Periodic Reporting for period 1 - ISLAND (An integrated setup for in-vitro optogenetic experiments using AI to localize stimulation with a feedback of electrophysiological signals)
Reporting period: 2021-04-12 to 2023-04-11
The study of the brain remains one of the most challenging topics in science. A full understanding of the neuronal dynamics inside a living brain is still far from being achieved. Neuroscience tackles this problem by studying the neuronal functionality as individuals and in networks. For this purpose, different disciplines and approaches are used, e.g. electrophysiology, biochemistry, optogenetics and more.
The main aim of project "ISLAND" is to develop an experimental platform for optogenetic studies in-vitro with a feedback from electrophysiological signals. In-vitro studies provide a simplified model of the brain function, by probing the network in 2D or 3D, in the pursuit of obtaining a comprehensive picture of the brain activity. The system proposed in the project can be a very useful tool to perform optogenetic experiments in-vitro, to study neural network circuity in a controlled environment, conveniently and efficiently.
The setup we propose consists of:
- “Writing” unit- An optical system providing light for stimulating the neuronal culture (genetically modified for expressing light-sensitivity).
- “Reading” unit- A microelectrode array (MEA) which samples the electrophysiological signals generated by the neuronal culture.
- Processing and control unit- Analyses the electrophysiological signals and provides feedback to the optical stimulation system, which generates accordingly the stimulation pattern.
During the project we developed an integrated experimental setup which includes light stimulation system, based on digital light processor (DLP); and an electrophysiological recording system- microelectrode array (MEA). Then we developed an AI-based software which is able to map macroscopically the neuronal network under test and to simulate the dynamics of this network. This software serves as feedback from the electrophysiological data which provides information on the morphology, functionality and connectivity of the neuronal network.
The main aim of project "ISLAND" is to develop an experimental platform for optogenetic studies in-vitro with a feedback from electrophysiological signals. In-vitro studies provide a simplified model of the brain function, by probing the network in 2D or 3D, in the pursuit of obtaining a comprehensive picture of the brain activity. The system proposed in the project can be a very useful tool to perform optogenetic experiments in-vitro, to study neural network circuity in a controlled environment, conveniently and efficiently.
The setup we propose consists of:
- “Writing” unit- An optical system providing light for stimulating the neuronal culture (genetically modified for expressing light-sensitivity).
- “Reading” unit- A microelectrode array (MEA) which samples the electrophysiological signals generated by the neuronal culture.
- Processing and control unit- Analyses the electrophysiological signals and provides feedback to the optical stimulation system, which generates accordingly the stimulation pattern.
During the project we developed an integrated experimental setup which includes light stimulation system, based on digital light processor (DLP); and an electrophysiological recording system- microelectrode array (MEA). Then we developed an AI-based software which is able to map macroscopically the neuronal network under test and to simulate the dynamics of this network. This software serves as feedback from the electrophysiological data which provides information on the morphology, functionality and connectivity of the neuronal network.
The beginning of the project was dedicated to the establishment of the procedure of plating E17 cortical cells and maintaining them on top of MEA chips. Along with biology fellow researchers, we have consolidated the optimal protocol to plate neurons on top of the glass MEA chip and maintaining them until a sufficient age (more than 30 days in-vitro), including the culture infection with a viral vector for expressing ChR2 protein (for light sensitivity). In parallel, the MEA hardware and software (including different accessories) were installed in the lab, after which custom software tools for data processing and analysis were developed. During this period, the ER went to the University of Genoa (Italy) for training in topics related to neuro-engineering, including MEA measurements and data analysis.
Following these procedures, we started to perform basic electrophysiological recordings with the MEA system, where we tested basic neuronal activity starting from spontaneous activity and followed by evoked response to electrical stimuli and long-term potentiation (LTP) protocols by electrical stimulation.
In the next step we installed the optical setup for optogenetic stimulation of ChR2 infected cultures. The digital light processor (DLP) system, including the optical instrumentation were installed in the laboratory. The DLP is a programable light projecting device based on digital micromirror device (DMD), which can create custom illumination patterns. The patterns are then projected on top of the sample and constitute a stimulation trigger for the genetically-modified neurons. Later we integrated the DLP and MEA to a joint operation. We then tested and characterized the response of neurons to light stimulation. Then in further experiments we tackled the question of memory formation in neuronal culture following light stimulation.
The last part of the project was devoted to developing advanced analytical tools with the aim of building a processing and control unit that reads the electrophysiological signals from MEA recordings and creates a corresponding map of the tested network, which further can control the stimulation location and pattern. An AI-based computational model was developed. This model obtains a simplified network (or a graph model), using the data acquired in MEA measurements as training dataset in order to decode the spatio-temporal signal patterns of the neuronal network and to extract its structure. The model is based on Reservoir computer network (RCN) approach, as it uses the fact that around each electrode of MEA, a complex neuronal circuit is embedded. The complexity of these circuits cannot be easily understood from the standard measurement analysis, and hence they are modeled as nonlinear networks with inner random connections. With this model we are able to extract a macroscopic graph representing the structure of the culture under test, where each node of this network represents a neural circuit (population of neurons); and the connections (edges) between them represent the weighted interaction between the populations.
