I’ve been thinking a lot about my (potential) future in academia.

I want my research to have three main pillars - none of which are surprising if you know my trajectory.

Pillars

Three main pillars:

Neuro

Neuro makes sense to most people - my work has historically been anchored in neural systems. While I definitely don’t see myself as a neuroscientist, I do tend to tackle challenges in the neural domain. It might be better to say I’ll be focusing on neuroengineering and neuromedicine.

Medicine

Medicine also makes sense to most folks who know my MD/PhD training. But what may make less sense is that I have no interest in medicine as-is - I care much more about what the future of medicine should look like. Through my training, one simple truth became clear: medicine is engineering. It’s time we do it justice and bring math to anchor its more powerful inferential approaches 1.

Law

The broader context is always critical if you’re designing an effective, efficient intervention-machine. Law is that context for both neuro and medicine - we can’t ignore it or abstract it away. We need (a) to be aware of the bigger context our interventions and plants 2 operate in and, (b) encode that awareness in system diagrams that can plug in neatly to active research efforts 3.

Approaches

The pillars are distinct, but each will have the same columnar components - the base will be, of course, control theory. Much more on that later…

  1. Hint: null hypotheses have no role, while things like intuition and whole-system perturbation do. Basically, all the intuitions that are made reflexive in scientific training are antithetical to medicine-as-engineering. ↩︎

  2. Plant in the control-theoretic sense: the system we are trying to influence. The patient’s physiology is often the main “plant” in medicine, but my whole thing is going to be that the physiology-alone is never enough for effective or optimal control. ↩︎

  3. AI techniques will make this much more feasible. ↩︎