Altered dynamic functional network connectivity in levodopa‐induced dyskinesia of Parkinson's disease

Abstract Aims The aim of this study was to clarify the dynamic neural activity of levodopa‐induced dyskinesia (LID) in Parkinson's disease (PD). Methods Using dynamic functional network connectivity (dFNC) analysis, we evaluated 41 PD patients with LID (LID group) and 34 PD patients without LID (No‐LID group). Group spatial independent component analysis and sliding‐window approach were employed. Moreover, we applied a k‐means clustering algorithm on windowed functional connectivity (FC) matrices to identify reoccurring FC patterns (i.e., states). Results The optimal number of states was determined to be five, the so‐called State 1, 2, 3, 4, and 5. In ON phase, compared with No‐LID group, LID group occurred more frequently and dwelled longer in strongly connected State 1, characterized by strong positive connections between visual network (VIS) and sensorimotor network (SMN). When switching from OFF to ON phase, LID group occurred less frequently in State 3 and State 4. Meanwhile, LID group dwelled longer in State 2 and shorter in State 3. No‐LID group occurred more frequently in State 5 and less frequently in State 3. Additionally, correlation analysis demonstrated that dyskinesia's severity was associated with frequency of occurrence and dwell time in State 2, dominated by inferior frontal cortex in cognitive executive network (CEN). Conclusion Using dFNC analysis, we found that dyskinesia may be related to the dysfunctional inhibition of CEN on motor loops and excessive excitation of VIS and SMN, which provided evidence of the changes in brain dynamics associated with the occurrence of dyskinesia.


| INTRODUC TI ON
In Parkinson's disease (PD), levodopa treatments could alleviate motor symptoms. But after 4-6 years, levodopa-induced dyskinesia (LID), a disabling motor complication, can be induced in 40% of PD patients. 1,2 Idiopathic PD is considered as a disorder of response initiation characterized by excessive motor inhibition (i.e., akinesia and bradykinesia), while LID is primarily a clinical manifestation of disinhibition of movements. 3 Previous neuroimaging studies on LID have discovered cortical morphological and functional alterations in inferior frontal cortex (IFC), one fundamental component of executive control or motor inhibition network. [3][4][5][6][7][8] However, the neural mechanisms behind this motor disorder remain poorly understood.
Traditional static functional connectivity (FC) on LID was measured with the assumption that intrinsic fluctuations during magnetic resonance imaging (MRI) scan were stationary. 3 Nevertheless, recent studies have proposed that FC is dynamic or fluctuating within seconds to minutes, which highlights the need for detailed inspection of FC along discrete time windows. [9][10][11] Emerging data suggested the utility of dynamic functional network connectivity (dFNC) with sliding-window analysis for understanding functional neurodevelopment as well as PD pathogenesis. [12][13][14][15][16][17][18] Time-varying FC may reflect implied spontaneous changes in underlying networks, which might improve our knowledge of how neural systems flexibly coordinate to support cognitive and behavioral function.
In addition, the interactions between large-scale brain networks in LID have not been examined in previous studies. The exploration for dynamic internetwork connectivity can provide insights into the abnormal functional integration and specialization of the brain in LID of PD. Therefore, with dFNC analysis, this study was conducted to investigate the possible neural mechanisms behind LID. To be specific, we explored how large-scale functional network interactions changed dynamically in temporal domains (fractional windows, dwell time, and number of transitions) of PD patients with and without LID when levodopa worked or did not work. We hypothesized that compared with those without LID, PD patients with LID would display significantly different temporal properties in some specific brain states and help us understand the neural mechanism of dyskinesia.

| Participants
In this study, 90 PD patients were consecutively recruited, of whom 50 were clinically diagnosed as LID (LID group), and 40 were patients without LID (No-LID group). All participants were corresponding to the United Kingdom Parkinson's Disease Society Brain Bank diagnostic criteria for idiopathic PD. 19 Inclusion criteria were as follows: (1) a minimum 6 months duration of levodopa therapy; (2) presence or absence of LID after an acute levodopa test observed by the neurologist at the last examination; (3) stable levodopa medications dose for 1 month; (4) right handedness; (5) no use of anxiolytic, antidepressant, or antipsychotic drugs; (6) no evidence of severe cognitive impairment, especially dementia [mini-mental state examination (MMSE) score > 24]; (7) no contraindications for MRI; (8) no evidence of brain tumor, vascular brain lesions, or brain atrophy; (9) no excessive head movements during the MRI scan (see below); and (10) the ability to tolerate 12 h withdrawal of dopaminergic medications before MRI session.
In LID group, all patients presented with peak-dose dyskinesia rather than diphasic dyskinesia or off-period dyskinesia. All participants gave their written informed consents. The research was approved by the ethics committee of the First Affiliated Hospital of Nanjing Medical University and completed in line with the Declaration of Helsinki.

