Altered cerebellum functional network on newly diagnosed drug-naïve Parkinson’s disease patients with anxiety

Abbreviations

ACC

anterior cingulate cortex

ANOVA

one-way analysis of variance

FC

functional connectivity

GRF

Gaussian random field

HAMA

Hamilton anxiety rating scale

HAMD

Hamilton depression rating scale

H&Y

Hoehn & Yahr

MMSE

mini-mental state examination

MoCA

montreal cognitive assessment

MRI

magnetic resonance imaging

NCs

normal controls

NMS

non-motor symptom

NMSS

non-motor symptoms scale

PD

Parkinson’s disease

PD-A

PD patients with anxiety symptoms

PD-NA

PD patients without anxiety symptoms

ReHo

regional homogeneity

rs-fMRI

rest-state functional magnetic resonance imaging

ROI

region of interest

SPSS

statistical product and service solutions

UPDRS-III

unified PD rating scale part III

VMHC

voxel mirror homotopy connectivity

1 Introduction

Parkinson’s disease (PD), the second most prominent chronic progressive neurodegenerative disease in middle-aged and elderly people [1], has anxiety as one of its common neuropsychiatric symptoms. Studies have shown that up to 25–50% of the PD patients have varying degrees of anxiety [2]. It aggravates other motor symptoms, adversely affects the patient’s daily functioning and disease prognosis, reduces the patient’s quality of life, and increases the burden on the caregiver [3]. However, its insidious clinical symptoms make PD with anxiety difficult to identify and diagnose, and it has not received enough attention from researchers. Current neuromodulation system studies have suggested that striatum dopaminergic uptake levels [4], noradrenergic transmission [5], and serotonergic neuron systemic dysfunction [6] may be involved in the occurrence of PD with anxiety. Studies using magnetic resonance imaging (MRI) have found certain functional and structural changes in the cortical and subcortical areas of the brain of PD patients with anxiety, such as the gyrus return [7], the limbic system [8], and the cerebellum [9]. The neuronal circuit wiring plays a parallel role in anxiety symptoms in PD, in which the pathophysiological mechanism still has a large blind spot [10]. Therefore, characterizing the underlying neural mechanisms of patients with PD with anxiety symptoms using neuroimaging methods is urgent and imperative.

In PD patients, alteration of neuronal circuits such as subcortex-to-cortex pathways during neurodegeneration may explain the neuromechanism of anxiety symptoms [11]. Awareness has been increasing of a role for the cerebellum in affective behavior [12]; cerebellar dysfunction has been linked to motor symptoms in PD as well as to non-motor function, including anxiety symptoms. It is thus of direct interest to examine the connectivity within cerebellum and cortex to uncover neuroimaging evidence representative of anxiety symptoms in PD patients. This study’s aim was to explore aberrant functional connectivity (FC) throughout the cerebellar efferent pathways that project to the cerebral cortex using regional homogeneity (ReHo) and FC analysis among PD patients with anxiety symptoms.

2 Materials and methods

3 Results

3.2 ReHo analysis

The ANOVA showed that ReHo values of the left medial frontal gyrus and left cerebellum differed significantly among PD-A patients, PD-NA patients, and NCs (voxel level P < 0.001, cluster level P < 0.05, and GRF corrected). Further post hoc t comparison of each pair among the three groups showed that, compared with the PD-NA patients as well as the NCs, the PD-A patients had significantly decreased ReHo values in the left medial frontal gyrus and significantly increased ReHo values in the left cerebellum (Table 2, Figure 1).

Table 2

Comparison of ReHo differences among each pair among PD-A, PD-NA, and NCs

Anatomical region		Cluster size	MNI coordinates			Peak value (T value)
			X	Y	Z
PD-A vs PD-NA	Left Cerebellum	23	−27	−55	−54	4.166
	Left middle frontal gyrus	15	−47	35	21	−2.655
PD-A vs NCs	Left Cerebellum	29	−22	−61	−51	3.372
	Left middle frontal gyrus	18	−52	25	34	−4.852

PD-A, Parkinson’s Disease patients with anxiety; PD-NA, Parkinson’s Disease without anxiety; NCs, Normal controls. The positive or negative of T values represent increases and decreases in ReHo values.

Figure 1

ReHo changes based on one-way ANOVA (a) and the post hoc t comparison-independent samples (b and c). (a) (Column demonstrates regions with significantly different ReHo values showing left cerebellum (a) and the left medial frontal gyrus (b) among the PD-A, PD-NA, and NC groups). The images shown here from left to right are sagittal, coronal, and axial views, respectively. (b) Column demonstrates ReHo value comparison of left cerebellum between post hoc t comparison of each pair among the three groups. (c) Column demonstrates ReHo value comparison of left middle frontal gyrus between post hoc t comparison of each pair among the three groups. Red represented increased ReHo. L indicates the left side. The color bar shows F value of one-way ANOVA analysis. Statistical significance was set at voxel level P < 0.001 and cluster level P < 0.05 based on GRF correction. ^* P < 0.05 and ^** P < 0.001.

