NeuroData AI Summer School 2027

About this year

The 2027 school is the first installment supported by the NINDS R25 award and the first under the NeuroData AI name. It continues the format that NeuroDataReHack refined over five years: participants arrive with a reanalysis question, form teams around shared datasets, take instruction in the mornings, and spend most of the week on project work with faculty and dataset stewards on hand. Compared with earlier years, the week is longer and a larger share of the curriculum is devoted to AI methods for neural data.

The DANDI Archive now holds more than 550 neurophysiology datasets in the NWB format, spanning many species, brain areas, task designs, and recording modalities. Some of these are extraordinary resources, including releases from Allen Institute OpenScope, the MICrONS project, and the International Brain Laboratory Brain Wide Map. This school teaches participants how to find, access, and reanalyze these datasets so that data deposited in the archive is used to answer questions beyond those the original experiments were designed for.

Example projects include but are not limited to:

Determine whether your result is present in another species or brain area.
Showcase the capabilities of your tool or analysis on existing data from another lab.
Explore follow-up questions to a published study.
Apply a foundation model or decoding approach across datasets.
Create educational content that uses open data.

The school focuses on analyzing existing data in NWB and on DANDI, not on converting data to NWB. If you want to learn how to convert data, consider an NWB data conversion workshop instead.

Eligibility

Eligibility is broad. The school is intended for PhD students, postdoctoral researchers, principal investigators, and researchers in similar positions. Advanced undergraduates may be considered case by case. We welcome applicants from all areas of neuroscience, including experimentalists, computational neuroscientists, and data scientists, and we reserve roughly a fifth of the places for researchers from computational backgrounds who want to apply their skills to neurophysiology.

Applicants should have basic programming experience in Python or MATLAB and experience with neurophysiology research. Preference is given to applicants who have a specific project in mind that involves reanalyzing existing neurophysiology data in NWB format and who have not previously attended a NeuroDataReHack event.

Logistics

Thanks to the sponsorship of HHMI Janelia Research Campus and the NINDS award, the school is free to participants.

There is no registration or application fee.
Participants are provided a private room at Janelia for the duration of the event, checking in the evening of Sunday, July 18 and checking out on Saturday, July 24 or the morning of Sunday, July 25.
Breakfast, lunch, dinner, and coffee breaks are provided. Janelia is adept at accommodating dietary restrictions; please indicate any on your application.
We reimburse up to $550 of travel costs per participant, and participants who need more can apply for additional support. Details come with the acceptance letter.
You are responsible for booking your own flight and transportation to Janelia.
You are responsible for determining visa and healthcare requirements for travel to the United States. Plan early, since this process can take several months. We are happy to provide a letter of invitation to support a visa application.

What to bring

Bring a laptop with Python installed. MATLAB is optional. Installation instructions for PyNWB and MatNWB are on the resources page. Accepted participants receive preparatory materials six to eight weeks before the school.