Abnormal topological organization of structural covariance networks in early-stage Parkinson’s disease patients with autonomic dysfunction

Introduction

Autonomic dysfunction, a universal nonmotor symptom in Parkinson’s disease (PD), becomes more prominent as the disease develops (1-3). In many PD patients, autonomic symptoms often precede the onset of motor symptoms or present in the early stages of disease (4). Dysautonomia, a predictor for prognosis and survival, is an important determinant of quality of life itself and global motor severity (1). Previous studies demonstrated that PD patients had impairment in various autonomic domains, such as cardiovascular, gastrointestinal, urinary, sexual, and thermoregulatory function (2,3,5). It is logical to propose that the systemic pathogenesis of PD is widespread and encompasses both the central and peripheral components of the autonomic nervous system (6). However, the characteristics of central neural networks in PD patients with dysautonomia remain under debate and far from being fully identified.

A potential strategy for identifying biomarkers in PD involves the measurement of interactions within large-scale brain networks using magnetic resonance imaging (MRI). This method has been employed to investigate a variety of motor and nonmotor symptoms in PD (7-9). Recent neuroimaging studies have uncovered complicated regimes, containing interrupted functional connectivity within the lateral premotor-parietal (10) and thalamo-striatal-hypothalamic (11) loops, hypoperfusion in the frontal lobe and insula (12), damaged gray matter (GM) in the inferior temporal cortex (13), and white matter (WM) connectivity collapse in the frontal-subcortical regions (14), which might be associated with dysautonomia in PD patients. The existing studies have mainly involved abnormal alterations in local or zonal areas of the brain, whereas the large-scale global network alteration patterns in early-stage PD patients with autonomic dysfunction remain elusive.

Structural covariance analysis of similarities in GM between brain regions, referred to as the structural covariance network (SCN), can portray the large-scale structural brain network by identifying interregional covariance. A wealth of studies have implicated that the SCN might mirror inherited developmental harmonization or synchronized maturation between disparate brain regions, apart from closely resembling the gene expression profile and imminent functional network architecture of the human brain (15). Based on these basic characteristics, several investigations have uncovered that SCN could be utilized to explore network-level alterations in healthy aging populations (15) as well as those with neurodegenerative (7-9,16) and neuropsychiatric disorders (17,18). Graph theory-based analysis methods, an efficacious tool for characterizing the topological properties of large-scale brain networks, have been successfully used to investigate the topological malformation of the SCN in assorted neurological diseases, including amyotrophic lateral sclerosis (19) and PD (20). Wang et al. successfully applied graph theory to discover that the SCN in PD patients exhibited small-world property and abnormal topological organization (20). Hence, we conjectured that the SCN of early-stage PD patients with autonomic dysfunction would have suboptimal topological organization. To validate this hypothesis, we aimed to evaluate network topology changes in early-stage PD patients with autonomic dysfunction by the SCN established via brain gray matter volume (GMV). We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-2310/rc).

Methods

Participants

We obtained all data from the Parkinson’s Progression Markers Initiative (PPMI) database in April 2024 (http://www.ppmi-info.org/). A detailed description of the study’s objectives and methodology can be found elsewhere (21). The Institutional Review Boards sanctioned all procedures and all PD patients provided written informed consent. Study protocols and associated manuals (including inclusion criteria) are available online at www.ppmi-info.org/study-design. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

The initial search strategy comprised the following criteria: age: ≥50 years; research group: PD; study visit: baseline; image modality: MRI, field strength: 3 Tesla (3T); slice thickness: ≤1.5 mm; acquisition type: three-dimensional (3D); magnetic resonance (MR) system manufacturer: Siemens (Erlangen, Germany); weighting: T1. A total of 341 early-stage PD patients with baseline 3D T1-weighted scans acquired using Siemens MR systems were retrieved. The following exclusion criteria were used to screen patients: (I) without dopamine transporter (DAT) single-photon emission computed tomography (SPECT) or with negative DAT SPECT; (II) without Scales for Outcomes in Parkinson’s Disease-Autonomic (SCOPA-AUT), the Movement Disorders Society sponsored revision of Unified Parkinson’s Disease Rating Scale (MDS-UPDRS-III), Hoehn and Yahr stage (H-Y stage), or Montreal Cognitive Assessment (MoCA); (III) with cognitive impairment or dementia (MoCA <24). A total of 279 PD patients with 3D T1-weighted scans (Siemens scanners) were selected. It should be noted that although the SCOPA-AUT scale has been previously validated in PD (22-24), the cut-off values are not reported. Dayan et al. used quartiles of a SCOPA-AUT score to divide PD patients in the PPMI database into PD patients with a higher burden of autonomic symptoms (AUT_high) and lower burden (AUT_low), and successfully found disrupted thalamo-striato-hypothalamic functional connectivity in the AUT_high group. Correspondingly, the histograms of the distribution of SCOPA-AUT for these selected cases are shown in Figure S1, which was similar to the distribution reported in Dayan et al.’s research (11). According to the interquartile spacing, PD patients with the upper [≥75% (upper quartile), n=73] and lower [≤ lower quartile (25%), n=89] score on the SCOPA-AUT were assigned to the AUT_high and AUT_low groups, respectively (Figure 1) in our study. The ID numbers of all cases can be obtained in Table S1.

Figure 1 Subject selection process. AUT_high, Parkinson’s disease patients with the highest (≥75%) quartiles score on the SCOPA-AUT; AUT_low, Parkinson’s disease patients with the lowest (≤25%) quartiles score on the SCOPA-AUT; PD, Parkinson’s disease; SCOPA-AUT, Scales for Outcomes in Parkinson’s Disease-Autonomic.

