Learning in public —
from the beginning.

When I first heard about deep brain stimulation, I couldn't understand how it was real. Surgeons drill through the skull, thread electrodes deep into a structure called the subthalamic nucleus, and run wires under the skin to a pacemaker-like device in the chest. They turn it on. For many Parkinson's patients, the tremors stop — sometimes within seconds. It seemed like science fiction. It's been in clinical use since the 1990s.

The subthalamic nucleus is a small, lens-shaped structure that plays a key role in motor control. In Parkinson's, the loss of dopaminergic neurons in the substantia nigra throws the basal ganglia circuit out of balance. The subthalamic nucleus becomes overactive — firing in abnormal, synchronized bursts that disrupt smooth movement. DBS works by interrupting that abnormal signaling. The exact mechanism is still debated, but the effect is clear: in the right patients, it dramatically reduces tremor, rigidity, and slowness of movement.

Here's what DBS doesn't do: it doesn't stop neurons from dying. The underlying neurodegeneration continues. Alpha-synuclein still misfolds and accumulates. Lewy bodies still spread. Many patients do well for years, then find that DBS becomes less effective as the disease progresses into brain regions not targeted by the electrodes. The surgery also carries risks — infection, lead migration, hardware failure, and in rare cases, intracranial hemorrhage.

What I found most interesting was the research happening now around closed-loop DBS — systems that read the brain's electrical activity in real time and adjust stimulation automatically, rather than running at a fixed setting. This is where neurotechnology and neuroscience intersect most directly. The brain tells the device what it needs. The device responds. It's a glimpse of what brain-computer interfaces could eventually become for patients who can't speak or move.

In January 2023, the FDA gave full approval to lecanemab (brand name Leqembi) — an antibody drug that targets amyloid-beta, one of the two hallmark proteins that accumulate in Alzheimer's disease. It was the first drug to demonstrate that clearing amyloid from the brain actually slows cognitive decline. The clinical trial showed a 27% reduction in decline on a standard cognitive assessment over 18 months. The headlines called it a breakthrough. I wanted to understand what that actually means.

Lecanemab is a monoclonal antibody — an engineered protein designed to bind to and tag amyloid-beta protofibrils for removal by the immune system. The idea behind it comes from the amyloid hypothesis: that the accumulation of amyloid plaques between neurons triggers a cascade that eventually kills them. For decades, drugs targeting amyloid kept failing in trials. Lecanemab is the first to show clear clinical benefit — not just biomarker improvement, but measurable slowing of real-world cognitive decline.

But I wanted to understand the limits too. Lecanemab works only in early-stage Alzheimer's — before too many neurons are already dead. It comes with serious risks: about 35% of patients in the trial experienced ARIA (amyloid-related imaging abnormalities), which includes brain swelling and micro-bleeds. It requires infusions every two weeks, costs roughly $26,000 per year, and isn't accessible to most patients globally. And it doesn't reverse any damage already done. It slows a process. It doesn't stop or undo it.

What struck me most was the implications for early detection. If lecanemab only works before significant neuronal loss, then identifying Alzheimer's years before symptoms begin becomes critical. This is driving intense research into blood biomarkers — particularly plasma p-tau217, which can now predict Alzheimer's pathology with high accuracy from a simple blood test. If you could detect amyloid accumulation a decade before memory loss begins, you could intervene when intervention actually matters.

I've been interested in Meta glasses since I first read about them — not because I wanted to wear them at lunch, but because of what they actually are: a device that overlays digital information onto the physical world in real time. That's not a consumer feature. That's a scientific instrument. And the more I read about how the brain processes what we see, the more I became convinced these devices have an important role to play in neuroscience research.

The brain doesn't passively receive visual input. It actively predicts and constructs what it expects to see, then updates that model when sensory data conflicts with the prediction. This is called predictive processing, and it's one of the most influential frameworks in modern neuroscience. When you put on AR glasses and overlay a digital object onto a real table, you're creating a conflict between prediction and input. How the brain resolves that conflict — and how resolution changes with age, disease, or injury — is scientifically interesting.

