VWV Volumes #30: Analyst-in-Training Daniel Lee on Artificial Intelligence & Neurology
Artificial Intelligence & Neurology: Two Converging Fields with Unlimited Possibilities
Introduction
The brain is our black box of complete mystery. A three pound mass of electrical signaling, an amalgamation of chemicals and hormones, and what seems like just simple white and dark matter–yet, it plays a vital role in every moment of our existence, from our perception of sensations to the transient thoughts that cross our minds.
The realization of how little we understand the core part of our being may be quite daunting, especially when thinking of our vulnerability to diseases in the brain. Yet, it also comes with excitement of possible discoveries and treatments–perhaps even greater excitement when imagining the convergence of Artificial Intelligence (AI) and neurology and what that means for the future neuro-related healthcare and research.
A brief look into the Neurology/AI industry
The boom of the artificial intelligence industry had a cascading effect, forming many other segmented industries, case in point, the neurology/AI industry. This industry has been reported by an early market research study to be valued at a current 48 million USD with a predicted compound annual growth rate (CAGR) of around 46% in the years 2024 to 2030.
Such large growth is attributed to an equally large need for AI to assist in procedures from data filtration to rapid diagnosis of brain diseases. Furthermore, an article by Deloitte states that neurological and neuropsychiatric disorders account for 17 million annual deaths (around 30% of annual deaths), evidencing a need for intervention. Thankfully, intervention has already started its course, and AI has already made great strides in the neurological field.
AI and Neurology Linkup
To understand how AI has been an instrumental tool in bettering practices for various neurological disorders, one study explains AI involvement in various facets of neurological disorders. Let’s delve into the following neurological disorders, to glean deeper insights into how AI has enhanced their treatment and management:
Brain Tumors:
Definition:
A brain tumor is defined as an abnormal growth of cells in the brain that can either be cancerous or non-cancerous.
Current AI Involvement
AI, particularly machine learning and deep learning algorithms, have assisted in more efficient tumor detection, segmentation (identifying spatial location of the tumor), and classification.
The methodology is through a Convolutional Neural Network (CNN) which involves processing pixel data for image recognition through deep learning algorithms. CNN is particularly useful for finding patterns in images, allowing for deep learning to take in the patterns and output results.
Degenerative Disorders:
Definition:
Degenerative disorders occur through progressive death of cells and neural connections that impair movement, cognition, and sensation. Two of the most common neurodegenerative disorders are Parkinson’s Disease and Huntington’s Disease, both characterized by motor dysfunction.
Current AI Involvement
AI algorithms have been used quite creatively to better diagnose patients with Parkinson’s and Huntington’s. Simple tasks like drawing and handwriting have been incorporated into various algorithms to detect early signs of neuromuscular decline. Other algorithms, like naïve Bayesian, decision tree, SVM have contributed to more accurate diagnosis of degenerative disorders.
Neurological Infections:
Definition:
Neurological infections are caused by viruses, bacteria, fungi, etc. and are often attributed with a high mortality rate in the central nervous system.
Current AI Involvement
An issue with neurological infections is the varying symptoms that make classification/diagnosis difficult. Meningitis, a common neurological infection, causes inflammation in the brain and spinal cord areas. Current detection for meningitis occurs through invasive procedures like the lumbar puncture. By taking various parameters like temperature and cerebrospinal fluid concentration, current AI predicts meningitis at around 97% accuracy. As different types of meningitis and other neurological infections have high mortality rates, AI can play a key role in early detection–and as an added bonus, is non-invasive.
Limitations to AI & Neurology
While AI continues to develop (and at a rapid pace), its promise for unending applications comes with just as much excitement as it does challenges. For one, AI relies on large amounts of data for algorithmic computing, and data can often be explicitly and/or implicitly biased by the individuals creating the algorithms. Furthermore, AI’s Black Box Problem, explained as our inability to derive a step-by-step explanation for an AI’s solutions, poses its own set of challenges. At a legal level, the Food and Drug Administration (FDA) has greatly limited the use of AI in clinical trials due to the legal ramifications and resources that may come with algorithmic errors.
Conclusion
As of now, the precise role of AI within the neurology industry remains uncertain. There's ongoing debate about whether it will revolutionize the field or merely serve as a novel yet limited tool. However, what is clear is the promising convergence of neurology and AI, offering avenues for preventing, treating, and studying neurological diseases, particularly within the complex and intriguing realm of the brain. 🧠
About the Author
Daniel is a freshman from the Bay Area in California planning to concentrate in Neuroscience and Computer Science. He has a keen interest in corporate culture and is excited to explore entrepreneurship from the people perspective. Daniel is also a huge soccer fanatic and is part of the club soccer team at Brown. For leisure, you can find him scaling the rock climbing walls at Rock Spot, vibing to the Marías, and indulging in one too many lavender lattes at the Underground.🍵
Sources:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10053494/
https://www.sciencedirect.com/science/article/pii/S0893608021003683
https://www.marknteladvisors.com/research-library/artificial-intelligence-neurology-market.html