Clinical evaluation

MDS-UPDRS-III and H-Y stage were applied to evaluate the severity of motor symptoms and disease staging. MoCA was used to assess the cognitive function (25). Specifically, the SCOPA-AUT questionnaire, a reliable and effective self-report scale of dysautonomia, was utilized to evaluate the existence of autonomic symptoms (22). The SCOPA-AUT contains 25 items, including gastrointestinal (7), urinary (6), cardiovascular (3), thermoregulatory (4), pupillomotor (1), and sexual (2 questions for men and 2 questions for women) symptoms. The symptoms of autonomic dysfunction are more severe and frequent as the scores increase, with the exception of sexual dysfunction. Additionally, the scale does not incorporate the asymmetrical distribution of symptoms in domains such as pupil motility or thermoregulatory dysfunction. The aforementioned scales were administered during the drug OFF state. Our study did not calculate levodopa equivalent dose as most PD patients at the baseline had not yet begun dopamine replacement therapy.

MRI acquisition and preprocessing

3D T1-weighted images were acquired with Siemens scanners with the following protocols: echo time (TE) =2.3–3.6 ms, repetition time (TR) =1,900–2,400 ms, flip angle (degree) =9, slice thickness ≤1.5 mm. The preprocessing of all 3D T1-weighted images was conducted by the computational anatomy toolbox (CAT12; http://dbm.neuro.uni-jena.de/cat/) in MATLAB 2018b (MathWorks, Natick, MA, USA) as the default pipeline: format conversion, correction for bias-field inhomogeneities, segmentation into GM and WM and cerebrospinal fluid, and normalization by DARTEL. Then, we generated 123 cortical and subcortical regions of interest (ROIs), excluding the WM, cerebrospinal fluid, and ventricles, from the Neuromorphometrics Atlas (http://Neuromorphometrics.com/) to extract the individual average GMV of each ROI for subsequent construction of the SCN (Table S1). Notably, individual total intracranial volume (TIV) values were also calculated for further analysis.

Construction of SCNs

Brain Connectivity Toolbox (https://www.nitrc.org/projects/bct) was applied to construct the GMV-based SCN. The Pearson correlation coefficients between the GMV in each ROI were calculated after regressing age, sex, MDS-UPDRS- III scores, and TIV, and then a 123×123 correlation matrix was constructed for each group (Figure 2A-2D). Subsequently, we binarized the matrix in both groups within the network sparsity range of 0.10 to 0.40 with an interval of 0.01 (Figure 2E,2F). Specifically, the correlation coefficient above the threshold was set to 1; otherwise, it was set to 0 (including weights on the main diagonal).

Figure 2 Correlation matrices with 123×123 for AUT_high and AUT_low. Original correlation matrices for (A) AUT_high and (B) AUT_low patients, adjusted correlation matrices (by age, sex, MDS-UPDRS-III scores and TIV) for (C) AUT_high and (D) AUT_low patients, and binary adjacency matrices thresholded at D_min=0.10 for (E) AUT_high and (F) AUT_low patients. Both x-axis and y-axis represent brain regions based on Neuromorphometrics Atlas. The color bar denotes the correlation coefficient and represents the strength of the connections. AUT_high, Parkinson’s disease patients with the highest (≥75%) quartiles score on the SCOPA-AUT; AUT_low, Parkinson’s disease patients with the lowest (≤25%) quartiles score on the SCOPA-AUT; MDS-UPDRS, Movement Disorders Society sponsored revision of Unified Parkinson’s Disease Rating Scale; SCOPA-AUT, Scales for Outcomes in Parkinson’s Disease-Autonomic; TIV, total intracranial volume.

Network parameters

As in previous studies (19,20), a succession of global and regional network parameters was generally utilized to portray the topological organization of the GMV-based SCN. Global parameters included the normalized clustering coefficient (Gamma), normalized characteristic path length (Lambda), small-world index (Sigma), global efficiency (Eglobal), and local efficiency (Elocal), which were the most widely used indicators for quantifying the small-world topology of a network. Gamma and Lambda were calculated by comparing the C and characteristic path length to the parallel mean values of 100 corresponding random networks. Nodal parameters containing betweenness centrality (BC) and nodal efficiency (Ne) were also computed in each separate node of the GMV-based SCN.

Statistical analysis

For demographic and clinical features, all data were analyzed using the software SPSS 25.0 (IBM Corp., Armonk, NY, USA). Kolmogorov-Smirnov tests were performed to assess for normality. Two-sample t-tests and Mann-Whitney U tests were carried out for continuous variables, paralleling with χ² test for discrete variables. A two-tailed P value <0.05 was considered significant.

For global and regional topological network parameters, a nonparametric permutation test with 1,000 repetitions was conducted to estimate the statistical differences between AUT_high and AUT_low groups. During each permutation, the corrected GM volumes of each participant were randomly redistributed to one of the two new groups, ensuring that both new groups had the same number of cases as the original AUT_high and AUT_low groups. Specifically, the ‘randperm’ function (in Matlab R2018b) randomly rearranges the sample numbers of all cases in each iteration. The first 73 samples of the permuted dataset were regarded as the AUT_high group, and the last 89 samples were regarded as the AUT_low group. Therefore, the permutation randomly assigns sample labels in each iteration to test the significance of the differences in network parameters. The correlation matrix for each randomized group was recalculated and binarized across sparsity thresholds (0.1–0.4, step: 0.01). Each network’s topological properties were estimated at each density. The intergroup disparities in network parameters for the two randomized groups were then calculated to create a permutation distribution of differences under the null hypothesis. To determine the significance level, the actual difference between AUT_high and AUT_low was placed in its corresponding permutation distribution for each network parameter. Ultimately, we calculated the area under the curve (AUC) for each network parameter to supply a scalar that was independent of particular threshold selection. For global network parameters, the significant level was set at P<0.05. For regional network parameters, the significant differences were set at P<0.05, calibrated by false discovery rate (FDR) correction.