3.3 FC analysis

The left cerebellum, showing the maximum peak value of the ReHo changes, was selected as the ROI to research FC with other brain regions. The ANOVA revealed significant differences between the left cerebellum’s FC and that of the right cerebellum, right caudate nucleus, and right anterior cingulate cortex (ACC) among PD-A patients, PD-NA patients, and NCs (voxel level P < 0.001, cluster level P < 0.05, and GRF corrected). In post hoc t comparisons, PD-A FC values between the left and right cerebellum were decreased, whereas FC values between the left cerebellum and the right caudate nucleus and right ACC were increased compared with PD-NA FC values. Compared with NCs, the FC values between left and right cerebellum in PD-A were significantly lower (voxel level P < 0.001, cluster level P < 0.05, and GRF corrected) (Table 3, Figure 2).

Table 3

Comparison of FC values among each pair among PD-A, PD-NA and NCs

Anatomical region		Cluster size	MNI coordinates			Peak value (T value)
			X	Y	Z
PD-A vs PD-NA	Right cerebellum	15	6	−50	−32	−4.251
	Right caudate nucleus	16	15	18	−15	4.267
	Right ACC	12	8	37	−8	4.158
PD-A vs NCs	Right cerebellum	14	5	−47	−35	−4.506

PD-A, Parkinson’s disease patients with anxiety; PD-NA, Parkinson’s disease without anxiety; NCs, normal controls.

Figure 2

FC changes based on one-way ANOVA (a) and the post hoc t comparison-independent samples (b). (a) (Column demonstrates regions with significantly different FC with the left cerebellum showing right cerebellum (a), right caudate nucleus (b), and right ACC (c) among the PD-A, PD-NA, and NC groups). The image from left to right shows sagittal, coronal, and axial views, respectively. (b) Column demonstrates FC value comparison between post hoc t comparison of each pair among the three groups. Warm represented increase. L indicates the left side. The color bar shows F value of one-way ANOVA analysis. Statistical significance was set at voxel level P < 0.001 and cluster level P < 0.05 based on GRF correction. ^* P < 0.05 and ^** P < 0.001.

3.4 Partial correlation analysis

The partial correlation analysis revealed a negative correlation between HAMA scores and ReHo values in the left middle frontal gyrus of the PD-A patients (r = −0.503 and P = 0.019). FC values between the left and right cerebellum showed a significantly negative correlation with HAMA scores of PD-A patients (r = −0.488 and P = 0.024) (Figure 3). However, there was no significant correlation between UPDRS-III score, NMSS score, and imaging data (ReHo and FC) of altered brain areas in PD-A patients when correcting for age, gender, MMSE, and HAMD scores.

Figure 3

Partial correlation analysis in the PD-A between imaging data and HAMA scores.

4 Discussion

The present study investigated the neuroimaging mechanism of changes in neuronal functional activities and functional connections on newly diagnosed drug-naïve PD patients with anxiety symptoms based on ReHo and ROI-based FC analysis. Consequently, we found that ReHo values in the left cerebellum were significantly increased, while it was significantly decreased in the left middle frontal gyrus. In addition, the FC values of the left and right cerebellum were significantly decreased, while the FC of the left cerebellum and right caudate nucleus as well as right ACC were higher in PD patients with anxiety than in PD patients without anxiety. The partial correlation analysis demonstrated that significant anxiety symptoms in PD patients were negatively correlated with ReHo values in the left middle frontal gyrus and decreased FC values within the left and right cerebellum. This suggests that the underlying neuropathologic mechanism of PD with anxiety was related to the dysfunction of the local cerebellum network.

In this study, it was found that the ReHo value of the left medial frontal gyrus was reduced and that of the left cerebellum was significantly increased. The left medial frontal gyrus has a greater top-down loop regulation function [19] in the PD-A group than in the PD-NA group. This suggests that the neural circuits between the cerebellum and medial frontal gyrus were dysfunctional in the anxiety state. The middle frontal gyrus participates in advanced brain functions such as emotion, memory, and thinking and often plays an instrumental role in emotional control [20]. Neuroimaging studies suggest the involvement of a similar neural circuit in the development of anxiety disorders: the frontal-subcortical and the cortical-marginal pathway [19,21]. In addition, the reduction in prefrontal lobe activity has been observed in patients with anxiety, which through top-down attenuation inhibits cortical and subcortical structures, leading to emotional regulation imbalance, abnormalities in processing negative emotions such as anxiety and fear, and the occurrence of anxiety [19]. Reduction in ventral activity in the prefrontal cortex and enhancement of dorsal activity were reportedly associated with anxiety, suggesting that various subregions of the prefrontal cortex play different roles in anxiety [22]. However, the specific mechanism by which the cerebellum regulates anxiety through the middle frontal gyrus is not yet clear.

