The six spines, and why you need one
Biomedical engineering is wide and shallow by design. The students who win pick one deep, hirable skill and use BME as the wings.
You are skimming: the title, the first figure, and the short version. Switch to Read in the header for the full page, or Deep to open every deep dive.
Here is the structural fact about biomedical engineering that nobody tells freshmen. Medical device companies mostly hire mechanical, electrical, and software engineers, and BME graduates compete against them for the same jobs. A BME degree, on its own, is a little of everything. The students who do well pick a spine, one deep engineering skill that an employer will pay for on its own, and use the biology, the clinical context, and the regulatory fluency as wings that the pure engineers do not have.
In neuroengineering the spines are sharper than in the rest of BME, because the field is so specialized. Pick one by the end of sophomore year. Not before, because you have not tried them yet. Not after, because junior summer is when it gets tested.
The six
Analog and mixed-signal hardware. Microvolt-level amplifiers, low-noise design, electrode interfaces, stimulation circuits, the firmware that runs them. The scarcest skill in the field. Every BCI company is short of people who can design a front end, and the ones who can are paid accordingly. Phase 2 is a long audition for this spine.
Decoding, signal processing, and machine learning. The algorithms that turn neural data into intent. The most crowded spine and still in huge demand. It is far stronger when you also have real hardware experience, because you will know where the noise comes from instead of treating the data as given. Phase 3 is the audition.
Computational neuroscience. Models of neurons and populations, the theory that says what a decoder should be. Academic-leaning; increasingly hired by BCI companies for decoder design. Phase 4.
Electrophysiology and the wet lab. Patch clamp, in-vivo recording, optogenetics, cell culture on electrodes. Needs a lab from the first semester and usually leads to a PhD. The research ladder is the whole path.
Materials and microfabrication. Electrode materials, flexible arrays, coatings, hermetic packagingHermetic packagingSealing implanted electronics against body fluid for decades, one of the hardest unsolved engineering problems in the field. Glossary entry. Underpopulated, deeply valuable, and it needs a cleanroom, which this campus has. Phase 6.
Clinical, regulatory, and human factors. Needs findingBiodesignStanford's process for medical technology innovation that starts from an observed clinical need rather than a technology looking for a use. Glossary entry, design controlsDesign controlsFDA's required process for developing a medical device: user needs, requirements, design, verification, validation, and the records of all of it. Glossary entry, risk analysisFMEAFailure mode and effects analysis: a table of every way the device could fail, how bad and how likely each is, and what you did about it. Glossary entry, FDA pathwaysFDA device classesClass I is low risk (a tongue depressor), Class II moderate (most EEG amplifiers), Class III highest (implanted BCIs), each with heavier review. Glossary entry, usability. Rare, valuable, and almost never chosen by undergraduates, which is exactly why it works. Phase 5. This one pairs with any of the others as a second spine.
How to actually choose
Not by which sounds most impressive. By which one you find yourself doing when nobody assigned it. The student who keeps redrawing the amplifier schematic at midnight has found the hardware spine. The one who cannot stop trying to beat the published accuracy has found decoding. The one who reads FDA recall reports for fun has found regulatory. This site is arranged so that you try each spine’s audition project before you decide.
Deep dive What each spine's first job looks like 4 min
Hardware: “electrical engineer, neural interfaces” at a device company; you own a board from schematic to test. Interviews ask about noise, CMRR, and what you have built and characterized.
Decoding: “algorithms engineer” or “data scientist, neural signals”; you own a model from data to deployment. Interviews ask about cross-validation, leakage, and a paper you reproduced.
Computational neuroscience: usually a PhD first; then research scientist at a company or a lab. Interviews ask you to derive things.
Electrophysiology: research technician, then PhD. Interviews ask about your hands and your notebook.
Materials: process engineer at a device company or a fab; or PhD. Interviews ask about cleanroom experience and characterization.
Clinical/regulatory: “quality engineer,” “regulatory affairs associate,” “clinical engineer.” Interviews ask about a design history file you have written and a recall you have studied. This is the spine where an undergraduate with a mock DHF sounds like a third-year employee.
Deep dive Two spines are better than one, but only in a specific way 2 min
The most valuable profiles in the field are hardware plus decoding (you know where the noise comes from and what to do about it), and anything plus clinical/regulatory (you can build it and you can get it to a patient). The least valuable is a little of all six, which is what a BME degree gives you by default. The wings are only worth something when they are attached to a spine.
Tell your roommate which two spines you are most drawn to right now and what evidence from your own behaviour supports each.