Progressive iron deposition and widespread neural dysfunction in Parkinson’s disease: a multimodal MRI study

Introduction

Parkinson’s disease (PD) is the second most common neurodegenerative disorder worldwide, with its epidemiological characteristics closely related to the aging population (1). The hallmark clinical features of PD include progressively worsening motor impairments, such as resting tremor, rigidity, and bradykinesia, as well as non-motor symptoms including cognitive dysfunction, sleep disturbances, and autonomic dysfunction (2). Approximately 5–10% of PD cases are associated with specific gene mutations, such as those in LRRK2 and GBA; however, the pathophysiological mechanisms underlying the more common sporadic form of the disease remain incompletely understood (3,4). Current research suggests that the pathological mechanisms of PD may involve several factors, including abnormal aggregation of α-synuclein leading to protein misfolding, mitochondrial dysfunction, oxidative stress, and disrupted brain iron metabolism (5-7). Notably, imaging studies have shown that abnormal iron deposition in the basal ganglia, particularly in regions like the substantia nigra (SN) and globus pallidus (GP), is closely associated with the progressive loss of dopaminergic neurons (8). However, the dynamic evolution of iron metabolism abnormalities from the early to late stages of the disease, and their specific role in neurodegeneration, remain areas requiring further investigation.

Iron metabolism dysregulation is considered one of the major pathological mechanisms contributing to neurodegenerative diseases (9). In the case of PD, numerous studies have demonstrated that abnormally high iron deposition in specific brain regions is a potential biological marker for tracking disease progression (10,11). Quantitative susceptibility mapping (QSM) is a highly sensitive imaging technique capable of detecting iron deposition in brain tissues of PD patients (12,13). Fu et al. showed that the magnetic susceptibility values of the SN, red nucleus (RN), and GP in PD patients were higher than those of healthy controls, with the magnetic susceptibility values of the SN showing a positive correlation with the Hoehn and Yahr (H&Y) stage of PD (14). Thomas et al. found, through a 3-year longitudinal study, that iron deposition in the right temporal cortex, Meynert’s basal nucleus, and the putamen was associated with cognitive decline, while iron deposition in the basal ganglia, SN, RN, insular cortex, and dentate nucleus correlated with worsening motor symptoms (15).

Resting-state functional magnetic resonance imaging (rs-fMRI) explores spontaneous fluctuations in blood oxygen level-dependent (BOLD) signals to investigate neuronal activity in the resting state, providing reliable measurements of baseline brain activity (16). rs-fMRI has been extensively used to study brain activity and connectivity changes in PD (17). The fractional amplitude of low-frequency fluctuations (fALFF), which is the ratio of low-frequency power to the total power spectrum across the entire frequency range, exhibits higher sensitivity and specificity in detecting spontaneous brain activity (18). Hou et al. demonstrated that fALFF values in specific brain regions, such as the caudate nucleus and precuneus, can predict the progression of motor symptoms in newly diagnosed PD patients who are unresponsive to medication, with high classification accuracy (19). Furthermore, cognitive impairment in PD has been associated with abnormal spontaneous brain activity, a relationship confirmed through fALFF analysis. Chen et al. demonstrated that there are significant fALFF differences in the parahippocampal gyrus between cognitively intact PD patients and those with mild cognitive impairment (MCI), suggesting the potential of this region as an early biomarker for PD (20).

Regional homogeneity (ReHo) is a reliable indicator of local functional synchronization of spontaneous neuronal activity (21). Previous studies have shown that changes in ReHo in PD are associated with the motor cortex, visual-related cortical areas, and the default mode network (DMN) (22,23). Nguyen et al. used machine learning techniques to predict longitudinal disease severity based on ReHo and fALFF, with an R² of 55.8% in explaining the variance of Movement Disorder Society-Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) scores after 1 year of follow-up. Their results indicated that the DMN and motor-related regions are key predictive factors, suggesting that functional reorganization precedes clinical deterioration (24). These findings highlight the value of multimodal MRI techniques, such as integrating ReHo, fALFF, and QSM, to capture complementary aspects of neural dysfunction across different stages of PD.

In PD, pathophysiological mechanisms change at different disease stages and the H&Y score provides a reliable means of charting the clinical progression of the disorder (25). Given the significance of integrating ReHo, fALFF, and QSM as multimodal neuroimaging biomarkers, we hypothesized that there would be distinct patterns in these measures between PD patients at early stage and advanced stage, as classified by the H&Y scale, as well as differences between these PD groups and healthy controls (HC). Therefore, ReHo, fALFF, and QSM were compared across the three groups, and their associations with motor and non-motor symptoms in early-stage PD (ESP) and advanced-stage PD (ASP) were investigated. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2070/rc).

Methods

Participants

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Research Ethics Committee (REC) of Henan Provincial People’s Hospital (HPPH) (No. 2018-064-02) and informed consent was obtained from all individual participants. All procedures strictly adhered to international standards for participant screening and evaluation. Patients were consecutively recruited from the outpatient clinic of the Neurology Department at HPPH between January 2019 and December 2020. The diagnosis of PD was established independently by two senior neurologists according to the 2015 International Parkinson and Movement Disorder Society (MDS) clinical diagnostic criteria (26), with final confirmation based on consensus.

Exclusion criteria were as follows: (I) secondary Parkinsonian syndromes (e.g., drug-induced or vascular) or atypical Parkinsonian-plus syndromes (e.g., progressive supranuclear palsy, multiple system atrophy); (II) contraindications to MRI, including claustrophobia, cardiac pacemakers, or other metallic implants; (III) other central nervous system disorders, such as cerebrovascular disease (infarction or hemorrhage) or intracranial space-occupying lesions (e.g., brain tumors); and (IV) Mini-Mental State Examination (MMSE) scores below the education-adjusted thresholds (<17 for illiterate participants, <20 for those with 1–6 years of education, and <23 for those with ≥7 years of education) (27). In addition, 53 age- and sex-matched HC participants were recruited from the community.

The same two neurologists also conducted comprehensive clinical evaluations for all participants, including: (I) disease severity assessed with the H&Y scale; (II) motor symptoms evaluated using the MDS-UPDRS (28); and (III) non-motor symptoms assessed with the MMSE (cognitive function), Hamilton Anxiety Scale (HAMA) (29), and Hamilton Depression Scale (HAMD) (30). Of the 116 consecutively recruited patients with PD, 12 were excluded due to poor-quality MRI data (mainly motion artifacts) or comorbid cerebrovascular pathology on structural MRI. Based on H&Y staging, PD patients were subdivided into two groups: 49 patients with H&Y stage <3 were classified as ESP, and 55 patients with H&Y stage ≥3 were classified as ASP. All PD patients underwent MRI scanning in the practical ‘on’ medication state to ensure patient comfort and minimize head motion during the acquisition.

