Pallidal spiketrain variability and randomness are the most important signatures to classify Parkinson's disease and cervical dystonia

Movement disorders such as Parkinson’s disease (PD) and cervical dystonia (CD) are associated with abnormal neuronal activity in the globus pallidus internus (GPi). Reduced firing rate and presence of spiking bursts are typical for CD, while PD is characterized by high frequency tonic activity. This research aims to identify the most important pallidal spiking parameters to classify these conditions. We analyzed the single unit activity of the external (GPe) and internal (GPi) segments of the globus pallidus in 11 CD and 10 PD patients who underwent standard DBS implantation. We compared firing rate, firing pattern and oscillatory characteristics of tonic, burst and pause cells and used logistic regression and random forest models to classify patients according to their pallidal activity. In the GPi we discovered prevalence of high firing rate tonic cells in patients with PD, while in dystonia burst neurons with high firing rate were predominant. GPi pause cells were mostly observed in CD patients and exhibited less spike variability compared to PD. Characteristics of neurons and their distribution in the GPe was similar. Logistic regression and random forest models identified spike variability and randomness as the key features for distinguishing between PD and CD, instead of firing rate or oscillation properties. Our study demonstrates that pallidal activity can predict Parkinson’s disease and cervical dystonia with high accuracy. Burst dynamics and characteristics of spiking randomness including entropy appear to be the most meaningful reflections of the neurophysiology of studied diseases.