Results

Demographic and clinical characteristics

PD patients in the AUT_high group showed significantly higher MDS-UPDRS-III scores than those in the AUT_low (P<0.001; Table 1), which uncovered that early-stage PD patients in the AUT_high group had more severe motor symptoms. No significant difference was found in other variables, including age, sex, education, MDS-UPDRS-I, Hoehn and Yahr clinical rating scale (H-Y) stage, and TIV, between the two groups, suggesting that the two groups were well matched.

Alterations in global network parameters

Figure 3 displays the variations and intergroup differences in global network parameters between the AUT_high and AUT_low groups, over densities ranging from 0.10 to 0.40. We found that the GMV-based SCN of both groups exhibited a small-world property at all network sparsity with Gamma >1, Lambda ≈1, and Sigma >1 (Figure 3). Further analysis uncovered that the AUC of Sigma and Gamma were significantly decreased (both P<0.001) in the AUT_high group compared with the AUT_low group (Figure 4).

Figure 3 Changes in global network parameters as a function of network density. (A) Sigma: small-world index; (B) Gamma: normalized clustering coefficient; (C) Lambda: normalized characteristic path length; (D) Eglobal: global efficiency, (E) Elocal: local efficiency in AUT_high and AUT_low patients. AUT_high, Parkinson’s disease patients with the highest (≥75%) quartiles score on the SCOPA-AUT; AUT_low, Parkinson’s disease patients with the lowest (≤25%) quartiles score on the SCOPA-AUT; SCOPA-AUT, Scales for Outcomes in Parkinson’s Disease-Autonomic.

Figure 4 Differences between AUT_high and AUT_low patients in global network parameters as a function of network density. The red line denotes the difference between AUT_high and AUT_low PD patients; the red line lying outside of the confidence intervals indicate the density where the difference is significant at P<0.05. The positive values indicate AUT_high > AUT_low and negative values indicate AUT_high < AUT_low. ***, P<0.001. AUT_high, Parkinson’s disease patients with the highest (≥75%) quartiles score on the Scales for Outcomes in Parkinson’s Disease-Autonomic; AUT_low, Parkinson’s disease patients with the lowest (≤25%) quartiles score on the Scales for Outcomes in Parkinson’s Disease-Autonomic; CI, confidence interval; Eglobal, global efficiency; Elocal, local efficiency; Gamma, normalized clustering coefficient; Lambda, normalized characteristic path length; Sigma, small-world index.

Alterations in regional network parameters

BC

BC is usually applied to measure the number of times a brain node appears on the shortest path between any two other pairs of brain nodes. We found significantly decreased BC in the right thalamus proper (P<0.0001, FDR corrected) and increased BC in the cerebellar vermal lobules VI-VII (P<0.0001, FDR corrected) in AUT_high compared with AUT_low (Figure 5A).

Figure 5 Differences between AUT_high and AUT_low patients in nodal network parameters. (A) The distribution of the significant BC between AUT_high and AUT_low groups. (B) The distribution of the significant Ne between AUT_high and AUT_low groups. Red color indicates increased nodal properties, and the blue color indicates decreased nodal properties. AUT_high, Parkinson’s disease patients with the highest (≥75%) quartiles score on the SCOPA-AUT; AUT_low, Parkinson’s disease patients with the lowest (≤25%) quartiles score on the SCOPA-AUT; BC, betweenness centrality; Ne, nodal efficiency; SCOPA-AUT, Scales for Outcomes in Parkinson’s Disease-Autonomic.

Ne

Ne characterizes the capacity of parallel information processing for local nodes of GMV-based SCN. We discovered that PD patients in the AUT_high group had increased Ne in the left middle temporal gyrus compared with AUT_low (P<0.0001, FDR corrected, Figure 5B).

Discussion

The present study investigated the topological organization properties of the GMV-based SCN in early-stage PD patients with autonomic dysfunction. We found that PD patients in the AUT_high group had decreased small-world index and normalized clustering coefficient compared with those in the AUT_low group. For regional network parameters, PD patients in the AUT_high group exhibited decreased nodal BC in the right thalamus proper, increased nodal BC in the cerebellar vermal lobules VI–VII, and enhanced Ne in the left middle temporal gyrus compared with those in the AUT_low group. These manifestations intensified the comprehension of the altered topological organization in the VBM-based SCN of early-stage PD patients with autonomic dysfunction.

