__ __ ______ ______ ______ ______ __ ______ __ _____
/\ "-./ \ /\ ___\ /\ ___\ /\ __ \ /\ ___\ /\ \ /\ ___\ /\ \ /\ __-.
\ \ \-./\ \ \ \ __\ \ \___ \ \ \ \/\ \ \ \ __\ \ \ \ \ \ __\ \ \ \____ \ \ \/\ \
\ \_\ \ \_\ \ \_____\ \/\_____\ \ \_____\ \ \_\ \ \_\ \ \_____\ \ \_____\ \ \____-
\/_/ \/_/ \/_____/ \/_____/ \/_____/ \/_/ \/_/ \/_____/ \/_____/ \/____/
Mesofield is a PyQt6-based framework for running real-time, multi-camera
neuroscience experiments. It coordinates hardware via serial connections
and MicroManager (through pymmcore-plus
custom MDAEngines and multi-CMMCorePlus instancing) and manages
experiment configuration, acquisition orchestration, and data logging.
The project is aimed at laboratory use and is not a full production
package; some specialised knowledge of device hardware and
MicroManager device configuration is necessary to get started.
Documentation lives at gronemeyer.github.io/mesofield and is split by audience:
- Tutorial — the fastest path from a fresh conda env to a working acquisition on your hardware.
- User Guide —
for experimenters running acquisitions: launching the GUI, writing
experiment.json, interpreting the on-disk output. - Developer Guide —
for developers extending mesofield: custom devices,
Proceduresubclasses, frame processors, threading models. - API Reference — auto-generated from docstrings.
conda create -n mesofield python=3.12 -y
conda activate mesofield
pip install -e .Launch an acquisition by pointing at your rig — experiment.json is optional:
mesofield launch path/to/hardware.yaml # rig only (author params in the GUI)
mesofield launch path/to/experiment.json # rig + params (sibling hardware.yaml auto-detected)
mesofield launch path/to/experiment/ # a directory containing eitherScaffold a new experiment:
mesofield new my-experimentFor end-to-end setup, follow the Tutorial.
Tested on Windows 10/11. For multi-camera acquisition with large files we recommend ≥ 32 GB RAM, a 12th-gen Intel i7 or equivalent, and fast local NVMe storage for the experiment directory.
MIT — see LICENSE.