphere or tract-specific targets could enhance treatment efficacy. Future research integrating imaging biomarkers with clinical outcomes could help refine these strategies further.

6. Statement of ethics

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study protocol was approved by the Institutional Review Board of Magna Graecia University in Catanzaro, Italy. Written informed consent was obtained from all participants prior to their inclusion in the study.

Funding support

Partial financial support was received from “MNESYS - A multiscale integrated approach to the study of the nervous system in health and disease” project (F63C22000640002 - DECRETO 1553.11-10-2022 PNRR MUR, PE0000006). The funder played no role in study design, data collection, analysis and interpretation of data, or writing the manuscript.

CRediT authorship contribution statement

Alessia Sarica: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Vera Gramigna: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Fulvia Arcuri: Formal analysis. Marianna Crasà: Data curation. Camilla Calomino: Data curation. Rita Nisticò: Data curation. Maria Giovanna Bianco: Data curation. Andrea Quattrone: Writing – review & editing, Supervision, Conceptualization. Aldo Quattrone: Writing – review & editing, Supervision, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary data to this article:

Data availability

The data that has been used is confidential.