A retrospective study on deep brain stimulation effects on impulsivity in Parkinsonian patients

Parkinson's disease (PD) is a neurodegenerative disorder characterized by motor symptoms such as tremors, rigidity and bradykinesia and non-motor symptoms such as mood disorders, cognitive issues, and impulse control disorders (ICDs). ICDs are often associated with dopamine agonists (DAs), although this association is still debated. DAs are prescribed as an alternative or in combination with levodopa, to reduce levodopa-induced complications such as dyskinesia. Deep brain stimulation (DBS) proves to be a viable alternative: although it is invasive, being localized, it provides a more targeted effect. Moreover, some studies have shown that DBS has an improvement effect not only on motor symptoms but also on impulsive behavior. However, other studies have found no positive impact or even the appearance of ICD after surgery (de novo ICD). This study tries to shed light on this issue by analyzing the effect of DBS in PD patients with ICD. To realize this purpose, the clinical and neural features of 24 PD patients from Careggi Hospital (Florence, Italy) have been analyzed. 12 of them were diagnosed with ICD, and all were treated with STN-DBS. Statistical analyses were conducted to compare patients, those who recovered from ICD after surgery (improved group) and other ICD+ patients (stable group), before, during and after surgery (1-month and 1-year follow-ups). The results obtained were exploited to cluster patients’ profiles before surgery and to classify patients with neural features (during surgery). Analysis before surgery pointed out that the improved group displayed significantly lower values in UPDRS III off (medication off), UPDRS III off-on (difference between off and on) and in bradykinesia subscore, and that clustering with high effect size features (Hoehn Yahr off and rigidity subscore together with the previous ones) allowed to differentiate the two groups with an accuracy of 83%. The significant difference in UPDRS III off remained the same at the 1-year follow-up. Another interesting finding, in contrast to many studies, was the lack of a relationship between dopamine replacement therapy (DRT) and recovery from ICD. After 1 month, DRT was decreased due to the positive effect of DBS, but this decrease was not significantly different between stable and improved groups, nor a DRT difference was found at the baseline condition. When data were available, comparisons between ICD+ and ICD- were conducted, and a significant difference in UPDRS III off and in the Barratt impulsiveness Scale (BIS) was found before surgery. Among neural features retrieved through microelectrode recordings (MERs) during surgery, beta oscillation amplitude, theta oscillation frequency and interburst interval (IBI) differed significantly between improved and stable groups. Beta oscillation amplitude was lower in the improved group, an evidence that was confirmed by literature since a greater beta activity is considered as a pathological oscillation. IBI and theta oscillation frequency showed significantly higher values in the improved group. Beta oscillation amplitude and IBI were selected to train a classifier to recognize individual improved or stable neurons and subsequently, by majority voting, to define the patient's class. The prediction of patients’ conditions reached an accuracy value of 83%, as did clustering performance. This may have clinical relevance, as it allows to conduct an analysis of the patient’s profile and future outcomes before surgery, better assessing the procedure to be followed.