Data acquisition

MRI was performed on a 3.0 T MRI system (MAGNETOM Prisma, Siemens Healthcare, Erlangen, Germany) equipped with a 64-channel head and neck coil. Participants were instructed to keep their eyes closed, remain awake, and avoid directed thought, while head motion was minimized using foam padding. First, a 3D T1-weighted magnetization-prepared rapid acquisition gradient-echo (MPRAGE) sequence was acquired with the following parameters: repetition time (TR) =2,300 ms, echo time (TE) =2.28 ms, inversion time (TI) =900 ms, receiver bandwidth =200 Hz/Px, flip angle (FA) =8°, number of slices =192, slice thickness =1 mm, field of view (FOV) =260×260 mm², matrix size =256×256, yielding an isotropic voxel size of 1.0×1.0×1.0 mm³. The acquisition time was 5 min 21 s. Second, QSM images were generated using the STrategically Acquired Gradient Echo (STAGE) platform (SpinTechMRI, Bingham Farms, MI, USA). Two 3D gradient-echo sequences were acquired with FA =6°/27°, TR =25 ms, TE1 =8.5 ms, TE2 =8.75 ms, slice thickness =2 mm, matrix size =256×256, and a total scan time of 308 s. Finally, rs-fMRI data were collected using a gradient-echo echo-planar imaging (GRE-EPI) sequence, acquiring 180 volumes with the following parameters: TR =2,000 ms, TE =35 ms, FA =80°, number of slices =75, slice thickness =2.2 mm, FOV =207×207 mm², matrix size =94×94, corresponding to a voxel size of 2.2×2.2×2.2 mm³.

rs-fMRI data preprocessing

The rs-fMRI data were preprocessed using the Resting-State fMRI Data Analysis Toolkit (RESTplus) (31). The pipeline included the following steps: (I) conversion of DICOM to NIfTI format; (II) removal of the first 10 time points; (III) slice-timing correction using the middle slice as the reference; (IV) realignment was performed to correct for head motion. Head motion was controlled by censoring individual volumes with FD >0.2 mm and excluding participants whose mean FD exceeded 0.2 mm; (V) co-registration of each participant’s functional images to their high-resolution T1-weighted MPRAGE image was performed, followed by spatial normalization to the Montreal Neurological Institute (MNI) space using the DARTEL algorithm, with resampling to 3×3×3 mm³ voxel size; (VI) spatial smoothing using a 6-mm full-width at half-maximum (FWHM) Gaussian kernel; (VII) linear detrending; (VIII) nuisance covariate regression was performed to remove the 24 head motion parameters (Friston 24-parameter model), white matter signal, and cerebrospinal fluid signal. Global signal regression was not performed to preserve group differences in overall signal amplitude; and (IX) band-pass filtering (0.01–0.08 Hz). Importantly, non-filtered data were used for fALFF calculations, and spatial smoothing was applied after metric computation in the case of ReHo. Unless otherwise specified, the default settings of the RESTplus toolkit were used.

STAGE data processing

The raw STAGE data were imported into the STAGE software (version 3.2.5, SpinTech, Bingham Farms, MI, USA) for automated post-processing. First, skull signals were removed, and susceptibility-weighted images were calculated using the conventional phase mask approach. Both SWI and T2 maps were automatically generated (32-34). Prior to QSM reconstruction, the first echo was phase unwrapped and this was used in a step-by-step fashion to unwrap all echoes. R2* relaxation maps were employed to selectively mask gray matter nuclei and venous vessels within the basal ganglia region. QSM maps were also automatically reconstructed within the STAGE software pipeline, along with the automated generation of SWI and T2 maps. Phase unwrapping was performed using the STAGE-implemented multi-echo Laplacian-based method, in which the first echo served as the phase reference for sequential unwrapping across remaining echoes (35-39). Regions of interest (ROIs) for QSM quantification, including the SN and GP, were automatically segmented using the atlas-based deep-gray-matter parcellation integrated within the STAGE software. This method employs an MNI-space template incorporating prior anatomical knowledge to ensure reproducible, operator-independent delineation of subcortical nuclei, consistent with established QSM workflows in PD research (40,41).

rs-fMRI analysis

ReHo analysis was conducted on unsmoothed, filtered data using the RESTplus software. The local temporal synchronization of each voxel was quantified by calculating Kendall’s coefficient of concordance (KCC) between the time series of the voxel and those of its 26 nearest neighbors, forming a 27-voxel cube (21). Individual ReHo maps were generated by computing the KCC for each voxel, and then standardized by dividing each voxel’s KCC value by the global mean KCC across the whole brain. Finally, spatial smoothing was applied with a Gaussian kernel (FWHM =6×6×6 mm³) to enhance the signal-to-noise ratio of the ReHo maps.

For fALFF computation, the time series of each voxel was first band-pass filtered (0.01–0.08 Hz). A Fourier transform was then performed to obtain the power spectrum, from which the amplitude of low-frequency fluctuation (ALFF) was calculated as the square root of the power within the 0.01–0.08 Hz band (42). The fALFF was derived as the ratio of ALFF to the total power across the full frequency range (0–0.25 Hz), thereby reducing the influence of physiological noise (18). Similar to ReHo, the resulting fALFF maps were spatially smoothed with a 6-mm FWHM Gaussian kernel.

Both ReHo and fALFF maps were subsequently co-registered to the MNI standard space for group-level comparisons.

Statistical analysis

Clinical and demographic data for PD patients and HCs were analyzed using SPSS software (version 25.0; IBM Corp., Armonk, NY, USA). All continuous variables—including age, disease duration, clinical scale scores, and QSM values for the nigrostriatal ROIs—were expressed as mean ± standard deviation, and normality was assessed using the Shapiro-Wilk test. For normally distributed data, independent-samples t-tests were applied for pairwise group comparisons, while one-way analysis of variance (ANOVA) was used for comparisons across multiple groups, with Bonferroni correction applied for multiple comparisons. For non-normally distributed data, the Mann-Whitney U test or Kruskal-Wallis test was employed. For voxel-wise group comparisons of ReHo and fALFF maps, statistical significance was assessed using cluster-level false discovery rate (FDR) correction (P<0.05) to control for multiple comparisons across the entire brain.

To examine associations between imaging measures and clinical outcomes, mean values were extracted from brain regions showing significant group differences in the primary FDR-corrected voxel-wise analyses. ANOVA was first performed to compare differences across H&Y stages, followed by post hoc two-sample t-tests. For QSM analyses, age and sex were included as covariates in all between-group comparisons. Associations between QSM measures and clinical scores were examined using Pearson correlation, implemented as partial correlations controlling for age and sex. For rs-fMRI analyses, voxel-wise group comparisons of ReHo and fALFF were conducted using a general linear model, including age, sex, years of education, and mean FD as covariates to control for demographic and motion-related effects. Associations between rs-fMRI metrics and clinical measures were assessed using Pearson correlation, implemented as partial correlations adjusting for age, sex, years of education, and FD. The covariate structure was modality-specific: QSM analyses adjusted for age and sex, while rs-fMRI analyses additionally controlled for education level and FD. Supplementary partial correlation analyses were performed for all reported Pearson correlations to verify robustness, with consistent results provided in Figures S1-S4.