For a long time, more attention has been given to abnormalities in the peripheral autonomic nervous system for autonomic dysfunction in PD. In effect, the term central autonomous neural networks (CANs) was first proposed in 1993 by Benarroch, stemming from former investigations in animals and humans using electrical stimulation or tracer techniques (26). CAN, an internal regulation large-scale system composed of widespread cortical and subcortical nuclei, regulates visceromotor, neuroendocrine, pain, and behavioral responses indispensable for survival and is portrayed by reciprocal interconnections, parallel organization, state-dependent activity, and neurochemical complexity (26,27). The CAN receives and processes visceral, humoral, and external sensory signals through multiple parallel circuits to regulate the homeostasis of the autonomic nervous system (26). Put simply, autonomic dysfunction is an inevitable consequence of the disruption of CAN network information processing capacities. Previous postmortem investigations have uncovered that autonomic dysfunction in PD might be linked to the α-synuclein accumulation in certain brain regions, including the hypothalamus, insula, anterior cingulate, and dorsal motor nucleus of the vagus (14,28-30). Using graph theory methodologies, we could conduct large-scale topological investigations of the VBM-based SCN to identify and describe the crucial “weak networks and nodes” involved in the pathophysiological mechanisms of PD patients with autonomic dysfunction.

The small-world properties were defined by a greater local clustering coefficient and shorter path length, allowing for an optimal balance between local segregation and global integration (31). Meanwhile, small-world properties demonstrate both local and global efficiency in transmitting information, resulting in great efficiency at a low cost of wiring (32). Previous studies had shown abnormal global topological organization of cortical morphological SCN in PD patients (16,20,33). In this study, we found that all early-stage PD patients in both AUT_high and AUT_low exhibited small-world organization, implying that the brain was optimally structured to facilitate the efficient flow of information in distributed processing. Correspondingly, we found that PD patients in the AUT_high group had impaired small-world properties characterized by significant decreases in σ and γ values. Considering the similar pattern of λ between the AUT_high and AUT_low group, the lower σ organization might be explained by the decreased γ value. A high SCN clustering indicates the capacity of tightly coupled groups of brain areas to execute specialized processing procedures associated with autonomic function in early-stage PD (34,35). The drop in γ that we detected in PD patients with a higher burden of autonomic symptoms might be a result of a disruption in the neuronal segregation that occurs between linked areas of the brain.

Hubs are nodes in the brain network that have a high degree of influence on the hierarchical organization of the whole system. Without these heavily weighted brain regions, the whole network’s robustness and efficiency would be severely diminished. It is possible to identify fragile brain areas and establish a connection between functional deficiencies and neuroanatomical abnormalities by analyzing the profile of hub distribution and changes in network properties associated with these hub nodes (36). For nodal topological network parameters, BC is used to measure the number of times a brain node appears on the shortest path between any two other pairs of brain nodes (37). In other words, the brain nodes with enormous BC values are judged to be hubs of the network (38). Hence, BC acts as a vital control of information transit and is mainly used to measure the degree to which a node plays a “bridge” role in a brain network. Under this framework, we found that early-stage PD patients with a higher burden of autonomic symptoms exhibited decreased BC values in the right thalamus proper and increased BC values in the cerebellar vermal lobules VI–VII compared to those with a lower burden of autonomic symptoms. Thalamus proper, which encompasses all thalamic nuclei except for the lateral and medial geniculate bodies (39), serves as a central “bridge” for peripheral homeostatic information ascending to the cortex and for bidirectional thalamic-cortical (40,41) and thalamic-subcortical (40) communication. Beyond being a bridge, the thalamus actively and dynamically gates salient inputs by minimizing the materiality of currently irrelevant inputs, thus playing a crucial role in information integration (40). Emerging studies uncovered that the integrity and connectivity of the thalamus might be correlated with autonomic function (42-44). Valenza et al. combined functional MRI with instantaneous autonomic outflow estimates and characterized the extensive CAN in a resting state, including ventrolateral posterior thalamic nuclei (45). De Looze et al. found that delayed heart rate recovery after an orthostatic test is intimately linked to smaller thalamic volume (43). Another study suggested that patients with neurogenic orthostatic hypotension had weakened activation in the bilateral thalamus during Valsalva maneuver (46). Accordingly, the detected decreased BC values of the right thalamus proper in the early-stage PD patients with a higher burden of autonomic symptoms might reflect the collapse of the thalamic ‘bridge’. We speculated that the thalamus lost its crucial function in the regulation of CAN balance and resource allocation, leading to autonomic dysfunction in PD. However, this inference needs to be substantiated in future cohort studies.

The cerebellum, profiting from its dense projections with key CAN brain regions (such as the prefrontal lobe, cingulate gyrus, thalamus, and nucleus tractus solitarius), plays a crucial role in the autonomic function, including cardiovascular responses, vesical function, postural control of blood pressure and heart rate, and vestibule-sympathetic reflexes (47-50). In PD, the cerebellum has been gradually acknowledged as an important structure for explaining both motor and non-motor symptoms (51) due to the observation of extensive alpha-synuclein aggregates and Lewy bodies formation as in other CAN regions (52,53). Iniguez et al. discovered that the clinical severity of dysautonomia was correlated with the presence of a functional break at rest between the chronotropic cardiac function and the activity of CAN, including the cerebellum cortex, in PD patients (54). Chung et al. found that the decreased function of the cerebellum in PD patients increased the severity of autonomic dysfunction (14). Meanwhile, a recent study with ¹⁸F-fluoropropyl-2β-carbomethoxy-3β-4-iodophenyl nortropane positron emission tomography (¹⁸F-FP-CIT PET) uncovered that the de novo PD patients with dysautonomia had increased regional perfusion in the pontocerebellar area (55). Hence, based on the above neuroimaging evidence, we speculated that there might be abnormal functional and metabolic disorders in the cerebellum related to autonomic dysfunction in PD. Accordingly, our study, from a perspective of large-scale structural networks, observed that the hub role of the cerebellar vermal lobules VI–VII was significantly enhanced in the early-stage PD patients with a higher burden of autonomic symptoms. Regarding the vermis of the cerebellum, existing research has mostly discovered an intimate junction with the autonomic nervous regulation of the heart and blood pressure (48,50). Considering that the presence of cardiovascular dysautonomia is mainly characterized by orthostatic hypotension in PD patients (56), the enhanced hub role of the cerebellar vermal lobules VI–VII might be a central compensation for peripheral sympathetic nerve injury (27,56) and an impaired ‘bridge’ role in the right thalamus.

