The main findings of the current study were, first, that in patients with PD, higher FC between two given brain regions is associated with stronger interregional homogeneity of dopamine and serotonin levels as assessed by 11C-PE2I and 11C-DASB PET. In other words, brain regions with high 11C-PE2I and 11C-DASB PET uptake are likely to have higher FC with regions showing similar high uptake. Conversely, brain regions with low 11C-PE2I and 11C-DASB PET uptake tend to have higher FC with regions exhibiting similarly low uptake. Second, we found that highly functionally connected brain regions show similar longitudinal changes in 11C-PE2I and 11C-DASB PET uptake in PD patients. Brain regions with greater longitudinal changes in PET uptake were preferentially connected to regions with comparable longitudinal changes in PET uptake, whereas brain regions with smaller longitudinal changes in PET uptake were highly connected to other regions with smaller longitudinal changes in PET uptake. Third, the association between PET uptake in target ROIs and their FC to the seed region was correlated with PD motor and non-motor severity across different brain regions that was dependent on the neurotransmitter system evaluated.
To the best of our knowledge, this is the first study to evaluate the association between the spatial distribution of dopamine and serotonin and FC in patients with PD. Our first finding confirms our hypothesis that dopamine and serotonin levels covary among brain regions that are highly functionally connected, suggesting that the distribution of neurotransmitters follows the organisational principles of the brain's functional connectomes. Notably, we observed a stronger correlation between FC and dopamine covariance in the subcortical network, especially in the basal ganglia nuclei. Conversely, in the visual network, the correlation was more pronounced between FC and serotonin covariance. This finding aligns with the fact that the dopamine depletion primarily occurs in subcortical regions in PD.43 Additionally, this corresponds with a previous study suggesting a significant role for serotonin in mediating visual hallucinations associated with cognitive dysfunction through the visual network in PD.44 These differential correlations across brain networks underscore the complexity of neurotransmitter interactions and their relationships with FC, highlighting that dopamine and serotonin play distinct roles in various brain networks.
Longitudinally, we observed that brain regions highly functionally connected at baseline displayed similar longitudinal changes in 11C-PE2I and 11C-DASB PET uptake at follow-up, echoing our cross-sectional findings in PD. Again, there were network-specific differences in the strength of correlations between FC and the change covariances of 11C-PE2I and 11C-DASB PET. Our findings suggest that the degeneration of the dopamine system over time has a stronger relationship with FC at the whole-brain level and in the brainstem network than does the serotonin system. This observation is consistent with the fact that dopamine depletion is thought to be central in PD, especially the progressive loss of dopaminergic neurons in the substantia nigra.45 Although the primary clinical focus has long been on motor symptoms, there is increasing recognition of the importance of studying non-motor symptoms in patients with PD.9,46 The stronger correlation between FC and 11C-DASB PET change covariance in the limbic and default mode networks underscores the role of the serotonin system in emotional and self-referential processing.47,48 This supports the idea that serotonin dysfunction in patients with PD is associated with the development of non-motor symptoms.49
A number of earlier studies have demonstrated that neighbouring brain regions often exhibit stronger intrinsic relationships.50–53 Salvador et al.54 found that adjacent brain regions show stronger FC than those farther apart. Similarly, Honey et al.55 noted a decline in FC as the interregional distance increases. Furthermore, structural connectivity (SC) also exhibits a distance dependence, where brain regions in close proximity are more likely to be connected, while those farther apart are less so.56,57 Therefore, when evaluating associations between distinct brain regions, it is crucial to account for physical distance to avoid mistakenly attributing spatial relationships to functional or neurotransmitter associations. Here, we found that the association between FC and PET covariance, as well as between FC and PET change covariance, remained significant even when controlling for the Euclidean distance between brain regions. This suggests that the distribution of neurotransmitter and FC patterns is intrinsically linked in the PD brain, independent of spatial proximity.