The results of the ISLAND project will allow further applications and studies in-vitro and ex-vivo. The integrated optogenetic setup that has been developed in the project is already in use for further research of epileptic seizure suppression by light at NanoLab. The AI-model that has been developed during the project is able to retrieve functional connectivity of the tested neuronal network with high accuracy and hence various applications might arise from this work, such as pharmacological studies, brain mapping and more.
Dissemination:
1. 11th Optoelectronics and Photonics Summer School: NMP2021 – Neuromorphic Photonics, 20-26 June 2021, Monte Bondone, Trento, Italy. A talk introducing MSCA individual fellowship program and the ISLAND project was given (title: “Between biological and artificial neurons and MSCA-IF ISLAND project”).
2. International School on Bio hybrid interfaces, organic bio electronics and bio photonics Lake Como School of Advanced Studies, 5-9 September 2022, Como, Italy. A poster of the recent work in the ISLAND project was presented (title: “An integrated setup for in-vitro optogenetic experiments using AI to localize stimulation with a feedback of electrophysiological signals”).
3. SPIE Photonics West, 28 January - 3 February 2023, San Francisco, CA, USA. A poster about the ISLAND project was presented (title: “An integrated setup for in-vitro optogenetic experiments using AI to localize stimulation”).
4. (SUBMITED PAPER) 20th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (29-31 August 2023 – Eindhoven, The Netherlands)- a talk about the developed AI model is planed.
Following these procedures, we started to perform basic electrophysiological recordings with the MEA system, where we tested basic neuronal activity starting from spontaneous activity and followed by evoked response to electrical stimuli and long-term potentiation (LTP) protocols by electrical stimulation.
In the next step we installed the optical setup for optogenetic stimulation of ChR2 infected cultures. The digital light processor (DLP) system, including the optical instrumentation were installed in the laboratory. The DLP is a programable light projecting device based on digital micromirror device (DMD), which can create custom illumination patterns. The patterns are then projected on top of the sample and constitute a stimulation trigger for the genetically-modified neurons. Later we integrated the DLP and MEA to a joint operation. We then tested and characterized the response of neurons to light stimulation. Then in further experiments we tackled the question of memory formation in neuronal culture following light stimulation.
The last part of the project was devoted to developing advanced analytical tools with the aim of building a processing and control unit that reads the electrophysiological signals from MEA recordings and creates a corresponding map of the tested network, which further can control the stimulation location and pattern. An AI-based computational model was developed. This model obtains a simplified network (or a graph model), using the data acquired in MEA measurements as training dataset in order to decode the spatio-temporal signal patterns of the neuronal network and to extract its structure. The model is based on Reservoir computer network (RCN) approach, as it uses the fact that around each electrode of MEA, a complex neuronal circuit is embedded. The complexity of these circuits cannot be easily understood from the standard measurement analysis, and hence they are modeled as nonlinear networks with inner random connections. With this model we are able to extract a macroscopic graph representing the structure of the culture under test, where each node of this network represents a neural circuit (population of neurons); and the connections (edges) between them represent the weighted interaction between the populations.
The results of the ISLAND project will allow further applications and studies in-vitro and ex-vivo. The integrated optogenetic setup that has been developed in the project is already in use for further research of epileptic seizure suppression by light at NanoLab. The AI-model that has been developed during the project is able to retrieve functional connectivity of the tested neuronal network with high accuracy and hence various applications might arise from this work, such as pharmacological studies, brain mapping and more.
Dissemination:
1. 11th Optoelectronics and Photonics Summer School: NMP2021 – Neuromorphic Photonics, 20-26 June 2021, Monte Bondone, Trento, Italy. A talk introducing MSCA individual fellowship program and the ISLAND project was given (title: “Between biological and artificial neurons and MSCA-IF ISLAND project”).
2. International School on Bio hybrid interfaces, organic bio electronics and bio photonics Lake Como School of Advanced Studies, 5-9 September 2022, Como, Italy. A poster of the recent work in the ISLAND project was presented (title: “An integrated setup for in-vitro optogenetic experiments using AI to localize stimulation with a feedback of electrophysiological signals”).
3. SPIE Photonics West, 28 January - 3 February 2023, San Francisco, CA, USA. A poster about the ISLAND project was presented (title: “An integrated setup for in-vitro optogenetic experiments using AI to localize stimulation”).
4. (SUBMITED PAPER) 20th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (29-31 August 2023 – Eindhoven, The Netherlands)- a talk about the developed AI model is planed.
The results of the ISLAND project will allow further applications and studies in-vitro. The integrated optogenetic setup that has been developed in the project is already in use for further research of epileptic seizure suppression by light. The AI-model that has been developed during the project is able to retrieve functional connectivity of the tested neuronal network with high accuracy. It is expected to be among the state-of-the-art algorithms for retrieval of neuronal functional connectivity, therefore it can be applied to different studies in relying on multisite electrophysiological signals (for example pharmacology and brain mapping).