Results

Comparison of clinical and demographic characteristics

Table 1 summarizes the clinical and demographic characteristics of the 104 patients with PD—classified into ESP (n=49) and ASP (n=55) groups based on H&Y scores—together with 53 HCs. As expected, the HC group did not undergo UPDRS assessments. Significant differences were observed in UPDRS-III scores between the ESP and ASP groups, reflecting the clinical progression of motor impairment. Evaluations of anxiety and depression, assessed using the HAMA and HAMD scales, showed significantly higher scores in both PD groups compared with HCs; however, no significant differences were found between ESP and ASP. Cognitive performance, measured by the MMSE, differed significantly across all three groups, with progressive decline from HCs to ESP and ASP patients.

Analysis of SN and GP QSM values

Values of QSM for the SN and GP ROIs in the ESP, ASP, and HC groups are presented in Table 2 and Figure 1. ANOVA demonstrated significant group differences in QSM values across the three cohorts, which remained significant after Bonferroni correction for multiple comparisons. Specifically, bilateral iron concentrations in the SN and GP were significantly elevated in both ESP and ASP patients relative to HCs, and further increased in the ASP group compared with the ESP group. Furthermore, Pearson correlation analyses revealed robust associations between iron deposition and motor symptom severity. Within the PD cohort, QSM values in the bilateral SN and GP showed significant positive correlations with UPDRS-III scores, indicating that greater iron accumulation was closely linked to more severe motor impairment (Figure 2).

Figure 1 Comparison of QSM values in the SN and GP among the study groups. *, P<0.05; **, P<0.01; ***, P<0.001. ASP, advanced-stage Parkinson’s disease; ESP, early-stage Parkinson’s disease; GP, globus pallidus; HC, healthy control; LGP, left globus pallidus; LSN, left substantia nigra; RGP, right globus pallidus; QSM, quantitative susceptibility mapping; RSN, right substantia nigra; SN, substantia nigra.

Figure 2 Correlation between QSM values in the SN and GP and UPDRS-III scores in all PD patients. Each dot represents one patient, with the regression line indicating the linear fit. Correlation coefficients (r) and P values are displayed on the plots. GP, globus pallidus; LGP, left globus pallidus; LSN, left substantia nigra; PD, Parkinson’s disease; QSM, quantitative susceptibility mapping; RGP, right globus pallidus; RSN, right substantia nigra; SN, substantia nigra; UPDRS-III, Unified Parkinson’s Disease Rating Scale part III.

rs-fMRI analysis

Compared with HCs, patients with PD exhibited widespread abnormalities in local neural activity, as measured by fALFF and ReHo, across multiple brain regions. Figure 3 illustrates regions showing significant group differences in fALFF among ESP, ASP, and HC participants, while Figure 4 depicts corresponding differences based on ReHo. These alterations were distributed across the visual network, basal ganglia-thalamocortical circuits, prefrontal-parietal regions, and the limbic system. In ESP, fALFF and ReHo abnormalities were predominantly localized to visual-associated cortices, including the bilateral lingual gyri (Lingual_L/R), calcarine cortices (Calcarine_L/R), cuneus (Cuneus_L/R), and the right fusiform gyrus (Fusiform_R). With disease progression, these aberrant patterns expanded to involve subcortical nuclei (bilateral putamen, caudate, and thalamus), sensorimotor cortices (precentral and postcentral gyri), and limbic structures such as the anterior cingulate cortex (Cingulum_Ant), insula, and hippocampus.

Figure 3 Brain regions showing significant group differences in fALFF among the three groups. Group differences in fALFF across HC, ESP, and ASP were assessed using a one-way ANOVA to identify voxels with a significant main effect of group. Post-hoc pairwise t-tests (ESP vs. HC, ASP vs. HC, ASP vs. ESP) were conducted within the significant clusters identified by the ANOVA, with cluster-level FDR correction at P<0.05. The resulting statistical maps are shown in (B-D). Color scales are consistent across panels, with color bars indicating the corresponding t value ranges. Warmer colors indicate increased fALFF, while cooler colors indicate decreased fALFF. (A) Comparison of fALFF difference brain regions among HC, ESP, and ASP groups. (B) Comparison between ESP and HC. (C) Comparison of ASP and HC. (D) Comparison of ASP and ESP. ANOVA, analysis of variance; ASP, advanced-stage Parkinson’s disease; ESP, early-stage Parkinson’s disease; fALFF, fractional amplitude of low-frequency fluctuation; FDR, false discovery rate; HC, healthy control.

Figure 4 Brain regions showing significant group differences in ReHo among the three groups. Group differences in ReHo across HC, ESP, and ASP were assessed using a one-way ANOVA to identify voxels with a significant main effect of group. Post-hoc pairwise t-tests (ESP vs. HC, ASP vs. HC, ASP vs. ESP) were conducted within the significant clusters identified by the ANOVA, with cluster-level FDR correction at P<0.05. The resulting statistical maps are shown in (B-D). Color scales are consistent across panels, with color bars indicating the corresponding t value ranges. Warmer colors indicate increased ReHo, while cooler colors indicate decreased ReHo. (A) Comparison of ReHo difference brain regions among HC, ESP, and ASP groups. (B) Comparison between ESP and HC. (C) Comparison of ASP and HC. (D) Comparison of ASP and ESP. ANOVA, analysis of variance; ASP, advanced-stage Parkinson’s disease; ESP, early-stage Parkinson’s disease; FDR, false discovery rate; HC, healthy control; ReHo, regional homogeneity.

As summarized in Table 3, consistent intergroup differences across both fALFF and ReHo were observed in several regions, suggesting that these areas may constitute core hubs of stable functional impairment in PD. In contrast, Table 4 shows that certain abnormalities were identified by only one of the two metrics, possibly reflecting more localized or subtle disruptions in neural activity. Furthermore, correlation analyses revealed that fALFF and ReHo values in aberrant regions were significantly associated with UPDRS-III scores in PD patients (Figure 5). In addition, significant positive correlations were observed between MMSE scores and ReHo values in both the left and right hippocampus in patients with PD, as shown in Figure 6.