In Parkinson's disease, visual processing and spatial perception are often disrupted early — sometimes before motor symptoms appear. Patients report difficulty judging distances, navigating crowded spaces, and tracking moving objects. Could AR-based assessments detect these subtle perceptual changes earlier than current clinical tests? There's early research suggesting yes. Some labs are using VR and AR environments to track eye movement patterns, reaction times, and spatial navigation as biomarkers for neurological disease.

In Alzheimer's, spatial disorientation is one of the earliest and most distressing symptoms — patients get lost in familiar places because the hippocampus, which encodes spatial maps, is one of the first structures affected. Virtual and augmented environments offer a way to test spatial memory precisely and repeatedly, creating a longitudinal picture of how navigation ability changes over time.

CRISPR-Cas9 can find a specific sequence of DNA — a single sentence in a three-billion-letter book — and cut it, replace it, or silence it with near-surgical precision. Since Jennifer Doudna and Emmanuelle Charpentier described the system in 2012, it has transformed biology. The question I wanted to answer: can it do anything for Parkinson's disease?

About 10–15% of Parkinson's cases are genetic — caused by mutations in specific genes. The most common are LRRK2 (leucine-rich repeat kinase 2), which in its mutant form appears to hyperactivate a cellular cleaning process in ways that damage neurons, and PINK1/Parkin, which disrupts mitochondrial quality control. In principle, CRISPR could correct these mutations — either in cells before transplantation or, more ambitiously, delivered directly to neurons in the brain.

The obstacles are real. Getting CRISPR into the brain requires a delivery vehicle — usually an adeno-associated virus (AAV) — that can cross the blood-brain barrier and infect neurons without triggering immune responses. Even if delivery works, editing efficiency in neurons is low: most cells don't get edited. Off-target effects — unintended cuts in other parts of the genome — remain a concern, though newer base-editing and prime-editing systems have dramatically improved precision. And for the 85–90% of Parkinson's cases that aren't clearly genetic, CRISPR's role is less obvious.

There's also a timing problem. By the time Parkinson's is diagnosed, roughly 60–80% of dopaminergic neurons in the substantia nigra are already dead. Fixing a gene mutation at that point doesn't restore what's been lost. Like lecanemab for Alzheimer's, the real power of genetic intervention may lie in prevention — correcting mutations in people who carry them before neurodegeneration begins.

I grew up with four grandparents who loved me in four completely different ways. My maternal grandfather was the patient one — the one who let me sit beside him and ask questions and never made me feel like a burden for being curious. My paternal grandfather was warmer in a different way, more expressive, the one who made every visit feel like a celebration. Both of them shaped who I am. Both of them are losing something essential now.

My maternal grandfather has Parkinson's disease. I notice it most in his hands — the tremor he tries to control, the way he holds things differently now. He doesn't complain. He adapts. But I see the effort it takes, and it costs something to watch. The man with the steadiest hands I knew now has to think about things that used to be automatic. Parkinson's does that: it makes the unconscious conscious, and the effortless effortful.

My paternal grandfather has Alzheimer's disease. His loss is different — quieter in some moments, more disorienting in others. There are times he doesn't immediately know who I am. He'll look at me and search — I can see him searching — and then something clicks. Or sometimes it doesn't. I've learned not to show that it affects me, because it doesn't help him if it does. I just hold his hand and talk about things he remembers. We go back to places that still exist for him.

I'm a boarding student at Mercersburg Academy in Pennsylvania, thousands of miles from both of them. Distance has a way of sharpening things. When you can't be there, you find other ways to be useful. For me, that meant reading. Learning. Building toward something that might, in some small way, matter.

I don't know yet what my contribution will look like. Maybe I'll end up in a research lab. Maybe in clinical medicine. Maybe in biotech, building the tools that researchers and doctors use. What I know is the problem I want to work on — and that knowing that, at 16, is something I don't take for granted. Most people spend years figuring out what they care about. I had two grandfathers who showed me.