The Ne characterizes the capacity of parallel information processing for local nodes of GMV-based SCN (57). In this study, significantly increased Ne values located in the left middle temporal gyrus were detected in early-stage PD patients with a higher burden of autonomic symptoms, indicating abnormally elevated nodal information processing efficiency. Accumulated evidence suggests that the temporal lobe is also an indispensable part of the CAN (14,58,59). An intriguing piece of evidence suggested that autonomic dysfunction in obstructive sleep apnea patients was associated with impaired integrity of WM fiber bundles in the temporal lobe (58). In the prodromal phase of PD, Li et al. observed reduced functional connectivity between the temporal lobe and brainstem, which was significantly related to the severity of autonomic dysfunction (59). Another neuroimaging study showed extensive temporal-frontal WM network collapse in de novo PD patients with autonomic dysfunction (14). Hence, combined with our study, we inferred that the abnormal structure and functional network in the temporal lobe might underlie the central nervous mechanisms of autonomic dysfunction in PD. On the one hand, the increased Ne in the left temporal gyrus in the GMV-based SCN network could be a compensation for the breakdown of WM and functional networks in PD patients with autonomic dysfunction. On the other hand, considering that restoring high-frequency electrical activity homeostasis in the temporal lobe could improve autonomic regulation in postural orthostatic tachycardia syndrome (60), the left temporal gyrus might transmit excessive information flow through the temporal-frontal and temporal-brainstem pathways, disrupting CAN homeostasis and causing autonomic dysfunction in PD.

In our study, we also found that PD patients in the AUT_high group exhibited higher MDS-UPDRS-III scores than those in the AUT_low group, indicating that early-stage PD patients with a higher burden of autonomic symptoms had more severe motor symptoms. Consistent with our findings, a previous study found that motor symptoms and pathological burden were more severe in de novo PD patients with autonomic dysfunction than those without autonomic dysfunction (55). It should be pointed out that we used the UPDRS-III score as a covariate to correct the results during the GMV-based SCN network analysis. Therefore, the malformation in the CAN that we found was not implicated by the interference of brain regions related to PD motor symptoms.

Several limitations of our study should be acknowledged. First, the autonomic symptoms were evaluated by a screening questionnaire called SCOPA-AUT rather than quantitative autonomic tests since the data we had analyzed originated from PPMI database. However, the SCOPA-AUT scale has been validated as a dependable and accurate measure for autonomic dysfunction in PD (22). Second, a group of healthy controls was not included in our study. Considering that grouping was based on the quartiles of SCOPA-AUT scale in PD patients which cannot be directly used for the selection of healthy controls, we referred to the previous research (11) and did not include healthy controls. Moreover, the inclusion of patients with other central nervous system diseases with autonomic dysfunction is helpful for our understanding of autonomic dysfunction in PD, and future studies need to explore it. Third, the GMV-based SCN and network parameters were established at the group level and therefore the relationship between SCOPA-AUT scores and network parameters could not be analyzed. Nevertheless, the construction of SCN provided us with the possibility to explore the CAN characteristics of autonomic dysfunction in PD from the perspective of large-scale GM networks. Fourth, our cross-sectional study limited the ability to portray the central pathogenesis of PD patients with autonomic dysfunction using the topology attributes of the GMV-based SCN network. Fifth, PPMI is a multicenter cohort study having differences in MRI acquisition. Despite our efforts to standardize the acquisition equipment upon enrollment, we still cannot avoid discrepancies in parameters such as TE and TR. Finally, although no statistical difference existed, the age of the AUT_high group was higher than that of the AUT_low group, potentially introducing interference in the analysis of SCN. Nevertheless, we had incorporated age as a covariate in the construction of the SCN network to mitigate its influence as much as possible. Future investigation into the neuroimaging relationship with dysautonomia in PD is warranted and should incorporate objective quantification of autonomic nerve function and longitudinal sequential assessments.

Conclusions

This study demonstrates abnormal global and local topological properties in the GMV-based SCN of early-stage PD patients with autonomic dysfunction. We offered novel network-level evidence for elucidating the pathophysiological mechanisms of dysautonomia in early PD. Neuroimaging studies would be an effective tool to explore the central neural substrates involved in processing autonomic function in early-stage PD.

Acknowledgments

Data used in the preparation of this article were obtained from the Parkinson’s Progression Markers Initiative database (https://ida.loni.usc.edu/home/projectPage.jsp?project=PPMI). For up-to-date information on the study, visit www.ppmi-info.org. PPMI, a public-private partnership, is funded by the Michael J. Fox Foundation for Parkinson’s Research and funding partners, including Abbott, Avid Radiopharmaceuticals, Biogen Idec, Bristol-Myers Squibb, Covance, lan, GE Healthcare, Genentech, GSK-GlaxoSmithKline, Lundbeck, Lilly, Merck, MSD-Meso Scale Discovery, Pfizer, Piramal, Roche, Servier, and UCB (www.ppmi-info.org/fundingpartners).