Figure 5 Partial correlations between resting-state fMRI metrics (fALFF and ReHo) and motor severity (UPDRS-III) in patients with PD (n=104). (A) fALFF values in the putamen (bilateral), caudate nucleus (right), and thalamus (right) show significant negative correlations with UPDRS-III scores. (B) ReHo values in the putamen, caudate nucleus, pallidum and thalamus show significant negative correlations with UPDRS-III scores. Each dot represents one patient; solid lines indicate linear regression fits, and shaded bands denote 95% confidence intervals. fALFF, fractional amplitude of low-frequency fluctuation; fMRI, functional magnetic resonance imaging; PD, Parkinson’s disease; ReHo, regional homogeneity; UPDRS-III, Unified Parkinson’s Disease Rating Scale part III.

Figure 6 Correlation between resting-state fMRI metric (ReHo) and cognitive performance (MMSE) in patients with PD (n=104). (A) Scatterplot showing a significant positive correlation between the ReHo value in the Hippocampus_R and MMSE scores. (B) Scatterplot showing a significant positive correlation between the ReHo value in the Hippocampus_L and MMSE scores. Each dot represents one patient; solid lines indicate linear regression fits, and shaded bands denote 95% confidence intervals. fMRI, functional magnetic resonance imaging; MMSE, Mini-Mental State Examination; PD, Parkinson’s disease; ReHo, regional homogeneity.

Discussion

This study integrates multiple MRI modalities, including QSM, ReHo, and fALFF, to systematically reveal the dynamic evolution of brain iron deposition and neuronal activity during the progression of PD. Importantly, combining iron-sensitive and functional metrics enabled us to relate subcortical iron burden to distributed network dysfunction, providing mechanistic insights that single-modality analyses cannot easily capture. Whereas fALFF reflects the magnitude of spontaneous activity and ReHo reflects local synchrony, the divergence between the two suggested dissociable disruptions in neural intensity versus coordination, supporting their complementary value in multimodal interpretation. Our findings indicate that iron deposition in the SN and GP significantly increases as the disease progresses and correlates positively with the UPDRS-III score. Additionally, abnormal neuronal activity, initially confined to visual-related brain regions such as the lingual gyrus and calcarine cortex, gradually expands to involve the basal ganglia-thalamocortical motor loop, the prefrontal-parietal network, and the limbic system structures in the later stages of the disease. Notably, in regions such as the putamen, caudate nucleus, and thalamus, fALFF and ReHo values are negatively correlated with the UPDRS-III score. The spatial distribution of functional abnormalities across disease stages shows a pattern compatible with the caudal-to-cranial progression described in Braak’s neuropathological staging; however, this cross-sectional design cannot establish temporal progression, and longitudinal studies are needed to confirm this sequence in vivo (43). Taken together, these cross-modality patterns suggest that iron accumulation and functional disruption may evolve as partially coupled processes: iron dysregulation may amplify susceptibility in dopaminergic structures, whereas network-level dysfunction may arise in parallel through α-synuclein propagation or long-range circuit effects. Although causality cannot be inferred, this multimodal convergence strengthens the mechanistic interpretation beyond what either modality could provide alone. Importantly, the joint use of QSM and rs-fMRI explained a larger proportion of UPDRS-III variance than either modality individually, indicating added clinical explanatory value of the multimodal framework.

In the early stages of the disease, neuronal activity abnormalities primarily affect the visual network. As PD progresses, these abnormalities gradually involve multiple brain systems, including the motor network, cognitive control network, and limbic system, which is better characterized as a caudal-to-cranial progression, starting in the locus coeruleus and brainstem and gradually advancing cranially, consistent with Braak’s staging model, rather than a strictly posterior-to-anterior or external-to-internal spread. Although longitudinal validation is required, the progressive increase in iron deposition across disease stages suggests its biological contribution to PD pathology. Excess iron may promote neurodegeneration through mitochondrial impairment and oxidative stress (44,45). Nevertheless, our cross-sectional design cannot determine whether iron dysregulation precedes or follows neurodegeneration (46), and dopamine metabolites may further intensify toxicity in high-iron environments (47). However, whether iron dysregulation is an upstream trigger or a downstream consequence of synucleinopathy remains unresolved. Early-stage iron dysregulation may signal impaired cellular buffering, while progressive accumulation could later directly drive degeneration via ferroptosis. The spatial divergence between focal iron deposition and widespread functional abnormalities further suggests that these processes may evolve in parallel, rather than in a strictly linear sequence. The ESP-to-ASP increase in QSM may partially reflect impaired iron clearance. Glymphatic dysfunction and BBB leakage—both reported in PD—can reduce perivascular efflux, creating a feed-forward cycle in which impaired clearance exacerbates iron retention. Thus, stage-dependent QSM elevation may arise not only from increased import or neuronal release but also from failed clearance, suggesting that glymphatic- or vascular-targeted strategies could complement iron-focused interventions (48). Region-specific vulnerability may arise from interactions between neurotransmitter systems and iron dysregulation. The SN, with its high dopamine turnover and relatively low antioxidant capacity, may experience a toxic synergy between iron and dopamine auto-oxidation. In contrast, functional abnormalities in the early visual cortex may reflect α-synuclein propagation along Braak pathways or downstream network effects, rather than local iron deposition. Furthermore, differences in iron-handling capacities across dopaminergic, cholinergic, and glutamatergic systems may shape the spatiotemporal pattern of dysfunction, potentially explaining the dissociation between QSM and rs-fMRI abnormalities in the early stages of the disease.

In early PD, functional disruption is most pronounced in the visual network, consistent with prior reports of reduced ALFF and ReHo in the lingual and calcarine cortices. Our ESP findings reproduce this pattern, suggesting that early visual dysfunction arises before major motor-loop impairment (49). We also observed that iron deposition in the GP and SN of early PD patients was significantly higher than in HCs, consistent with previous studies (14,50). However, not all studies report elevated GP iron levels in the early stage (51), suggesting that such heterogeneity may arise from differences in clinical staging criteria, cohort characteristics, or QSM methodologies.

In the ASP, QSM values in the SN and GP significantly increased compared with early PD, and both regions showed positive correlations with UPDRS-III scores. Previous studies have similarly reported stage-dependent increases in nigral and pallidal iron and stronger associations with motor severity in more advanced PD, supporting the link between iron burden and clinical progression (52,53). The convergence of increased SN/GP iron deposition with reduced ReHo and fALFF in the putamen, caudate, and thalamus suggests that iron accumulation may impair local synchrony and disrupt basal ganglia-thalamocortical communication. This cross-modality association provides mechanistic coherence beyond what unimodal analyses can offer. From a mechanistic perspective, these findings suggest that iron may initially accumulate as a marker of metabolic stress but could transition into an active driver of degeneration once ferroptotic thresholds are exceeded. Determining whether this inflection precedes or follows functional network breakdown will require longitudinal imaging; however, the rapidly rising QSM signal in advanced PD implies that a stage-specific shift from compensatory to toxic iron handling is plausible. The convergence of elevated SN/GP iron with reduced local synchrony in connected motor regions suggests that iron dysregulation may impair network efficiency, supporting mechanistic coherence across modalities (54-56).

