A Python toolbox for large scale Calcium Imaging data Analysis and behavioral analysis.
CaImAn implements a set of essential methods required in the analysis pipeline of large scale calcium imaging data. Fast and scalable algorithms are implemented for motion correction, source extraction, spike deconvolution, and component registration across multiple days. It is suitable for both two-photon and one-photon fluorescence microscopy data, and can be run in both batch and online modes. CaImAn also contains some routines for the analysis of behavior from video cameras. A list of features as well as relevant references can be found here.
Right now, CaImAn works and is supported on the following platforms:
- Linux on Intel CPUs
- MacOS on Intel CPUs
- Windows on Intel CPUs
16G RAM is required for a good experience, and depending on datasets, 32G or more may be necessary.
CaImAn presently targets Python 3.8. Parts of CaImAn are written in C++, but apart possibly during install, this is not visible to the user. There is also an older implementation of CaImAn in Matlab (unsupported). That version can be used with MCMC spike inference
- ARM-based versions of Apple hardware work (if on a 16G model), but currently happen under x86 emulation and we cannot support them as well. A native OSX port is planned for late 2021/early 2022.
- Support for Linux on ARM (e.g. AWS Graviton) is not available (but it may work with the port of conda, if you compile Caiman yourself - we do not have binary packages and this is untested). If you care about this, please let us know.
The supported ways to install CaImAn use the Anaconda python distribution. If you do not already have it, first install a 3.x version for your platform from here. Familiarise yourself with Conda before going further.
We strongly recommend installing the mamba package into your base environment, with 'conda install -c conda-forge mamba', using it to build your conda environment. Mamba performs the same environment creation tasks that the base conda tool does, but far faster. In the instructions below, we assume you're using mamba, but if you're not, you can run the same commands with the conda tool instead.
There are a few supported install methods.
The easiest (and strongly recommended on Windows) is to use a binary conda package, installed as the environment is built. Install this with 'mamba create -n caiman -c conda-forge caiman'. This is suitable for most use, if you don't need to change the internals of the caiman package. You do not need to fetch the source code with this approach.
Another option is to build it yourself; you will need a working compiler (easy on Linux, fairly easy on OSX, fairly involved on Windows). Clone the sources of this repo, create an environment with all the prereqs with 'mamba env create -n caiman -f environment.yml', activate the environment, and then do a 'pip install .' or 'pip install -e .' The former is a user install, the latter is more suitable for active development on the caiman sources.
There are other ways to build/use caiman, but they may get less or no support depending on how different they are.
More detailed docs on installation can be found here.
After installing the software, the caimanmanager.py script (which will be put in your path on Linux and OSX) is used to unpack datafiles and demos into a directory called caiman_data.
If you want to use GPU functionality and have a GPU where you're running CaImAn (most likely a Linux system), you'll want, after you build your conda environment, to switch to a GPU build of the tensorflow package (conda list will tell you, after the version string, what build variant you have - you most likely will get an mlk build, but a "conda search tensorflow" will probably show you some gpu variants you can switch to - pick one appropriate for your conda version, ideally of the same version of tensorflow you otherwise got). If you need help switching versions, reach out to us on the gitter channel.
If you used caimanmanager to unpack the demos and data files, you will find in the caiman_data folder a set of demos and jupyter notebooks. demo_pipeline.py and demo_behavior.py (or their notebook equivalents) are good introductions to the code.
A paper explaining most of the implementation details and benchmarking can be found here.
@article{giovannucci2019caiman,
title={CaImAn: An open source tool for scalable Calcium Imaging data Analysis},
author={Giovannucci, Andrea and Friedrich, Johannes and Gunn, Pat and Kalfon, Jeremie and Brown, Brandon L and Koay, Sue Ann and Taxidis, Jiannis and Najafi, Farzaneh and Gauthier, Jeffrey L and Zhou, Pengcheng and Khakh, Baljit S and Tank, David W and Chklovskii, Dmitri B and Pnevmatikakis, Eftychios A},
journal={eLife},
volume={8},
pages={e38173},
year={2019},
publisher={eLife Sciences Publications Limited}
}
All the results and figures of the paper can be regenerated using this package. For more information visit this page.
CaImAn implements a variety of algorithms for analyzing calcium (and voltage) imaging data. A list of references that provide the theoretical background and original code for the included methods can be found here.
If you use this code please cite the corresponding papers where original methods appeared as well the companion paper.
Our online algorithms can be used for real-time analysis of live-streaming data. An example for real-time analysis of microendoscopic 1p data is shown in the notebook demos/notebooks/demo_realtime_cnmfE.ipynb.
For more information about the approach check the paper.
VolPy is an analysis pipeline for voltage imaging data. The analysis is based on following objects:
MotionCorrect: An object for motion correction which can be used for both rigid and piece-wise rigid motion correction.volparams: An object for setting parameters of voltage imaging. It can be set and changed easily and is passed into the algorithms.VOLPY: An object for running the spike detection algorithm and saving results.
The object detection network Mask R-CNN in VolPy is now compatible with tensorflow 2.4.1.
To see examples of how these methods are used, please consult the demo_pipeline_voltage_imaging.py script in the demos/general folder. For more information about the approach check the preprint.
There is also a general paper on this pipeline
Documentation of the code can be found here.
Other docs:
- Eftychios A. Pnevmatikakis, Flatiron Institute, Simons Foundation
- Andrea Giovannucci, University of North Carolina, Chapel Hill
- Johannes Friedrich, Flatiron Institute, Simons Foundation
- Changlia Cai, University of North Carolina, Chapel Hill
- Pat Gunn, Flatiron Institute, Simons Foundation
A complete list of contributors can be found here.
Currently Pat Gunn and Johannes Friedrich are the most active maintainers.
For support, you can create a Github issue describing any bugs you wish to report, or any feature requests you may have.
You may also use the gitter chat room for discussion.
Finally, you may reach out via email to one of the primary maintainers (above).
Special thanks to the following people for letting us use their datasets in demo files:
- Weijian Yang, Darcy Peterka, Rafael Yuste, Columbia University
- Sue Ann Koay, David Tank, Princeton University
- Manolis Froudarakis, Jake Reimers, Andreas Tolias, Baylor College of Medicine
- Clay Lacefield, Randy Bruno, Columbia University
- Daniel Aharoni, Peyman Golshani, UCLA
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.