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-2310/rc

Funding: This work was funded by the National Natural Science Foundation of China (No. 82271273), the Jiangsu Social Development Project (No. BE2022808), and Postgraduate Research & Practice Innovation Program of Jiangsu Province (No. SJCX24_0769).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-2310/coif). All authors report the funding from the National Natural Science Foundation of China (No. 82271273), the Jiangsu Social Development Project (No. BE2022808), and Postgraduate Research & Practice Innovation Program of Jiangsu Province (No. SJCX24_0769). The authors have no other 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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

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. Schapira AHV, Chaudhuri KR, Jenner P. Non-motor features of Parkinson disease. Nat Rev Neurosci 2017;18:509. [Crossref] [PubMed]
2. Kim JB, Kim BJ, Koh SB, Park KW. Autonomic dysfunction according to disease progression in Parkinson's disease. Parkinsonism Relat Disord 2014;20:303-7. [Crossref] [PubMed]
3. Pfeiffer RF. Autonomic Dysfunction in Parkinson's Disease. Neurotherapeutics 2020;17:1464-79. [Crossref] [PubMed]
4. Asahina M, Vichayanrat E, Low DA, Iodice V, Mathias CJ. Autonomic dysfunction in parkinsonian disorders: assessment and pathophysiology. J Neurol Neurosurg Psychiatry 2013;84:674-80. [Crossref] [PubMed]
5. Park JW, Okamoto LE, Kim SH, Lee CN, Park KW, Baek SH, Sung JH, Jeon N, Koh SB, Gamboa A, Shibao CA, Diedrich A, Kim BJ, Biaggioni I. Sympathetic dysfunction as an early indicator of autonomic involvement in Parkinson's disease. Clin Auton Res 2024;34:269-79. [Crossref] [PubMed]
6. Djaldetti R, Lev N, Melamed E. Lesions outside the CNS in Parkinson's disease. Mov Disord 2009;24:793-800. [Crossref] [PubMed]
7. Chang YT, Lu CH, Wu MK, Hsu SW, Huang CW, Chang WN, Lien CY, Lee JJ, Chang CC. Salience Network and Depressive Severities in Parkinson's Disease with Mild Cognitive Impairment: A Structural Covariance Network Analysis. Front Aging Neurosci 2017;9:417. [Crossref] [PubMed]
8. Chen YS, Chen HL, Lu CH, Lee CY, Chou KH, Chen MH, Yu CC, Lai YR, Chiang PL, Lin WC. The corticolimbic structural covariance network as an early predictive biosignature for cognitive impairment in Parkinson's disease. Sci Rep 2021;11:862. [Crossref] [PubMed]
9. Zhou C, Gao T, Guo T, Wu J, Guan X, Zhou W, Huang P, Xuan M, Gu Q, Xu X, Xia S, Kong D, Wu J, Zhang M. Structural Covariance Network Disruption and Functional Compensation in Parkinson's Disease. Front Aging Neurosci 2020;12:199. [Crossref] [PubMed]
10. Liou LM, Ruge D, Chang YP, Wu MN, Hsu CY, Lin CW, Tsai CL, Lai CL. Functional connectivity between lateral premotor-parietal circuits and the cardiac autonomic system in Parkinson's disease. J Neurol Sci 2013;326:48-52. [Crossref] [PubMed]
11. Dayan E, Sklerov M, Browner N. Disrupted hypothalamic functional connectivity in patients with PD and autonomic dysfunction. Neurology 2018;90:e2051-8. [Crossref] [PubMed]
12. Lin WC, Chen PC, Huang CC, Tsai NW, Chen HL, Wang HC, Chou KH, Chen MH, Chen YW, Lu CH. Autonomic Function Impairment and Brain Perfusion Deficit in Parkinson's Disease. Front Neurol 2017;8:246. [Crossref] [PubMed]
13. Kim JB, Kim H, Lee CN, Park KW, Kim BJ. Regional Gray Matter Volume Changes in Parkinson's Disease with Orthostatic Hypotension. Brain Sci 2021;11:294. [Crossref] [PubMed]
14. Chung SJ, Bae YJ, Jun S, Yoo HS, Kim SW, Lee YH, Sohn YH, Lee SK, Seong JK, Lee PH. Dysautonomia is associated with structural and functional alterations in Parkinson disease. Neurology 2019;92:e1456-67. [Crossref] [PubMed]
15. Kuo CY, Lee PL, Hung SC, Liu LK, Lee WJ, Chung CP, Yang AC, Tsai SJ, Wang PN, Chen LK, Chou KH, Lin CP. Large-Scale Structural Covariance Networks Predict Age in Middle-to-Late Adulthood: A Novel Brain Aging Biomarker. Cereb Cortex 2020;30:5844-62. [Crossref] [PubMed]
16. Yan S, Lu J, Li Y, Tian T, Zhou Y, Zhu H, Qin Y, Zhu W. Impaired topological properties of cortical morphological brain networks correlate with motor symptoms in Parkinson's disease. J Neuroradiol 2024;51:101155. [Crossref] [PubMed]
17. Li M, Li X, Das TK, Deng W, Li Y, Zhao L, Ma X, Wang Y, Yu H, Meng Y, Wang Q, Palaniyappan L, Li T. Prognostic Utility of Multivariate Morphometry in Schizophrenia. Front Psychiatry 2019;10:245. [Crossref] [PubMed]
