Files Format

Supported source files

Use File → Open and select either a source file or a folder, depending on the format. Ephyr detects the format, converts it into a Ephyr experiment folder when needed, then opens a session.

Format	Typical extensions / markers	How to open
Axon ABF	.abf	Select the file
EDF	.edf	Select the file
DAQ	.daq	Select the file
XDAT	.xdat, *.xdat.json	Select the file
NWB	.nwb	Select the file
Neuralynx	.ncs (also .nev, related text)	Prefer the folder that contains .ncs files. Selecting a Neuralynx file is also accepted; Ephyr resolves to the parent folder of the .ncs set.
Open Ephys	session folder (settings.xml, continuous streams, etc.)	Select the session folder, or a file inside it (Ephyr walks parent directories until a valid session is found).
Intan RHD	.rhd, optional .xml	Select the folder that contains the .rhd files, or an .rhd/.xml file (resolved to the parent folder).
Intan RHS	.rhs, optional .xml	Same as RHD: prefer the folder with .rhs files.
Legacy Ephyr	paired *.lfp + *.header.json	Select the folder that contains the matching pair.
Existing Ephyr experiment	header.json inside *_ephyr	Select the Ephyr experiment folder directly (no conversion).

Conversion notes

• Opening a non-Ephyr source creates a sibling folder named {stem}_ephyr next to the chosen path (for example, exp.abf → exp_ephyr, folder my_rec → my_rec_ephyr).
• For Intan recordings, the conversion dialog may offer Group all Intan files into one sweep.
• If a valid {stem}_ephyr folder already exists next to the source, Ephyr loads it instead of converting again.
• NWB conversion requires a regular TimeSeries / ElectricalSeries with a usable sample rate (typically under acquisition or processing). Files that only contain spike-sorted Units (no continuous voltage series) are not supported for conversion. Large HD-MEA NWB files are quantized without a full-file peak scan; conversion still streams the entire series once into data/.

After conversion, electrode order and (for NWB) spatial layout can be restored from the original source via Channels layout → Import from source and Layout settings → Import from source (see Settings Panel).

Ephyr files

After conversion (or when you open an existing experiment), Ephyr uses this directory layout:

$EXPERIMENT_ephyr/
├── header.json
├── data/
│   └── sweep_NNNNN/
│       ├── 0.samples
│       ├── 1.samples
│       └── ...
├── sessions/
│   └── $SESSION_NAME.json
└── add_ons/
    └── data/
        ├── $ADD_ON_MODULE_NAME/
        └── ephyr/

header.json

Experiment metadata: sample rate / interval, number of channels and sweeps, points per sweep, channel names, units, analog/digital ranges, provenance of the original source, and related fields. This file identifies a valid Ephyr experiment folder.

data/

Signal samples stored as read-only int16 memory-mapped files.

• One subdirectory per sweep: sweep_00000, sweep_00001, …
• Inside each sweep, one *.samples file per channel (0.samples, 1.samples, …), indexed by channel index.

Scripts and the GUI read these arrays through ExperimentData.data_memmaps and convert them to voltage with from_int16_to_voltage_val (see Analysis).

sessions/

User annotation and GUI state. Each session is a JSON file (for example my_session.json) that stores:

• event and period vocabularies and placements
• experiment description text
• GUI setup (visible window, channel groups, filters, add-on toggles, layer visibility such as traces / channel names / events / periods, and related view state)

Created when you create or save a session. Switch between sessions without reloading the signal data.

add_ons/ and add_ons/data/

Persistent storage for add-ons:

• add_ons/data/$ADD_ON_MODULE_NAME/ — results and parameters for a specific add-on module (for example spike detection payloads).
• add_ons/data/ephyr/ — shared Ephyr-side add-on state (for example preprocessing pipelines used by tools).

Conversion creates the add_ons/ tree; individual module folders appear when add-ons run and write data.