The Complex Role of Dopamine in Parkinson’s Disease Resting Tremor

Resting tremor, a primary symptom of Parkinson’s disease (PD), typically appears at rest with a frequency of 4–6 Hz but varies widely among patients. The way brain circuits are affected in PD leading to the development of resting tremor remains difficult to understand. While dopamine (DA) loss is essential for tremor, many patients show no improvement with dopamine replacement therapy, hinting at a more complex relationship. Imaging studies reveal inconsistent findings in dopamine transporter (DaT) levels across tremor and non-tremor forms of PD. This inconsistency may reflect studies with small sample sizes and varied patient groups, highlighting the need for symptom-focused research on tremor in PD. In this study we aimed to investigate clinical and DaT imaging datasets, enriched with data from movement sensors, to clarify the role of dopamine loss in two regions which are typically associated with Parkinson’s disease – the caudate and the putamen.

Caudate Dopamine as a Relevant Feature

From the data available at the Parkinson’s Progression Markers Initiative (PPMI) – an open-access PD database – we found that patients with a resting tremor, when compared to those without, had significantly higher caudate binding ratio (CBR) during a 2-year period. No differences were seen in putamen binding ratio (PBR). If we distinguish patients by their tremor trajectories, we also see higher CBR (but not PBR) in patients who end up developing tremor over time. Furthermore, we saw that patients presenting without resting tremor at the first evaluation showed an increased integrity of caudate dopaminergic terminals, which was associated with an increased risk of developing tremor at follow-up, even after adjusting for other PD symptoms such as bradykinesia and rigidity. These results suggest that caudate dopamine terminal sparing might be relevant for the development of resting tremor.

Enhancing Tremor Assessment with Movement Sensors

Although resting tremor is typically assessed by movement specialists with a reliable instrument, it may not capture the complete, unbiased attributes of this symptom. We therefore took advantage of two datasets enriched with data from movement sensors - one consisting of patients referred for a DaT-SPECT and another composed of PD patients and age-matched controls. We analysed acceleration data provided by these sensors and found an increased power in the 4-6 Hz band for PD patients and for patients with an abnormal DaT-SPECT (evidence of dopamine terminal loss) but not for controls or patients with normal DaT-SPECT. We extracted the maximum power in this band to obtain an unbiased and sensitive accelerometer-based tremor quantification. This metric was positively associated with clinically assessed resting tremor in the PD and DaT-SPECT groups, meaning that it can be used to capture an approximation of parkinsonian tremor. Interestingly, this metric was correlated with CBR in patients with evidence of dopamine loss, but not with PBR. These results present further evidence that caudate dopamine may be relevant for the genesis of resting tremor in PD, and addressing this relationship can be enhanced through the use of sensor data.

Computational Models to Disentangle Lateralized Associations

In PD, symptoms frequently start on one side and the disease remains relatively asymmetric during the first years. If direct causality exists between caudate terminals integrity and tremor severity, a lateralized finding is expected to emerge. We were struck by finding that an association between CBR and resting tremor was mostly driven by positive ipsilateral correlations, an unexpected result that is not aligned with current knowledge on basal ganglia motor control circuits. We hypothesized that this finding could be spurious and would emerge from a combination of factors: 1) patients with tremor have higher CBR, 2) ipsi- and contralateral CBR are highly correlated, 3) PD patients with unilateral tremor have lower contralateral CBR, but because the disease can be asymmetric and unilateral, patients without tremor on the reference side have, on average, lower ipsilateral CBR. 
We created a computational model with these assumptions, without presuming a correlation between tremor amplitude and CBR, which replicated the correlations found in our data and in other publications. This supported our hypothesis by demonstrating that an ipsilateral correlation between CBR and resting tremor can emerge spuriously from other well-established characteristics of these patient populations.

Moving Forward – To Understand the Mechanisms of Tremor with Advanced Tools

In summary, we found that in Parkinson’s disease a higher integrity of dopamine terminals in the caudate, but not the putamen, is related to the presence of resting tremor. Although we do not show that there is a causal mechanism between the sparing of dopamine terminals in the caudate and the development of tremor, we believe this robust observation sheds some light into the motor control circuits involved in the pathophysiology of resting tremor. We also address previous inconsistencies that emerge from available datasets, using a computational approach to show that spurious results may naturally emerge in this context. Finally, we believe our results support the use of sensors to potentiate biologically relevant findings, as they may offer greater sensitivity in detecting clinically meaningful symptom characteristics compared to traditional assessment instruments.