Introduction to ONE (Open Neurophysiology Environment)

Neurophysiology needs data standardization. A scientist should be able to analyze data collected from multiple labs using a single analysis program, without spending untold hours figuring out new file formats. Substantial efforts have recently been put into developing neurodata file standards, with the most successful being the Neurodata Without Borders (NWB) format. The NWB format has a comprehensive and careful design that allows one to store all data and metadata pertaining to a neurophysiology experiment. Nevertheless, its comprehensive design also means a steep learning curve for potential users, which has limited its adoption.

How the open neurophysiology environment works

Here we provide a solution to this problem: a set of four simple functions, that allow users to access and search data from multiple sources. To adopt the standard, data providers can use any format they like - all they need to do is implement these functions to fetch the data from their server and load it into Python or MATLAB. Users can then analyze this data with the same exact code as data from any other provider. The learning curve will be simple, and the loader functions would be enough for around 90% of common use cases.

By a data provider we mean an organization that hosts a set of neurophysiology data on an internet server (for example, the International Brain Lab). The Open Neurophysiology Environment (ONE) provides a way for scientists to analyze data from multiple data providers using the same analysis code. There is no need for the scientist to explicitly download data files or understand their format - this is all handled seamlessly by the ONE framework. The ONE protocol can also be used to access a scientist’s own experiments stored on their personal computer, but we do not describe this use-case here.

Data structure

Every experiment a data provider releases is identified by an experiment ID (eID) – a small token that uniquely identifies a particular experiment. It is up to the data provider to specify the format of their eIDs.

The data files have to follow the Alyx file name specification (ALF). The ONE package is concerned with parsing and loading files that follow this specification.

Data browsing and loading

When a user wants to analyze data released by a provider, they first import that provider’s loader functions. In python, to analyze IBL data, they would type

from one.api import ONE

Because it is up to data providers to maintain the loader functions, all a user needs to do to work with data from a specific provider is import their loader module. To analyze Allen data, they could instead type import one_allen. After that, all other analysis code will be the same, regardless of which provider’s data they are analyzing.

Once ONE is installed, the user can follow these tutorials to get acquainted with:

• searching for experimental session through the database

• listing available datasets for an experimental session

• loading datasets of a specific session (e.g. given an eID)

Further tutorials can be found in the following sections: introductory and advanced.