| MRI data acquisition
Images were obtained on a 3.0 T Siemens MRI system (Siemens Medical Solutions). Structural 3D T1-weighted high-resolution images were obtained using the following sequence (repetition time    x, y, or z was higher than 3 mm or in rotation indexes was higher than 3°. 14,22-24

| Group independent component analysis
To create intrinsic networks, we performed spatial group independent component analysis (GICA) implemented in the GIFT toolbox (GIFT v4.0b; http://icatb.sourc eforge.net). Subject-specific principal component analysis was applied to reduce data into 50 independent components (ICs), with the minimum description length criterion. 15 The stability and validity of ICs was ensured by repeating the ICA algorithm 20 times in ICASSO. 25,26 Subject-specific spatial maps and time courses of each IC were created by the GICA backreconstruction algorithm. 27 Of the 50 extracted ICs, relevant intrinsic connectivity networks were identified based on the criteria from Allen. 9 Firstly, we manually confirmed whether the peak activation clusters of spatial maps were located primarily in gray matter, no spatial overlap with vascular, ventricular, or susceptibility artifacts.
Then, we selected ICs with time courses dominated mainly by low-frequency fluctuations (ratio of power below 0.1 Hz to 0.15-0.25 Hz), and with a high dynamic range in spectrum. This procedure resulted in 22 meaningful ICs, which were sorted into seven networks according to the spatial correlation values between ICs and the resting-state networks template. 12,14,28 As shown in Figure 1A and Table 2

| Dynamic functional connectivity computation
The dFNC analysis was examined by applying two approaches: a sliding-window approach and k-means clustering. A sliding-window approach was adopted to estimate changes of FC over time, and the resulting windowed segments were used to calculate transient FC patterns. In our study, data were segmented into a sliding window of 30 TR, 30-33 with a Gaussian window alpha value of 3 TR and a step size between windows of 1 TR. To promote sparsity, a penalty on the L1 norm was imposed utilizing the graphic LASSO framework by 100 repetitions. 34 To extract reoccurring FC patterns or clusters (also described as states), we applied k-means clustering algorithm and iterated it 500 times. The similarity between window FC matrix and k-means cluster centroids was calculated using the L1 distance. 9 Based on the elbow criterion, 9,17 the optimal number of cluster centroids (i.e., states) was determined to be five (k = 5), the so-called State 1, 2, 3, 4, and 5.
According to the similarity with obtained five cluster centroids, all FC matrices of each subject were classified into one of the five states.

| Intrinsic connectivity networks
Spatial maps of all 22 ICs are shown in Figure 1A.  Figure 1B displayed the static FC. The detailed information of ICs is presented in Table 2.

| Clustering analysis and dFNC patterns
Utilizing k-means clustering, we identified five FC matrices states over the entire MRI scans. As shown in Figure 2A, the percentages of total occurrences of these states were a bit different, with State 5 more frequent (42%) than State 1 (12%), State 2 (13%), State 3 (11%), or State 4 (22%). To better visualize the patterns of outstanding FC differences between states, we kept connections with the most 5% strength of each state ( Figure 2B). 14,15 State 1-5 shared the similarity in within-network connections of SMN and VIS, while they still possessed some unique between-network interconnection patterns. State

| Correlations between dFNC and the severity of dyskinesia in LID in ON phase
After controlling for age at onset, LEDD, and disease duration, we found that fractional windows and dwell time in State 2 of LID group were respectively positively correlated with AIMS score in ON phase

| DISCUSS ION
In this study, dFNC analysis was used to explore the differences of dy- In this work, we did not focus on definite brain areas of interest as in previous LID studies, 3 13 and Alzheimer's disease. 16 These findings implied that dFNC was a promising approach for clinical neuroimaging and might provide greater insights into the neural mechanism of LID.
In this study, no significant differences of temporal properties were found between PD subgroups in OFF phase. In ON phase, However, the exact mechanism of LID needs to be confirmed by further research.
The pathogenesis of LID is still not clear, involving the dopaminergic and serotonergic system. 44,45 It is believed that aberrant do- in key interactive functional networks for coordinating motor function. 51 It was reported that patients with LID were profoundly relevant to the higher impulsivity score and lower inhibitory control than those without. 52,53 The SMN is involved in sensory perception and motor process. 54,55 Studies identified that the disinhibition of cortical-subcortical circuits in the SMN may contribute to the abnormal involuntary movements. 55 37 the disease duration did not match well in our study, so we have considered this as one of covariates in further analysis to mitigate its effect. Third, to ensure the security of patients and avoid the impact of head movements on image quality, we began fMRI scanning for those patients with LID before levodopa reached its peak in ON phase, which may not fully reflect the functional state of brain during the peak of dose. Last but not least, the data for duration of dyskinesia had not been collected.

FU N D I N G I N FO R M ATI O N
This study was funded by the National Natural Science Foundation of China (grant numbers 81671258 and 81901297).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restriction.