We found that the association between FC and PET uptake correlates with motor severity in patients with PD. This correlation was observed in more brain regions for 11C-PE2I PET compared to 11C-DASB PET, suggesting that these two radioligands differ in their sensitivities and specificities when assessing brain regions affected by PD. The correlations observed for both radioligands align with the fact that both the dopaminergic and serotonergic systems are affected in PD,58 though they vary in the degree to which they are disrupted. Our findings are consistent with the current understanding that the dopaminergic system plays a more significant role than the serotonergic system in the development of motor symptoms.59 However, it is worth mentioning that there was a lack of correlation between motor scales and FC-PET β-values for 11C-PE2I PET in the striatum. The possible reason for this lack of correlation is that the striatum, particularly the putamen, is that it is not a uniform structure; different parts of it are involved in different aspects of motor control. By subdividing the putamen into anterior or posterior regions, we might be able to achieve a more accurate correlation with motor scales. Furthermore, the observed correlation in more brain regions for the bradykinesia-rigidity sub-score than for the total UPDRS-III score suggests that bradykinesia and rigidity are more sensitive to alterations in neurotransmitter systems and functional connectomes in patients with PD. This finding is in agreement with our previous study, which showed that both striatal 11C-PE2I and 18F-DOPA uptakes are more strongly correlated with the bradykinesia-rigidity sub-score than with UPDRS-III.16 Similarly, Kerstens et al. using 18F-FE-PE2I PET, found that DAT availability in the motor striatum is more strongly correlated with bradykinesia and rigidity than with UPDRS-III.13 Increasing evidence suggests that PD is not a single entity, rather, it encompasses various clinical subtypes.60,61 Based on clinical symptoms, some PD patients are tremor-dominant, while others exhibit more pronounced bradykinesia-rigidity.62 The total UPDRS-III score provides a comprehensive assessment of motor symptoms in patients with PD, including tremor, bradykinesia-rigidity, and postural abnormalities.63 However, this broad assessment may dilute the effects of specific symptoms, resulting in weaker correlations compared to focusing solely on bradykinesia-rigidity sub-scores. Taken together, our findings suggest that future studies might benefit from focusing on clinical motor sub-scores, rather than composite scores, to achieve a more specific and accurate evaluation of associations between different symptoms and neuroimaging measures.
Previous studies have demonstrated a lack of correlation between 11C-DASB binding and both the total score and sub-scores of the UPDRS-III, suggesting that the serotonergic system may not be involved in the development of motor symptoms in PD.21,64 In contrast, we found that the FC-PET \(\beta\)-values for 11C-DASB (i.e., the association between FC and 11C-DASB PET uptake) were significantly correlated with motor severity in PD patients. This suggests that the development of motor symptoms in PD is not solely due to the degeneration of the dopaminergic system but may also arise from the functional interaction of other neurotransmitter systems and brain networks. Altogether, our findings highlight the complexity of the underlying pathology of PD and underline the value of using multimodal imaging techniques to study the disease.
We also found that the association between FC and PET uptake correlates with non-motor scales in PD patients. However, in contrast to motor scales, this correlation was observed in more brain regions for 11C-DASB PET compared to 11C-PE2I PET. Our results align with the prevailing view that the serotonergic system has a greater influence than the dopaminergic system in the emergence of non-motor symptoms in patients with PD. 65,66 Non-motor symptoms, such as depression, apathy, cognitive impairment, and sleep disorders are often overshadowed by the more overt motor symptoms but significantly affect PD patients’ quality of life.67 Our findings underscore that both dopaminergic and serotonergic systems are involved in the development of non-motor symptoms in PD, but they do so via distinct networks and mechanisms.
