How AI is Changing Neurology: A Developer’s Perspective

Prepared by: Alireza Minagar, MD, MBA, MS
software engineer
Intro:

As a neurologist-turned-software engineer, I’ve explored how AI models trained on radiological, genomic, and clinical data are transforming the landscape of neurology.

Let’s break down:

🧠 Use Case 1: Predicting Multiple Sclerosis Progression
Dataset: Longitudinal MRI scans + CSF profiles

Model: LSTM + random forest hybrid

Output: 5-year EDSS disability trajectory

🧬 Use Case 2: Genomic Pattern Mining in Alzheimer’s
Tools: scikit-learn, BioPython, pandas

Technique: SNP frequency clustering using unsupervised learning

Result: Highlighted APOE4 and potential novel modifiers

📊 Use Case 3: NLP in Neurology Notes
Stack: spaCy + SciKit-learn

Task: Flagging seizure risk factors from text

Result: 78% recall, 65% precision on clinical validation

Conclusion:
AI isn’t replacing neurologists—it’s becoming their algorithmic co-pilot.

Would love to hear from others building ML models in medicine. Let’s collaborate.