In ASP, functional abnormalities shift to the basal ganglia-thalamocortical motor loop, characterized by widespread fALFF and ReHo abnormalities in the putamen, thalamus, and sensory-motor cortex. The weakening of neuronal activity consistency in the putamen, caudate nucleus, and thalamus, associated with dopamine deficiency and iron deposition, directly leads to core symptoms such as bradykinesia and rigidity (57,58). The observed hyperactivity in sensorimotor cortex may represent an attempted compensatory response to basal ganglia dysfunction, whereby cortical motor regions recruit additional neural resources to maintain motor output. However, whether this increased activity is functionally adaptive remains unclear. Some evidence suggests cortical hyperactivity may contribute to symptom severity (e.g., rigidity, bradykinesia) rather than ameliorating deficits (59,60). Longitudinal studies correlating cortical activity changes with symptom trajectories are needed to resolve this question.

Our study further found that fALFF values in the bilateral putamen, caudate, and thalamus were negatively correlated with UPDRS-III scores; similarly, ReHo values in these regions were also negatively correlated with UPDRS-III scores. These results suggest that reduced spontaneous neuronal activity and decreased local functional synchronization in these key motor-related brain areas are closely related to the severity of motor disturbances in PD patients. Dysfunction in the putamen may impair the integration of dopaminergic signals and motor initiation, while thalamic dysfunction hinders the transmission and coordination of sensory-motor information. In addition to basal ganglia-thalamocortical loop dysfunction, disturbances in the cerebellum-thalamocortical pathway may also contribute to motor symptoms. Hu et al. reported that although right-cerebellar fALFF did not differ significantly between PD patients and controls, it correlated with UPDRS-III scores, indicating cerebellar involvement in motor impairment (61). Li et al. further showed that ReHo values in the right anterior/posterior central gyrus decrease with advancing H&Y stage, supporting progressive sensory-motor cortex dysfunction (62).

Previous studies have shown that the onset of motor and non-motor symptoms in PD is closely linked to dysfunction in the DMN and fronto-parietal network (FPN) (63,64). In line with this, levodopa has been shown to partially normalize DMN connectivity in PD (65). The FPN, which is highly dependent on dopaminergic input, also shows impaired functional integrity in PD and has been linked to deficits in executive function and motor control (66,67). Our findings are consistent with these observations, further supporting the notion that large-scale cognitive control networks are progressively disrupted as the disease advances.

In ASP, the limbic system is widely affected, providing a neurobiological explanation for memory and emotional disturbances. The hippocampus, as a key brain region involved in learning, memory, and emotional responses, is closely linked to cognitive performance (68,69).

In PD patients, ReHo values in the bilateral hippocampus positively correlated with MMSE scores, indicating that preserved local synchrony in this region supports global cognitive function. This aligns with prior work linking hippocampal dysfunction to non-motor symptoms in PD, such as daytime sleepiness (70). Taken together, these findings suggest that limbic system dysfunction plays a critical role in non-motor symptoms and may progress even before the exacerbation of motor symptoms. The abnormality in the limbic system may reflect potential influences from α-synuclein pathology and iron-related oxidative processes, although these mechanisms remain hypothetical in the context of cross-sectional data.

The spatiotemporal dynamics of iron deposition and neural activity changes observed in this study further support the conclusion by Ruppert et al. that the pathological mechanisms of PD involve multilayered abnormalities in neural networks (71). Our multimodal MRI approach delineated the progression of both iron deposition and functional disruption across early and advanced stages, offering a refined imaging basis for disease staging. The convergence between QSM increases and reductions in fALFF/ReHo indicates a stratified pattern of dysfunction—from subcortical motor circuits to cognitive and limbic networks—and provides mechanistic coherence that single-modality analyses could not achieve, underscoring the added explanatory value of our multimodal framework as an original contribution to characterizing PD pathophysiology.

Several limitations should be considered. First, the cross-sectional design prevents causal inferences between iron deposition and functional changes, requiring longitudinal validation. Second, scanning in a medicated state may influence fMRI metrics, masking disease effects. The absence of systematic levodopa equivalent daily dose (LEDD) data prevented its use as a covariate, potentially introducing medication-related variability. Third, we did not account for comorbidities or lifestyle factors, which could confound brain iron deposition and functional connectivity. Fourth, dichotomizing disease severity (H&Y <3 vs. ≥3) may miss finer progression patterns due to sample size constraints. Fifth, while QSM suggests iron deposition, contributions from other susceptibility sources (e.g., venous blood, myelin) cannot be excluded. Sixth, despite FDR correction, multiple comparisons increase the risk of false positives, underscoring the need for independent replication. Finally, the single-center Chinese cohort may limit generalizability, highlighting the need for multi-center studies with diverse populations.

Conclusions

Multimodal MRI metrics not only serve as sensitive imaging biomarkers for the early identification and staging of PD, but also provide novel insights into its pathological mechanisms. This study confirms that iron deposition is closely associated with dopaminergic neuronal degeneration and neural network dysfunction, and correlates with abnormalities in multimodal functional indicators. Previous studies have suggested that mechanisms such as oxidative stress and neuroinflammation may underlie these processes, which support the central role of iron metabolism abnormalities in PD progression. Although preclinical work supports iron chelation, clinical efficacy remains limited, emphasizing the need to define when iron shifts from compensatory to pathological. The stage-dependent iron increases observed here underscore the importance of identifying this therapeutic window, which will require longitudinal studies. Beyond chelation, integrating multimodal MRI with markers of ferroptosis, iron export, and glymphatic function may help clarify how iron dysregulation drives network-level disturbances in PD.