Longitudinally, significant correlations were found between changes in motor severity and the association between baseline FC and changes in PET uptake. These correlations were observed in more brain regions for 11C-PE2I than for 11C-DASB, suggesting that dopaminergic dysfunction is more responsible for motor function decline as the disease progresses than serotonergic dysfunction in PD, echoing our cross-sectional findings. For non-motor scales, longitudinal changes in the BDI were positively correlated with the association between baseline FC and changes in 11C-PE2I PET uptake in brain regions within the subcortical and brainstem network. Conversely, longitudinal changes in the BDI were negatively correlated with the association between baseline FC and changes in 11C-DASB PET uptake in brain regions within the sensorimotor, attention, limbic, frontoparietal, and default mode network. Our results align with a previous study that found that dopamine and serotonin have contrasting effects on brain networks. 68
The correlations between FC, PET uptake, and clinical measures observed in this study could have important therapeutic and clinical implications for PD. Traditional treatments in patients with PD often target dopamine levels exclusively, and the classic motor symptoms of bradykinesia and rigidity respond well to dopaminergic therapies.69 However, dopaminergic drugs are often ineffective against tremor and non-motor symptoms, and they may also lead to off-target effects and side effects, such as dyskinesias.70 In this study, we observed that the serotonergic system, traditionally thought to be associated with non-motor symptoms, also correlates with motor severity in PD when interacting with FC. This finding highlights the potential influence of factors such as the mood on PD patients at the time of assessment, which can affect motor severity ratings measured by the UPDRS-III.71 Consequently, treatments targeting both dopamine and serotonin may be more effective and comprehensive in managing PD symptoms than those focusing solely on dopamine. Furthermore, the region- and network-specific relationships we observed between the dopaminergic and serotonergic systems, FC, and specific symptoms could help clinicians develop more customized therapeutic strategies for patients with PD. For example, brain regions within the subcortical network are thought to be predominantly associated with bradykinesia and rigidity, and targeted interventions, such as dopaminergic-replacement therapies and modulation of FC within this network, could be used to address these particular symptoms. Similarly, if a specific brain region or network is identified as being primarily associated with non-motor symptoms such as visual hallucinations, mood and sleep problems, treatments should target the serotonergic system.
Over the past few decades, cell-based therapies focusing on dopaminergic neuron restoration have been considered as a potential treatment for PD.72 The objective of these therapies is to restore dopaminergic neuron functionality by transplanting dopamine-producing cells into the putamen, thereby increasing dopamine levels.30 By studying how dopamine levels covary with FC across brain regions, we can gain insights into which patients might benefit most from this transplantation. For instance, transplanting dopaminergic cells into the putamen of PD patients who already have a high FC between the putamen and the cortex prior to transplantation may increase dopamine level within the cortex. This could result in more positive treatment outcomes compared to patients who had a lower FC between the putamen and the cortex pre-transplantation. It is important to note that these are preliminary conclusions drawn from our findings in pre-transplantation PD patients. In the future, we aim to further validate these results with PD patients who have undergone dopamine cell transplantation to reach more definitive conclusions.
This study has some limitations. First, due to the unavailability of 11C-PE2I and 11C-DASB PET data for HCs, we could not determine whether the distribution of these neurotransmitters was associated with FC in HCs, or if such an association was exclusive to patients with PD. Second, the current study focuses strictly on FC, which is largely matched by SC as assessed by diffusion tensor imaging (DTI), but not wholly 55. Future studies should combine both rs-fMRI and DTI in order to assess the joint contribution of FC and SC to the distribution of neurotransmitters, thereby enhancing our understanding of the association between the distribution of neurotransmitter and brain connectivity.
In summary, the current study demonstrates that highly functionally connected brain regions in PD patients exhibit similar dopamine and serotonin levels, as well as similar changes in dopamine and serotonin levels over time. Our findings suggest a significant correspondence between the patterns of FC and the spatial distribution of dopamine and serotonin that is associated with several features of the disease. An understanding of the complex interaction between dopaminergic and serotonergic systems and functional networks may lead to more comprehensive therapeutic approaches to treat the motor and non-motor symptoms of PD.