Our findings demonstrated that the spontaneous neuronal activity of the cerebellum was disrupted in PD patients with anxiety symptoms. It is well-known that the cerebellum participates in motor function; however, previous studies have confirmed that the cerebellum is closely associated with the cortical and subcortical brain regions associated with anxiety, such as the amygdala [23], prefrontal cortex [24], and cingulate gyrus [25]. Structural MRI studies found that the cerebellar tonsils and cerebellar right lobes in PD patients with anxiety had significantly reduced gray matter volume compared with those of healthy controls [26]. The cerebellum may be an important component of the neural network of anxiety [27]. The existing research suggests that the cerebellum not only participates in the regulation of muscle tension while coordinating autonomous activities but also finely adjusts the process of sensory information acquisition, which has been widely recognized. Our results showed that the ReHo value of the left cerebellum increased in the PD-A patient group, suggesting that the cerebellum is an important anatomical structure in anxiety regulation.

The results – increased connectivity within left cerebellum and right caudate nucleus as well as right ACC and decreased connectivity between the left cerebellum and right cerebellum – also suggested that there is imbalance in the local neural brain functional network of PD patients with anxiety symptoms and that the cerebellum may be an important component of the neural network of anxiety.

The ACC is involved in the formation of emotional networks and in the integration of negative emotions [28]. Whalen et al. found that the ACC is significantly activated after stimulating the emotion of disgust in generalized anxiety patients, which was related to a good treatment prognosis in anxiety patients with high ACC activity [29]. This is consistent with our study’s finding for PD patients with anxiety symptoms. Further speculation is that the ACC perceives involuntary nervous and tremor motility based on abnormal activities in the cerebellum, thereby strengthening the functional network connection with the cerebellum so as to regulate anxiety symptoms. However, both the specific mechanism of FC changes in the cerebellum and the involvement of ACC in brain functional networks with regard to PD anxiety symptoms need further study.

Similarly, we found enhanced FC between the left cerebellum and right caudate nucleus in PD-A patients, but no similar changes in functional connections were reported in the previous study. Nevertheless, it is well-known that as subcortical structures, the cerebellum and striatum affect motor symptoms and emotional dysfunction [30]. Previous research has found that the density of dopamine transporters was increased in the caudate nucleus of PD-A patients, which was significantly positively correlated with the severity of anxiety symptoms [31]. It is confirmed that the substantia nigra striatum pathway was related to PD anxiety symptoms. Thus, as the fundings in the study, in PD-A patients, we speculated that the abnormal functional connections of the cerebellum and caudate nucleus were related to the impaired integrity of the cerebellar–striatum nerve loop in PD patients with anxiety symptoms.

Similarly, we found the presence of decreased FC between left and right cerebellum in PD-A patients, but not in NCs and PD-NA patients and that it was negatively correlated with HAMA score. Voxel mirror homotopy connectivity (VMHC) analysis found that bilateral cerebellum VMHC value decreased and was negatively correlated with HAMA scores in patients with generalized anxiety disorder [32]. Overall, our finding suggests that PD anxiety may be associated with decreased synchronization of bilateral cerebellar functions. Geometrically isotopic brain region connectivity decreases as anxiety severity increases, and abnormalities in the mirror bilateral cerebellar functional connection network may be involved in the brain functional network mechanism of PD anxiety symptoms.

We all know that the changes in the functional connections between the amygdala, prefrontal lobe, and cerebellum can be widely involved in regulating anxiety symptoms [33]. The present study did not find abnormal amygdala function in PD anxiety patients, which may be due to the mild degree of PD anxiety symptoms of patients recruited in our study; and the compensatory enhancement of left cerebellum functional activities in the early stage and the functional changes in the amygdala were inhibited. In addition, it is important to note that the H&Y grade and UPDRS III score had no significant relationship with left cerebellum function connection, which suggested that movement had little influence on early PD patients with anxiety. Abnormal neural networks in the cerebellum may be the result of anxiety symptoms rather than movement.

In conclusion, all patients included in the study were newly diagnosed, drug-naïve PD patients, thus effectively reducing drug bias in the outcome. Anxiety symptoms are less affected by the course of disease, severity, or exercise symptoms in the early stage of PD. In summary, our study shows that anxiety symptoms in PD patients are associated with changes in neural functional activity in the bilateral cerebellum, left middle frontal gyrus, right caudate nucleus, and right ACC.

5 Conclusion

There is abnormal spontaneous electrical activity in the left cerebellum and left middle frontal gyrus. Changes in the functional connections between the left and right cerebellum and in the left cerebellum, the right caudate nucleus, and the right ACC may represent underlying mechanisms of PD with anxiety symptoms and may thus be important potential biomarkers for PD with anxiety symptoms. Nonetheless, further clarification of long-term neural implications of anxiety symptoms in PD is needed.