Project D: Neuromatch Academy

Neuromatch AcademyNeuromatch Academy An online, three-week summer school in computational neuroscience with a large alumni network, on a country-scaled tuition. Glossary entry started in 2020 as an emergency online replacement for a summer school and became the largest computational neuroscience course in the world. Three weeks, every weekday, in a small pod of students with a teaching assistant, working through tutorials written by leading researchers and building a group project on real data. It used to be free. It is not: the fee is scaled to your country, and for a US student in academia it is $1,033 plus a$50 processing fee, halved to $516 plus$50 if you confirm that no supervisor or grant will reimburse you, with a further hardship discount available at enrolment. It is taught in many time zones, and its alumni are now in labs and companies everywhere. For a student on this site, it is the single most efficient way to get the computational neuroscience spine’s fundamentals in one go.

The tracks

Computational Neuroscience is the original: model fitting, linear algebra for neuroscience, dynamical systems, biological neuron models, network dynamics, Bayesian inference, reinforcement learning, and a project with a dataset from the Allen Institute, Steinmetz, or others. Deep Learning covers modern machine learning with a neuroscience flavour. NeuroAI is the newest. For this curriculum, Computational Neuroscience first; Deep Learning the following year if you want the decoding spine sharpened.

Getting in

Applications open in the spring, with a deadline typically in March or April (check the calendar and the site). The interactive track needs commitment to the daily schedule; an observer track lets you follow the materials without a pod. The fee is the part to plan for: ask a supervisor, a department, or a fellowship first, since most people who go have it paid from a grant, and take the unreimbursed rate and the hardship discount only after that fails. Applying costs nothing. The application asks about your background and why; be specific about what you have already built, which is where this site’s build logs pay off.

Arriving ready

The course assumes comfortable Python with NumPy, basic linear algebra (vectors, matrices, eigenvectors), basic probability, and calculus through derivatives and integrals. It runs pre-course refreshers, which are excellent. If Phase 1 to 3 are behind you, you are more than ready on the coding side; spend the month before on linear algebra and probability from the math thread.

Getting the most out of it

Do the tutorials before the pod session, not during it; the session is for the parts you got stuck on. Pick a project that uses real neural data (spikesSpike Shorthand for an action potential as seen by an electrode, a brief blip about a millisecond wide. Related: Action potentialAction potential The brief, all-or-nothing voltage spike a neuron fires, about a millisecond long and about 100 millivolts tall across its membrane. , Spike sortingSpike sorting Deciding which spikes on an electrode came from which neuron, by clustering their waveform shapes. , Extracellular recordingExtracellular recording Recording with an electrode near neurons rather than inside one; you see small, fast blips from nearby cells rather than the full membrane swing. . Glossary entry or LFPLocal field potential (LFP) The slow, summed electrical activity of many neurons near an electrode, below a few hundred hertz, as opposed to individual spikes. Related: Electroencephalography (EEG)Electroencephalography (EEG) Recording the brain's electrical activity from electrodes on the scalp, a few tens of microvolts of blurred, summed cortical activity. , Extracellular recordingExtracellular recording Recording with an electrode near neurons rather than inside one; you see small, fast blips from nearby cells rather than the full membrane swing. . Glossary entry) rather than a purely theoretical one, because that is the experience the next steps need. Talk to your teaching assistant about their lab; they are graduate students and postdocs and they know who is hiring undergraduates. Keep the pod’s chat group afterward.

If you cannot attend

All materials are public year-round on the Neuromatch site, with videos and notebooks. Working through the Computational Neuroscience tutorials alone, one day a week for a semester, gets most of the content without the pod. Pair it with the Coursera Computational Neuroscience course (Rao and Fairhall) for a second angle on the same material.