18. Wen B, Xu Y, Fang K, Guo HR, Liu H, Liu L, Wei Y, Zhang Y, Cheng J, Han S. Gray matter morphological abnormities are constrained by normal structural covariance network in OCD. Prog Neuropsychopharmacol Biol Psychiatry 2024;129:110884. [Crossref] [PubMed]
19. Zhang Y, Qiu T, Yuan X, Zhang J, Wang Y, Zhang N, Zhou C, Luo C, Zhang J. Abnormal topological organization of structural covariance networks in amyotrophic lateral sclerosis. Neuroimage Clin 2019;21:101619. [Crossref] [PubMed]
20. Wang E, Jia Y, Ya Y, Xu J, Mao C, Luo W, Fan G, Jiang Z. Abnormal Topological Organization of Sulcal Depth-Based Structural Covariance Networks in Parkinson's Disease. Front Aging Neurosci 2020;12:575672. [Crossref] [PubMed]
21. The Parkinson Progression Marker Initiative (PPMI). Prog Neurobiol 2011;95:629-35. [Crossref] [PubMed]
22. Visser M, Marinus J, Stiggelbout AM, Van Hilten JJ. Assessment of autonomic dysfunction in Parkinson's disease: the SCOPA-AUT. Mov Disord 2004;19:1306-12. [Crossref] [PubMed]
23. Rodriguez-Blazquez C, Forjaz MJ, Frades-Payo B, de Pedro-Cuesta J, Martinez-Martin P. Longitudinal Parkinson's Disease Patient Study, Estudio Longitudinal de Pacients con Enfermedad da Parkinson Group. Independent validation of the scales for outcomes in Parkinson's disease-autonomic (SCOPA-AUT). Eur J Neurol 2010;17:194-201. [Crossref] [PubMed]
24. Bostantjopoulou S, Katsarou Z, Danglis I, Karakasis H, Milioni D, Falup-Pecurariu C. Self-reported autonomic symptoms in Parkinson's disease: properties of the SCOPA-AUT scale. Hippokratia 2016;20:115-20.
25. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 2005;53:695-9. [Crossref] [PubMed]
26. Benarroch EE. The central autonomic network: functional organization, dysfunction, and perspective. Mayo Clin Proc 1993;68:988-1001. [Crossref] [PubMed]
27. Chen Z, Li G, Liu J. Autonomic dysfunction in Parkinson's disease: Implications for pathophysiology, diagnosis, and treatment. Neurobiol Dis 2020;134:104700. [Crossref] [PubMed]
28. Wakabayashi K, Takahashi H. Neuropathology of autonomic nervous system in Parkinson's disease. Eur Neurol 1997;38:2-7. [Crossref] [PubMed]
29. Braak H, Del Tredici K, Rüb U, de Vos RA, Jansen Steur EN, Braak E. Staging of brain pathology related to sporadic Parkinson's disease. Neurobiol Aging 2003;24:197-211. [Crossref] [PubMed]
30. Papapetropoulos S, Mash DC. Insular pathology in Parkinson's disease patients with orthostatic hypotension. Parkinsonism Relat Disord 2007;13:308-11. [Crossref] [PubMed]
31. Bassett DS, Bullmore ET. Small-World Brain Networks Revisited. Neuroscientist 2017;23:499-516. [Crossref] [PubMed]
32. Latora V, Marchiori M. Efficient behavior of small-world networks. Phys Rev Lett 2001;87:198701. [Crossref] [PubMed]
33. Xu J, Zhang J, Zhang J, Wang Y, Zhang Y, Wang J, Li G, Hu Q, Zhang Y. Abnormalities in Structural Covariance of Cortical Gyrification in Parkinson's Disease. Front Neuroanat 2017;11:12. [Crossref] [PubMed]
34. Suo XS, Lei DL, Li LL, Li WL, Dai JD, Wang SW, He MH, Zhu HZ, Kemp GJK, Gong QG. Psychoradiological patterns of small-world properties and a systematic review of connectome studies of patients with 6 major psychiatric disorders. J Psychiatry Neurosci 2018;43:427. [Crossref] [PubMed]
35. Li J, Chen H, Fan F, Qiu J, Du L, Xiao J, Duan X, Chen H, Liao W. White-matter functional topology: a neuromarker for classification and prediction in unmedicated depression. Transl Psychiatry 2020;10:365. [Crossref] [PubMed]
36. Yuan W, Wade SL, Babcock L. Structural connectivity abnormality in children with acute mild traumatic brain injury using graph theoretical analysis. Hum Brain Mapp 2015;36:779-92. [Crossref] [PubMed]
37. Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage 2010;52:1059-69. [Crossref] [PubMed]
38. Odish OF, Caeyenberghs K, Hosseini H, van den Bogaard SJ, Roos RA, Leemans A. Dynamics of the connectome in Huntington's disease: A longitudinal diffusion MRI study. Neuroimage Clin 2015;9:32-43. [Crossref] [PubMed]
39. Seidman LJ, Faraone SV, Goldstein JM, Goodman JM, Kremen WS, Toomey R, Tourville J, Kennedy D, Makris N, Caviness VS, Tsuang MT. Thalamic and amygdala-hippocampal volume reductions in first-degree relatives of patients with schizophrenia: an MRI-based morphometric analysis. Biol Psychiatry 1999;46:941-54. [Crossref] [PubMed]
40. Wolff M, Vann SD. The Cognitive Thalamus as a Gateway to Mental Representations. J Neurosci 2019;39:3-14. [Crossref] [PubMed]