Acknowledgments

We thank all the participants for their cooperation in this study.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2070/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2070/dss

Funding: This work was sponsored by the National Natural Science Foundation of China (Nos. 82441022, 82441023, 82371934); Special Training Program for Clinical Research Physicians under the “Three 100” Initiative of Henan Province (No. HNCRD202406); National Key R&D Program of China (No. 2023YFC2414200); The Medical Science and Technology Research Project of Henan Province (No. SBGJ202303007); The Science and Technology Project of Henan Province (No. 242102311029).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2070/coif). B.W. reports being an employee of Shanghai Zhu Yan Medical Technology Ltd. K.L. reports being an employee of Siemens Healthineers Ltd. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Research Ethics Committee (REC) of Henan Provincial People’s Hospital (HPPH) (No. 2018-064-02) and informed consent was obtained from all individual participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Emamzadeh FN, Surguchov A. Parkinson's Disease: Biomarkers, Treatment, and Risk Factors. Front Neurosci 2018;12:612. [Crossref] [PubMed]
2. Ben-Shlomo Y, Darweesh S, Llibre-Guerra J, Marras C, San Luciano M, Tanner C. The epidemiology of Parkinson's disease. Lancet 2024;403:283-92. [Crossref] [PubMed]
3. Smith L, Schapira AHV. GBA Variants and Parkinson Disease: Mechanisms and Treatments. Cells 2022;11:1261. [Crossref] [PubMed]
4. Prasuhn J, Brüggemann N. Genotype-driven therapeutic developments in Parkinson's disease. Mol Med 2021;27:42. [Crossref] [PubMed]
5. Sohrabi T, Mirzaei-Behbahani B, Zadali R, Pirhaghi M, Morozova-Roche LA, Meratan AA. Common Mechanisms Underlying α-Synuclein-Induced Mitochondrial Dysfunction in Parkinson's Disease. J Mol Biol 2023;435:167992. [Crossref] [PubMed]
6. Guan X, Lancione M, Ayton S, Dusek P, Langkammer C, Zhang M. Neuroimaging of Parkinson's disease by quantitative susceptibility mapping. Neuroimage 2024;289:120547. [Crossref] [PubMed]
7. Villalón-García I, Povea-Cabello S, Álvarez-Córdoba M, Talaverón-Rey M, Suárez-Rivero JM, Suárez-Carrillo A, Munuera-Cabeza M, Reche-López D, Cilleros-Holgado P, Piñero-Pérez R, Sánchez-Alcázar JA. Vicious cycle of lipid peroxidation and iron accumulation in neurodegeneration. Neural Regen Res 2023;18:1196-202. [Crossref] [PubMed]
8. Wen J, Guo T, Wu J, Bai X, Zhou C, Wu H, Liu X, Chen J, Cao Z, Gu L, Pu J, Zhang B, Zhang M, Guan X, Xu X. Nigral Iron Deposition Influences Disease Severity by Modulating the Effect of Parkinson's Disease on Brain Networks. J Parkinsons Dis 2022;12:2479-92. [Crossref] [PubMed]
9. Ward RJ, Zucca FA, Duyn JH, Crichton RR, Zecca L. The role of iron in brain ageing and neurodegenerative disorders. Lancet Neurol 2014;13:1045-60. [Crossref] [PubMed]
10. Naduthota RM, Honnedevasthana AA, Lenka A, Saini J, Geethanath S, Bharath RD, Christopher R, Yadav R, Gupta AK, Pal PK. Association of freezing of gait with nigral iron accumulation in patients with Parkinson's disease. J Neurol Sci 2017;382:61-5. [Crossref] [PubMed]
11. Foley PB, Hare DJ, Double KL. A brief history of brain iron accumulation in Parkinson disease and related disorders. J Neural Transm (Vienna) 2022;129:505-20. [Crossref] [PubMed]
12. Li G, Tong R, Zhang M, Gillen KM, Jiang W, Du Y, Wang Y, Li J. Age-dependent changes in brain iron deposition and volume in deep gray matter nuclei using quantitative susceptibility mapping. Neuroimage 2023;269:119923. [Crossref] [PubMed]
13. Meng H, Zhang D, Sun Q. The applied value in brain gray matter nuclei of patients with early-stage Parkinson's disease : a study based on multiple magnetic resonance imaging techniques. Head Face Med 2023;19:25. [Crossref] [PubMed]
14. Fu X, Deng W, Cui X, Zhou X, Song W, Pan M, Chi X, Xu J, Jiang Y, Wang Q, Xu Y. Time-Specific Pattern of Iron Deposition in Different Regions in Parkinson's Disease Measured by Quantitative Susceptibility Mapping. Front Neurol 2021;12:631210. [Crossref] [PubMed]
15. Thomas GEC, Hannaway N, Zarkali A, Shmueli K, Weil RS. Longitudinal Associations of Magnetic Susceptibility with Clinical Severity in Parkinson's Disease. Mov Disord 2024;39:546-59. [Crossref] [PubMed]
16. Biswal BB. Resting state fMRI: a personal history. Neuroimage 2012;62:938-44. [Crossref] [PubMed]
17. Hohenfeld C, Werner CJ, Reetz K. Resting-state connectivity in neurodegenerative disorders: Is there potential for an imaging biomarker? Neuroimage Clin 2018;18:849-70. [Crossref] [PubMed]
18. Zou QH, Zhu CZ, Yang Y, Zuo XN, Long XY, Cao QJ, Wang YF, Zang YF. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF. J Neurosci Methods 2008;172:137-41. [Crossref] [PubMed]
19. Hou Y, Zhang L, Ou R, Wei Q, Gu X, Liu K, Lin J, Yang T, Xiao Y, Gong Q, Shang H. Motor progression marker for newly diagnosed drug-naïve patients with Parkinson's disease: A resting-state functional MRI study. Hum Brain Mapp 2023;44:901-13. [Crossref] [PubMed]
20. Chen P, Tang G, Wang Y, Xiong W, Deng Y, Fei S, Zhang J. Spontaneous brain activity in the hippocampal regions could characterize cognitive impairment in patients with Parkinson's disease. CNS Neurosci Ther 2024;30:e14706. [Crossref] [PubMed]
21. Zang Y, Jiang T, Lu Y, He Y, Tian L. Regional homogeneity approach to fMRI data analysis. Neuroimage 2004;22:394-400. [Crossref] [PubMed]
22. Pan P, Zhan H, Xia M, Zhang Y, Guan D, Xu Y. Aberrant regional homogeneity in Parkinson's disease: A voxel-wise meta-analysis of resting-state functional magnetic resonance imaging studies. Neurosci Biobehav Rev 2017;72:223-31. [Crossref] [PubMed]
