top of page
  • Writer's pictureTriple Helix

The Future of Deep Brain Stimulation: AI?



Writer: Anthony Kharrat ’26

Editor: Jamie Saito ’25

DBS

Over the past two decades, Deep Brain Stimulation (DBS) has gained widespread acceptance for treating movement disorders associated with neurological conditions including Parkinson’s disease, essential tremor, dystonia, and epilepsy. In recent years, mounting evidence suggests the potential application of DBS in treating select neuropsychiatric disorders such as Obsessive-compulsive disorder (OCD) and depression. [1] By 2019, over 160,000 patients worldwide have undergone DBS, and with the continuous rise in average life expectancy, the DBS market is projected to grow by 14% annually, approaching a $4 billion market by 2030. [2]

 Now, what exactly is DBS? Deep brain stimulation is a minimally invasive neurosurgical intervention that uses implanted electrodes, or leads, and electrical stimulation to disrupt irregular brain signaling that causes pathophysiological motor symptoms including tremors, freezing of gait, and other uncontrolled movements. Typically, two electrodes, one for each brain hemisphere, are precisely implanted into neural regions such as the basal ganglia or cerebellum, which are responsible for motor control and coordination. These electrodes are connected to a small neurostimulator, usually implanted under the person’s collarbone, that delivers electrical impulses at a programmed intensity and rate. [1]

Since DBS requires surgery, physicians typically recommend the therapy in cases where patients respond poorly to medications, serving as a second-line treatment with clinically proven efficacy. Nevertheless, DBS outcomes can vary due to multiple factors. Some of these are inherent to the patient, such as baseline disease characteristics and brain circuitry differences, while others are externally modifiable, including electrode positioning and electrical programming parameters. [3]



AI

In the midst of the burgeoning AI revolution, researchers are now exploring artificial intelligence-based strategies to optimize these externally modifiable factors and improve DBS outcomes. Currently, AI is being used to both facilitate brain target localization and refine electrical stimulation parameters. For DBS therapy to be effective, initial targeting for electrode placement is critical. If placed incorrectly, not only will a patient's debilitating symptoms remain, but other detrimental brain-related symptoms could arise. Traditional algorithms rely on brain atlas templates to define neuroanatomical structures for lead placement; however, these are often inaccurate for small brain nuclei with high individual variability which are common DBS targets. Now, researchers are leveraging neural networks and deep learning applications, trained on large datasets, to accurately identify and segment these intricate deep brain nuclei. Promising laboratory research suggests that implementing these trained AI models to guide electrode placement could revolutionize DBS, significantly improving its accuracy and efficacy. [3]  

In the realm of DBS programming, several groups have developed AI algorithms to guide and automate the electrical programming of the neurostimulators. While conventional DBS involves the delivery of stimulation at constant electrical parameters, over the past decade, there has been increased interest in closed-loop or adaptive systems that adjust stimulation parameters based on the patient’s neural activity. However, those approaches have mostly involved simple algorithms that instruct the stimulation to be switched on and off or to a different intensity when recorded signals reach a specified threshold. Despite their simplicity, these algorithms have already produced encouraging results. Now, deep learning applications hold the potential to integrate a broader range of neural data and deliver more complex parameters of stimulation, further enhancing DBS adaptability. [3]

Ultimately, artificial intelligence stands poised to revolutionize DBS protocols by optimizing target localization and personalizing stimulation protocols in real time, leading to better outcomes. [4] The potential applications are vast, as advanced deep learning algorithms are capable of addressing the multi-dimensional complexities of individual brains and diseases that were previously unsolvable. Nonetheless, the practical use of these AI models in the field still needs refinement and validation in long-term, prospective studies before it can be fully integrated into human treatment. In regards to health equity, it is crucial to ensure that the anticipated increase in therapy costs does not hinder patient accessibility to DBS. [3]









References 

[1]  Johns Hopkins Medicine medical professionals. 2023. “Deep Brain Stimulation.” Johns Hopkins Medicine (October): https://www.hopkinsmedicine.org/health/treatment-tests-and-therapies/deep-brain-stimulation 


[2]  Market Analysts. 2022. “Deep Brain Stimulation Devices Market.” Global Market Insights (December): https://www.gminsights.com/industry-analysis/deep-brain-stimulation-devices-market 


[3]  Limousin, Patricia, Harith Akram. 2023. “AI and deep brain stimulation: what have we learned?” Nature Reviews Neurology (December): https://doi.org/10.1038/s41582-023-00836-9   


[4]  Allen, Ben. 2023. “Discovering Themes in Deep Brain Stimulation Research Using Explainable Artificial Intelligence.” Biomedicines (December): https://doi.org/10.3390/biomedicines11030771 




7 views0 comments

Comentários


bottom of page