41. Sherman SM, Guillery RW. The role of the thalamus in the flow of information to the cortex. Philos Trans R Soc Lond B Biol Sci 2002;357:1695-708. [Crossref] [PubMed]
42. Reinelt J, Uhlig M, Müller K, Lauckner ME, Kumral D, Schaare HL, Baczkowski BM, Babayan A, Erbey M, Roebbig J, Reiter A, Bae YJ, Kratzsch J, Thiery J, Hendler T, Villringer A, Gaebler M. Acute psychosocial stress alters thalamic network centrality. Neuroimage 2019;199:680-90. [Crossref] [PubMed]
43. De Looze C, Williamson W, Hirst R, O'Connor J, Knight S, McCrory C, Carey D, Kenny RA. Impaired orthostatic heart rate recovery is associated with smaller thalamic volume: Results from The Irish Longitudinal Study on Aging (TILDA). Hum Brain Mapp 2020;41:3370-8. [Crossref] [PubMed]
44. Rubio A, Van Oudenhove L, Pellissier S, Ly HG, Dupont P, Lafaye de Micheaux H, Tack J, Dantzer C, Delon-Martin C, Bonaz B. Uncertainty in anticipation of uncomfortable rectal distension is modulated by the autonomic nervous system--a fMRI study in healthy volunteers. Neuroimage 2015;107:10-22. [Crossref] [PubMed]
45. Valenza G, Sclocco R, Duggento A, Passamonti L, Napadow V, Barbieri R, Toschi N. The central autonomic network at rest: Uncovering functional MRI correlates of time-varying autonomic outflow. Neuroimage 2019;197:383-90. [Crossref] [PubMed]
46. Baker J, Paturel JR, Kimpinski K. Impaired cortical autonomic responses during sympathetic activation in neurogenic orthostatic hypotension characterized by postganglionic autonomic dysfunction. J Appl Physiol (1985) 2018;125:1210-7. [Crossref] [PubMed]
47. Sakakibara R. The cerebellum seems not a 'little brain' for the autonomic nervous system. Clin Neurophysiol 2019;130:160. [Crossref] [PubMed]
48. Critchley HD, Corfield DR, Chandler MP, Mathias CJ, Dolan RJ. Cerebral correlates of autonomic cardiovascular arousal: a functional neuroimaging investigation in humans. J Physiol 2000;523:259-70. [Crossref] [PubMed]
49. Baker J, Paturel JR, Kimpinski K. Cerebellar impairment during an orthostatic challenge in patients with neurogenic orthostatic hypotension. Clin Neurophysiol 2019;130:189-95. [Crossref] [PubMed]
50. Baker J, Kimpinski K. Evidence of Impaired Cerebellar Connectivity at Rest and During Autonomic Maneuvers in Patients with Autonomic Failure. Cerebellum 2020;19:30-9. [Crossref] [PubMed]
51. Wu T, Hallett M. The cerebellum in Parkinson's disease. Brain 2013;136:696-709. [Crossref] [PubMed]
52. Seidel K, Bouzrou M, Heidemann N, Krüger R, Schöls L, den Dunnen WFA, Korf HW, Rüb U. Involvement of the cerebellum in Parkinson disease and dementia with Lewy bodies. Ann Neurol 2017;81:898-903. [Crossref] [PubMed]
53. Coon EA, Cutsforth-Gregory JK, Benarroch EE. Neuropathology of autonomic dysfunction in synucleinopathies. Mov Disord 2018;33:349-58. [Crossref] [PubMed]
54. Iniguez M, Jimenez-Marin A, Erramuzpe A, Acera M, Tijero B, Murueta-Goyena A, Del Pino R, Fernandez T, Carmona-Abellan M, Cabrera-Zubizarreta A, Gómez-Esteban JC, Cortes JM, Gabilondo I. Heart-brain synchronization breakdown in Parkinson's disease. NPJ Parkinsons Dis 2022;8:64. [Crossref] [PubMed]
55. Shin HW, Chung SJ, Lee S, Cha J, Sohn YH, Yun M, Lee PH. Dysautonomia Is Linked to Striatal Dopamine Deficits and Regional Cerebral Perfusion in Early Parkinson Disease. Clin Nucl Med 2020;45:e342-8. [Crossref] [PubMed]
56. Palma JA, Kaufmann H. Orthostatic Hypotension in Parkinson Disease. Clin Geriatr Med 2020;36:53-67. [Crossref] [PubMed]
57. Huang Y, Zhang X, Cheng M, Yang Z, Liu W, Ai K, Tang M, Zhang X, Lei X, Zhang D. Altered cortical thickness-based structural covariance networks in type 2 diabetes mellitus. Front Neurosci 2024;18:1327061. [Crossref] [PubMed]
58. Chen HL, Huang CC, Lin HC, Lu CH, Chen PC, Chou KH, Su MC, Friedman M, Lin CP, Lin WC. White matter alteration and autonomic impairment in obstructive sleep apnea. J Clin Sleep Med 2020;16:293-302. [Crossref] [PubMed]
59. Li G, Chen Z, Zhou L, Zhao A, Niu M, Li Y, Luo N, Kang W, Liu J. Altered structure and functional connectivity of the central autonomic network in idiopathic rapid eye movement sleep behaviour disorder. J Sleep Res 2021;30:e13136. [Crossref] [PubMed]
60. Fortunato JE, Tegeler CL, Gerdes L, Lee SW, Pajewski NM, Franco ME, Cook JF, Shaltout HA, Tegeler CH. Use of an allostatic neurotechnology by adolescents with postural orthostatic tachycardia syndrome (POTS) is associated with improvements in heart rate variability and changes in temporal lobe electrical activity. Exp Brain Res 2016;234:791-8. [Crossref] [PubMed]