Advances in deep brain stimulation for parkinson's disease

Resumen

Parkinson's disease is the second most common neurodegenerative disorder. It is expected to grow to pandemic proportions by 2040, surpassing Alzheimer's disease. Nonetheless, it remains an idiopathic disease in about 95\% of cases, although it is known that it is caused by the degeneration of dopaminergic neurons of the Substantia Nigra Compacta (SNC). The loss of neurons in this brain area produces a disequilibrium, responsible for the symptoms of the disease, which include tremor of the limbs at rest (the so-called resting tremor (RT)), which is the most characteristic symptom of the disease, muscle rigidity, inability to initiate precise movements (akinesia) and slow motion, especially in complex voluntary movements (bradykinesia). To alleviate these symptoms, the first option is usually a pharmacological treatment, with or without dopaminergic effects. However, some patients have a tremor with high resistance to medication, even at the highest tolerable doses of levodopa. Besides, the use of levodopa leads to dyskinesias (LID), in which the patient suffers from involuntary movements that may ultimately be worse than the original PD's symptomatology. The second line of treatment in such cases is Deep Brain Stimulation (DBS). DBS consists of the surgical implantation of a neurostimulator, an implantable medical device (IMD) that uses an implanted pulse generator (IPG) to deliver electrical current through a set of electrodes to the surgical target; usually, the SubThalamic Nucleus (STN), modulating its functioning. Despite the growing trend in DBS use, the exact mode of operation and the effects it produces on brain networks remain confused. A higher level of understanding of the neurophysiological changes induced by DBS would be an essential step for several reasons. Firstly, to gain insights into the therapeutic mechanisms of DBS. Secondly, based on which, the accuracy of DBS could be improved, perhaps avoiding or reducing adverse effects and monitoring treatment response. It will lead to a better understanding of the functioning of the brain, under different conditions. For this reason, one of the two objectives of this thesis focuses on studying the effects that stimulation generates at the cortical level. However, the majority work of this thesis focuses on the functioning of the neurostimulator itself. In the context of cardiac illnesses, pacemakers have the ability to adapt the stimulation to perform event-response in real time. However, presently, neurostimulators, once implanted, provide continuous stimulation, which may induce adverse effects such as paresthesia, psychiatric or cognitive malfunction and even an increased risk of suicide. Real-time adaptive (closed-loop) DBS systems represent a better strategy, in which the IMD stimulates only when necessary, on demand, thereby reducing the adverse effects. These systems will sense continuously a feedback signal, through which to infer a biomarker, that correlates with the patient's symptomatology. The implementation of such a strategy in PD requires, therefore, the knowledge of what features of the STN activity change when (or ideally, shortly before) the clinical symptoms appear. This thesis includes four published works that aim to achieve a closed-loop DBS system with the highest possible level of accuracy. To this end, methods of different areas are explored, from chaos theory to machine learning, achieving 100% accuracy in the last work presented.