23. Wu T, Long X, Zang Y, Wang L, Hallett M, Li K, Chan P. Regional homogeneity changes in patients with Parkinson's disease. Hum Brain Mapp 2009;30:1502-10. [Crossref] [PubMed]
24. Nguyen KP, Raval V, Treacher A, Mellema C, Yu FF, Pinho MC, Subramaniam RM, Dewey RB Jr, Montillo AA. Predicting Parkinson's disease trajectory using clinical and neuroimaging baseline measures. Parkinsonism Relat Disord 2021;85:44-51. [Crossref] [PubMed]
25. Goetz CG, Poewe W, Rascol O, Sampaio C, Stebbins GT, Counsell C, Giladi N, Holloway RG, Moore CG, Wenning GK, Yahr MD, Seidl LMovement Disorder Society Task Force on Rating Scales for Parkinson's Disease. Movement Disorder Society Task Force report on the Hoehn and Yahr staging scale: status and recommendations. Mov Disord 2004;19:1020-8. [Crossref] [PubMed]
26. Postuma RB, Berg D, Stern M, Poewe W, Olanow CW, Oertel W, Obeso J, Marek K, Litvan I, Lang AE, Halliday G, Goetz CG, Gasser T, Dubois B, Chan P, Bloem BR, Adler CH, Deuschl G. MDS clinical diagnostic criteria for Parkinson's disease. Mov Disord 2015;30:1591-601. [Crossref] [PubMed]
27. Li H, Jia J, Yang Z. Mini-Mental State Examination in Elderly Chinese: A Population-Based Normative Study. J Alzheimers Dis 2016;53:487-96. [Crossref] [PubMed]
28. Goetz CG, Tilley BC, Shaftman SR, Stebbins GT, Fahn S, Martinez-Martin P, et al. Movement Disorder Society-sponsored revision of the Unified Parkinson's Disease Rating Scale (MDS-UPDRS): scale presentation and clinimetric testing results. Mov Disord 2008;23:2129-70. [Crossref] [PubMed]
29. Leentjens AF, Dujardin K, Marsh L, Richard IH, Starkstein SE, Martinez-Martin P. Anxiety rating scales in Parkinson's disease: a validation study of the Hamilton anxiety rating scale, the Beck anxiety inventory, and the hospital anxiety and depression scale. Mov Disord 2011;26:407-15. [Crossref] [PubMed]
30. Ziropadja Lj, Stefanova E, Petrovic M, Stojkovic T, Kostic VS. Apathy and depression in Parkinson's disease: the Belgrade PD study report. Parkinsonism Relat Disord 2012;18:339-42. [Crossref] [PubMed]
31. Jia XZ, Wang J, Sun HY, Zhang H, Liao W, Wang Z, Yan CG, Song XW, Zang YF. RESTplus: an improved toolkit for resting-state functional magnetic resonance imaging data processing. Sci Bull (Beijing) 2019;64:953-4. [Crossref] [PubMed]
32. Chen Y, Liu S, Wang Y, Kang Y, Haacke EM. STrategically Acquired Gradient Echo (STAGE) imaging, part I: Creating enhanced T1 contrast and standardized susceptibility weighted imaging and quantitative susceptibility mapping. Magn Reson Imaging 2018;46:130-9. [Crossref] [PubMed]
33. Wang Y, Chen Y, Wu D, Wang Y, Sethi SK, Yang G, Xie H, Xia S, Haacke EM. STrategically Acquired Gradient Echo (STAGE) imaging, part II: Correcting for RF inhomogeneities in estimating T1 and proton density. Magn Reson Imaging 2018;46:140-50. [Crossref] [PubMed]
34. Haacke EM, Chen Y, Utriainen D, Wu B, Wang Y, Xia S, et al. STrategically Acquired Gradient Echo (STAGE) imaging, part III: Technical advances and clinical applications of a rapid multi-contrast multi-parametric brain imaging method. Magn Reson Imaging 2020;65:15-26. [Crossref] [PubMed]
35. Haacke EM, Liu S, Buch S, Zheng W, Wu D, Ye Y. Quantitative susceptibility mapping: current status and future directions. Magn Reson Imaging 2015;33:1-25. [Crossref] [PubMed]
36. Schweser F, Deistung A, Lehr BW, Reichenbach JR. Quantitative imaging of intrinsic magnetic tissue properties using MRI signal phase: an approach to in vivo brain iron metabolism? Neuroimage 2011;54:2789-807. [Crossref] [PubMed]
37. Haacke EM, Tang J, Neelavalli J, Cheng YC. Susceptibility mapping as a means to visualize veins and quantify oxygen saturation. J Magn Reson Imaging 2010;32:663-76. [Crossref] [PubMed]
38. Tang J, Liu S, Neelavalli J, Cheng YC, Buch S, Haacke EM. Improving susceptibility mapping using a threshold-based K-space/image domain iterative reconstruction approach. Magn Reson Med 2013;69:1396-407. [Crossref] [PubMed]
39. Wharton S, Schäfer A, Bowtell R. Susceptibility mapping in the human brain using threshold-based k-space division. Magn Reson Med 2010;63:1292-304. [Crossref] [PubMed]
40. Jin Z, Wang Y, Jokar M, Li Y, Cheng Z, Liu Y, Tang R, Shi X, Zhang Y, Min J, Liu F, He N, Yan F, Haacke EM. Automatic detection of neuromelanin and iron in the midbrain nuclei using a magnetic resonance imaging-based brain template. Hum Brain Mapp 2022;43:2011-25. [Crossref] [PubMed]
41. Chai C, Qiao P, Zhao B, Wang H, Liu G, Wu H, Shen W, Cao C, Ye X, Liu Z, Xia S. Brain gray matter nuclei segmentation on quantitative susceptibility mapping using dual-branch convolutional neural network. Artif Intell Med 2022;125:102255. [Crossref] [PubMed]
42. Zang YF, He Y, Zhu CZ, Cao QJ, Sui MQ, Liang M, Tian LX, Jiang TZ, Wang YF. Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI. Brain Dev 2007;29:83-91. [Crossref] [PubMed]
43. Braak H, Ghebremedhin E, Rüb U, Bratzke H, Del Tredici K. Stages in the development of Parkinson's disease-related pathology. Cell Tissue Res 2004;318:121-34. [Crossref] [PubMed]
44. Wang ZL, Yuan L, Li W, Li JY. Ferroptosis in Parkinson's disease: glia-neuron crosstalk. Trends Mol Med 2022;28:258-69. [Crossref] [PubMed]
45. Zucca FA, Segura-Aguilar J, Ferrari E, Muñoz P, Paris I, Sulzer D, Sarna T, Casella L, Zecca L. Interactions of iron, dopamine and neuromelanin pathways in brain aging and Parkinson's disease. Prog Neurobiol 2017;155:96-119. [Crossref] [PubMed]
46. Fischbacher A, von Sonntag C, Schmidt TC. Hydroxyl radical yields in the Fenton process under various pH, ligand concentrations and hydrogen peroxide/Fe(II) ratios. Chemosphere 2017;182:738-44. [Crossref] [PubMed]
