The problem Deep brain stimulation (DBS) surgery is used to treat symptoms of neuropsychiatric and movement disorders for which other treatment options have been exhausted. The treatment is extremely invasive and not without risk. During DBS surgery, a microelectrode is lowered deep into the brain with the aim to stimulate small subcortical nuclei to alleviate symptoms such as rigidity and tremor in PD. Surgical planning involves combining magnetic resonance images (MRI) of the patient with brain anatomy atlases to optimize electrode placement during surgery. Studies using DBS in PD patients show that a suboptimal placement of electrodes in, for example, the Subthalamic Nucleus (STN) or Globus Pallidus (GP), can yield changes in cognitive processes (e.g. attention, mental speed, response inhibition) and affective states (e.g. depression, hypomania, anxiety, hypersexuality, and hallucinations). These unwanted side effects of DBS may be the result of the stimulation of non-motor zones within these nuclei, the stimulation of white matter connections, or the dysregulation of blood flow to neighboring areas (Horn et al., 2017). Crucially, electrode implantation changes the physiology (Noor et al., 2016) and can disrupt local blood vessels (Kozai et al., 2014), leading to poor outcomes. These effects are likely due to the limited information used to dictate the location of implantation. To date, leading commercial systems (e.g. Boston Scientific, Medtronics) rely on stereotactic coordinates and classical anatomy atlases of the target structures. Research software such as Lead-DBS are actively trying to remedy this by integrating information representing different important facets of the relevant structure and physiology.
The main challenge here is that the human subcortex is a highly crowded brain area, which consists of hundreds of unique, small grey matter nuclei constituting approximately ¼ of total human brain volume. These nuclei are inter-connected in complex networks and pathways involved in decision making, reward processing, attention modulation, as well as motor and sensory activity. They are richly vascularized with major arterial trees and veins passing through them to support their energy demands, which often change rapidly in disease, including in PD (Paul and Elabi, 2022). Importantly, the impact of mapping them in human brain MRI atlases has only recently been recognized, in part due to my efforts to raise awareness (Forstmann et al., 2017; Terra Incognita workshop; NeuroImage special issue).
