FOCIA: Fast, Open, Cellular Imaging Analysis toolbox

Lead Research Organisation: University of Edinburgh
Department Name: Sch of Informatics


Understanding brain activity remains a substantial scientific challenge which is essential both for basic science as well as for the diagnosis and potential treatments of neurological diseases, such as Alzheimer's, epilepsy, autism and stroke. A critical step towards understanding brain mechanisms, is to be able to measure activity from it. Due to the recent developments in imaging and biochemical tools, it is now possible to measure activity from populations of neurons in the living brain, using cellular calcium imaging. Fluorescent, activity-dependent indicators are used, which 'light up' (fluoresce) whenever the neuron is active.
This new technique allows scientists to study networks of hundreds of individual neurons in parallel and has already yielded crucial basic neuroscience results as well as critical insights in brain pathologies and disorders, through experiments that were thought impossible not too long ago. The signal from these dyes is however weak and noisy. In addition, because the fluorescence images are taken through a microscope, they are highly sensitive to even small motion of the tissue. Consequently the images taken require extensive image manipulation and data extraction procedures. In practice this mean that a 10 minute experiment is followed by hours or even days of computer analysis. In addition, current methods are not only slow, they are also poorly standardized, each lab developing its own analysis tools. This lack of standardization impairs reproducibility and comparison of the data across different laboratories. A direct consequence is that it leads to a duplication of efforts and experimental costs in terms of time, consumables and number of animals used for these experiments.

In this proposal, the Informatics and Neuroscience departments of the University of Edinburgh join strengths to develop a fast analysis software for calcium imaging experiments by using the latest insights from image processing and artificial intelligence. As in the scientific community an open source approach has proven to yield the highest quality and most reliable software, the software will be released in open source format so that laboratories across the country and across the world can use it and adapt it.
Thus, our specific objectives are:
1. To standardize and speed-up analysis methods of cellular imaging.
2. Validate the toolbox on in vivo data from different laboratories and experimental conditions.
3. Create and promote an integrated open source platform for the analysis of cellular imaging experiments.

Considering the rapidly increasing number of laboratories that are using cellular imaging across the world, this toolbox will substantially reduce duplication of efforts and experimental costs, will allow new experiments and will speed up discoveries in brain research.

Technical Summary

Cellular imaging of neuronal activity is rapidly gaining ground as the dominant technique to measure the activity of neurons in systems neuroscience, with important applications in basic science and translational research. By imaging population of neurons activity in the brain of awake behaving animals, this technique allows linking activity of sub-types of neurons to behaviour over temporal scales from milliseconds to months. However, compared to the rapid advances in experimental methods, computational analysis of imaging data remains poorly developed. Current methods are slow and poorly standardised, strongly limiting the experiments. This lack of standardization impairs reproducibility and, as a direct consequence, leads to increased experimental costs in terms of time, consumables and number of experimental animals.
The proposed collaboration between computer science and two-photon imaging laboratories aims at developing an analysis software for cellular imaging data, that allows for fast interpretation of neuronal activity.
Our objectives are:
1. To standardize and speed-up analysis methods of cellular imaging.
2. Validate the toolbox on in vivo data from different laboratories, species and imaging set-ups.
3. Create and promote an integrated open source platform for the analysis of cellular imaging experiments.

In order to dramatically reduce the analysis time of cellular imaging data, we are focusing on three bottlenecks: 1) motion correction, 2) regions of interest selection and 3) signal de-contamination from surrounding fluorescence. Preliminary results have already yielded a ten-fold speed-up compared to traditional tools. By releasing and supporting the toolbox as an open source package, the community will benefit maximally from this investment, improving the reproducibility of basic and translational neuroscience research, increasing the yield of experiments, and reducing the duplication of effort of many laboratories across the world

Planned Impact

The main beneficiaries of the project are researcher that use cellular calcium imaging to measure neural activity. They will benefit not only directly from having highly quality, standardized software but also by:
1) saving time and effort on writing their own tools,
2) increasing the reproducibility of their research and the ability to quickly re-analyze data from their own or from other laboratories,
3) being able to carry out experiments more interactively and with a higher yield.
4) reducing the number of experiments and, as a consequence, the cost of consumables and animals used for these studies

As a consequence, society will benefit from having research in basic and translational neuroscience carried out more quickly and more cost-effectively.

To maximize the take up of the developed software toolbox, we will 1) solicit feedback to increase usability and flexibility of the tools, 2) promote the tools at conferences through posters, presentation, and informal contacts with our large network of fellow scientists, 3) support users of the toolbox with documentation, example data sets, and a discussion forum.
The software will be released under the GNU public license (GPL), which means that anyone is free to use and modify the software, but not allowed to resell it. External organizations will also be able to benefit from the expertise developed through the project through consultancy agreements with the PIs and appointed staff.

Both the training of the RA and the exposure of other lab members to the latest computational approaches and tools, will help to address the current lack of computationally trained neuroscientists. The researchers involved in this project will learn computational analysis tools and imaging methods which they could implement in the future in other laboratories, including those of pharmaceutical companies that are using imaging approaches to screen the effect of potential medications in the living brain. These skills are also in high demand in microscopy companies such as Scientifica (UK based), whose recent and rapid expansion is directly related to the expanding market of two-photon microscopy. The training delivered in the context of this grant will thus contribute to the UK's competitiveness and further emphasize the UK's prime position for integrating computer and biological sciences.


10 25 50
Description In vivo calcium imaging has become a method of choice to image neuronal population activity throughout the nervous system. These experiments generate large sequences of images. Their analysis is computationally intensive and typically involves motion correction, image segmentation into regions of interest (ROIs), and extraction of fluorescence traces of each ROI. Out of focus fluorescence from surrounding neuropil and other cells can strongly contaminate the signal assigned
to a given ROI. In this study, we introduce the FISSA toolbox (Fast Image Signal Separation Analysis) for neuropil decontamination. Given pre-defined ROIs, the FISSA toolbox automatically extracts the surrounding local neuropil and performs blind-source separation with non-negative matrix factorization. Using both simulated and in vivo data, we show that this toolbox performs similarly or better than existing published methods. FISSA requires only little RAM, allowing for fast processing of large datasets even on a standard laptop. The FISSA toolbox is available in Python, with an option for MATLAB format outputs, and can easily be integrated into existing workflows. It is available from Github and the standard Python repositories.
Exploitation Route The project develops a toolbox to use in conjunction with cellular Calcium imaging. We envisage uptake be a large number of labs using such imaging techniques.
Sectors Digital/Communication/Information Technologies (including Software),Healthcare,Pharmaceuticals and Medical Biotechnology,Other

Title FISSA, Fast Image Signal Separation Analysis 
Description A neuropil decontamination toolbox for calcium imaging signals. 
Type Of Technology Webtool/Application 
Year Produced 2018 
Open Source License? Yes  
Impact This software will dramatically speed up certain steps in the computational processing of calcium imaging signals.