47. Hare DJ, Double KL. Iron and dopamine: a toxic couple. Brain 2016;139:1026-35. [Crossref] [PubMed]
48. Chen X, Pang X, Yeo AJ, Xie S, Xiang M, Shi B, Yu G, Li C. The Molecular Mechanisms of Ferroptosis and Its Role in Blood-Brain Barrier Dysfunction. Front Cell Neurosci 2022;16:889765. [Crossref] [PubMed]
49. Luo C, Guo X, Song W, Chen Q, Yang J, Gong Q, Shang HF. The trajectory of disturbed resting-state cerebral function in Parkinson's disease at different Hoehn and Yahr stages. Hum Brain Mapp 2015;36:3104-16. [Crossref] [PubMed]
50. Martin-Bastida A, Lao-Kaim NP, Loane C, Politis M, Roussakis AA, Valle-Guzman N, Kefalopoulou Z, Paul-Visse G, Widner H, Xing Y, Schwarz ST, Auer DP, Foltynie T, Barker RA, Piccini P. Motor associations of iron accumulation in deep grey matter nuclei in Parkinson's disease: a cross-sectional study of iron-related magnetic resonance imaging susceptibility. Eur J Neurol 2017;24:357-65. [Crossref] [PubMed]
51. Li KR, Avecillas-Chasin J, Nguyen TD, Gillen KM, Dimov A, Chang E, Skudin C, Kopell BH, Wang Y, Shtilbans A. Quantitative evaluation of brain iron accumulation in different stages of Parkinson's disease. J Neuroimaging 2022;32:363-71. [Crossref] [PubMed]
52. An H, Zeng X, Niu T, Li G, Yang J, Zheng L, Zhou W, Liu H, Zhang M, Huang D, Li J. Quantifying iron deposition within the substantia nigra of Parkinson's disease by quantitative susceptibility mapping. J Neurol Sci 2018;386:46-52. [Crossref] [PubMed]
53. Shahmaei V, Faeghi F, Mohammdbeigi A, Hashemi H, Ashrafi F. Evaluation of iron deposition in brain basal ganglia of patients with Parkinson's disease using quantitative susceptibility mapping. Eur J Radiol Open 2019;6:169-74. [Crossref] [PubMed]
54. Ding XS, Gao L, Han Z, Eleuteri S, Shi W, Shen Y, Song ZY, Su M, Yang Q, Qu Y, Simon DK, Wang XL, Wang B. Ferroptosis in Parkinson's disease: Molecular mechanisms and therapeutic potential. Ageing Res Rev 2023;91:102077. [Crossref] [PubMed]
55. Abeyawardhane DL, Lucas HR. Iron Redox Chemistry and Implications in the Parkinson's Disease Brain. Oxid Med Cell Longev 2019;2019:4609702. [Crossref] [PubMed]
56. Weinreb O, Mandel S, Youdim MBH, Amit T. Targeting dysregulation of brain iron homeostasis in Parkinson's disease by iron chelators. Free Radic Biol Med 2013;62:52-64. [Crossref] [PubMed]
57. Haber SN, Knutson B. The reward circuit: linking primate anatomy and human imaging. Neuropsychopharmacology 2010;35:4-26. [Crossref] [PubMed]
58. Helmich RC, Derikx LC, Bakker M, Scheeringa R, Bloem BR, Toni I. Spatial remapping of cortico-striatal connectivity in Parkinson's disease. Cereb Cortex 2010;20:1175-86. [Crossref] [PubMed]
59. Yu H, Sternad D, Corcos DM, Vaillancourt DE. Role of hyperactive cerebellum and motor cortex in Parkinson's disease. Neuroimage 2007;35:222-33. [Crossref] [PubMed]
60. Xu H, Wang Y, Song N, Wang J, Jiang H, Xie J. New Progress on the Role of Glia in Iron Metabolism and Iron-Induced Degeneration of Dopamine Neurons in Parkinson's Disease. Front Mol Neurosci 2017;10:455. [Crossref] [PubMed]
61. Hu XF, Zhang JQ, Jiang XM, Zhou CY, Wei LQ, Yin XT, Li J, Zhang YL, Wang J. Amplitude of low-frequency oscillations in Parkinson's disease: a 2-year longitudinal resting-state functional magnetic resonance imaging study. Chin Med J (Engl) 2015;128:593-601. [Crossref] [PubMed]
62. Li J, Liao H, Wang T, Zi Y, Zhang L, Wang M, Mao Z, Song C, Zhou F, Shen Q, Cai S, Tan C. Alterations of Regional Homogeneity in the Mild and Moderate Stages of Parkinson's Disease. Front Aging Neurosci 2021;13:676899. [Crossref] [PubMed]
63. Wang S, Xiao Y, Hou Y, Li C, Lin J, Yang T, Che N, Jiang Q, Zheng X, Liu J, Shang H. Altered gait speed and brain network connectivity in Parkinson's disease. Cereb Cortex 2024;34:bhae429. [Crossref] [PubMed]
64. Yeager BE, Twedt HP, Bruss J, Schultz J, Narayanan NS. Cortical and subcortical functional connectivity and cognitive impairment in Parkinson's disease. Neuroimage Clin 2024;42:103610. [Crossref] [PubMed]
65. Zhong J, Guan X, Zhong X, Cao F, Gu Q, Guo T, Zhou C, Zeng Q, Wang J, Gao T, Zhang M. Levodopa imparts a normalizing effect on default-mode network connectivity in non-demented Parkinson's disease. Neurosci Lett 2019;705:159-66. [Crossref] [PubMed]
66. Petersen SE, Posner MI. The attention system of the human brain: 20 years after. Annu Rev Neurosci 2012;35:73-89. [Crossref] [PubMed]
67. Lang S, Hanganu A, Gan LS, Kibreab M, Auclair-Ouellet N, Alrazi T, Ramezani M, Cheetham J, Hammer T, Kathol I, Sarna J, Monchi O. Network basis of the dysexecutive and posterior cortical cognitive profiles in Parkinson's disease. Mov Disord 2019;34:893-902. [Crossref] [PubMed]
68. Squire LR, Stark CE, Clark RE. The medial temporal lobe. Annu Rev Neurosci 2004;27:279-306. [Crossref] [PubMed]
69. Villar-Conde S, Astillero-Lopez V, Gonzalez-Rodriguez M, Villanueva-Anguita P, Saiz-Sanchez D, Martinez-Marcos A, Flores-Cuadrado A, Ubeda-Bañon I. The Human Hippocampus in Parkinson's Disease: An Integrative Stereological and Proteomic Study. J Parkinsons Dis 2021;11:1345-65. [Crossref] [PubMed]
70. Zi Y, Cai S, Tan C, Wang T, Shen Q, Liu Q, Wang M, Li J, Zhang L, Zhou F, Song C, Yuan J, Liu Y, Liu J, Liao H. Abnormalities in the Fractional Amplitude of Low-Frequency Fluctuation and Functional Connectivity in Parkinson's Disease With Excessive Daytime Sleepiness. Front Aging Neurosci 2022;14:826175. [Crossref] [PubMed]
71. Ruppert MC, Greuel A, Tahmasian M, Schwartz F, Stürmer S, Maier F, Hammes J, Tittgemeyer M, Timmermann L, van Eimeren T, Drzezga A, Eggers C. Network degeneration in Parkinson's disease: multimodal imaging of nigro-striato-cortical dysfunction. Brain 2020;143:944-59. [Crossref] [PubMed]