Periodic Reporting for period 1 - STOICISM (Stochastic Communication Inside Cortical Microcolumns)
Periodo di rendicontazione: 2019-09-02 al 2021-09-01
The STOICISM project aimed to model and quantify the communication inside microcortical columns by modelling the multi-scale neuronal activity from intracellular signalling, synaptic channel, and cortical microcolumns. STOICISM investigated an computational platform that models and analyses the multi-scale neuronal behaviours and long-term synaptic dynamics using information and communication theory. In addition to the scientific impact, this project explored an excellent set of tools and knowledge to widen and deepen the applicant’s expertise on computational neuroscience. This project combined the researcher’s experience, the host institution’s biophysical computational, in-vitro models and, lastly, emerging data and concepts from the neuroscience community. The developed platform will enable a further understanding of cortical microcolumns’ plasticity dynamics, and will contribute to large-scale brain projects by proposing a new method of analysing their neurological models and data. The role of astrocytes inside cortical microcolumns will also be further understood, and its impact on synaptic plasticity dynamics. This whole framework bridges the gap between resources and hypothesis testing through an extensive analysis of the long-term plasticity dynamics in neurons and astrocytes and their effects in cortical microcolumns. In this project, we used multi-scale modelling (computer science), neurological modelling (neuroscience) and information theory (communications engineering).
Based on the rationale previously introduced, we seek to address the following research questions: Can the multi-scale cortical microcolumns communication model be used to uncover hidden behavior of the cortical microcolumn and explain unknown physiological behavior? How is this heterogeneous communication environment affected by dynamic plasticity variation?
Based on the rationale previously introduced, we seek to address the following research questions: Can the multi-scale cortical microcolumns communication model be used to uncover hidden behavior of the cortical microcolumn and explain unknown physiological behavior? How is this heterogeneous communication environment affected by dynamic plasticity variation?
Theoretical Analysis of Different Cortical Microcolumn Scales - WP3
The main results:
• For the network scale we have achieved the modelling in multiple ways with different advantages.
o Pre-defined networks: Pre-defined synthetic network modelling was employed to study the effects of network flow, which open opportunities for the usage of biological substrate for information processing.
o Phenomenological modelling: A general network model was employed to analyse randomly placed neurons and astrocytes through a given pre-defined 2D-3D area. This model observes burst and synchronous activity.
o Probabilistic network modelling: A probabilistic network model was used to analyse the 25 morphological type (m-type) cells, and 14 electrical type (e-type) cells. Networks were built respecting connectivity data extracted from the Blue Brain Project.
In-silico-based approach for multi-scale Cortical Microcolumn characterization - WP4
The main results:
o Non-linear ODE modelling scale coupling: In favour of numerical results, we coupled models from intracellular, synapses and network using shallow coupling factors to facilitate description of the models and higher linkage with biological understanding.
o Stochastic modelling coupling: This method used the differences in scales to characterize the probabilities of events and predict the occurrence of those in a in-silico setting. We were able to use Guillespie algorithm to model the network level incorporating lower scales.
Information and Communication Theory - WP5
The main results:
• We were able to model the microcircuit neuronal channel as SISO, MISO, SIMO and MIMO models studying their capacity, mutual information and achievable throughput.
• Multiscale noise models: We modelled the impact of noise from multiple noise characterization and also their applications
o The synaptic noise
o The global brain noise
Synaptic Plasticity Dynamics and its effects on Cortical Microcolumn Communication - WP6
The main results:
• We show that drug such as 4AP and Gabazine can be used to change network dynamics and synaptical properties in in-vitro experiments of neurons-astrocytes. The dynamics observed were reduced and provoke more synchronous behaviour than expected.
• We show that synthetic genes such as pcDNA3.1-hGPR17 as well as chemical compounds such as MDL29,951 or T0510.3657 can also be used to also decrease dynamics and provide synchronicity now for goal-oriented communications directed for information processing.
• We also show that artificial intelligence can be used to model synaptic plasticity and behaviour of neuronal networks both in pre-defined and natural settings for reduced dynamics either by chemicals but also from mechanical structures integrated to multi-electrode arrays.
The main results:
• For the network scale we have achieved the modelling in multiple ways with different advantages.
o Pre-defined networks: Pre-defined synthetic network modelling was employed to study the effects of network flow, which open opportunities for the usage of biological substrate for information processing.
o Phenomenological modelling: A general network model was employed to analyse randomly placed neurons and astrocytes through a given pre-defined 2D-3D area. This model observes burst and synchronous activity.
o Probabilistic network modelling: A probabilistic network model was used to analyse the 25 morphological type (m-type) cells, and 14 electrical type (e-type) cells. Networks were built respecting connectivity data extracted from the Blue Brain Project.
In-silico-based approach for multi-scale Cortical Microcolumn characterization - WP4
The main results:
o Non-linear ODE modelling scale coupling: In favour of numerical results, we coupled models from intracellular, synapses and network using shallow coupling factors to facilitate description of the models and higher linkage with biological understanding.
o Stochastic modelling coupling: This method used the differences in scales to characterize the probabilities of events and predict the occurrence of those in a in-silico setting. We were able to use Guillespie algorithm to model the network level incorporating lower scales.
Information and Communication Theory - WP5
The main results:
• We were able to model the microcircuit neuronal channel as SISO, MISO, SIMO and MIMO models studying their capacity, mutual information and achievable throughput.
• Multiscale noise models: We modelled the impact of noise from multiple noise characterization and also their applications
o The synaptic noise
o The global brain noise
Synaptic Plasticity Dynamics and its effects on Cortical Microcolumn Communication - WP6
The main results:
• We show that drug such as 4AP and Gabazine can be used to change network dynamics and synaptical properties in in-vitro experiments of neurons-astrocytes. The dynamics observed were reduced and provoke more synchronous behaviour than expected.
• We show that synthetic genes such as pcDNA3.1-hGPR17 as well as chemical compounds such as MDL29,951 or T0510.3657 can also be used to also decrease dynamics and provide synchronicity now for goal-oriented communications directed for information processing.
• We also show that artificial intelligence can be used to model synaptic plasticity and behaviour of neuronal networks both in pre-defined and natural settings for reduced dynamics either by chemicals but also from mechanical structures integrated to multi-electrode arrays.
The main advancements beyond the state of the art are:
We defined novel ways to characterise mathematical networks of neurons in a microcolumns with: Pre-defined networks: Pre-defined synthetic network modelling was employed to study the effects of network flow, which open opportunities for the usage of biological substrate for information processing. Probabilistic network modelling: A probabilistic network model was used to analyse the 25 morphological type (m-type) cells, and 14 electrical type (e-type) cells. Networks were built respecting connectivity data extracted from the Blue Brain Project.
We observed that neuronal goal-oriented communication can induce synthetic biocomputing approaches. We explored the hypothesis that these networks can communicate with the objective of outputting logic functions in the form of gates and circuits.
We also proposed a new direction in modelling neuronal using artificial intelligence and non-linear ODEs under dynamical plasticity. We showed that AI can either classify neuronal types tested in more than 10,000 network configuration topologies, as well as configure networks of brain cells to influence their communication towards biocomputing solutions.
Neurological pathologies and neurodegeneration are increasingly affecting people’s lives in many different forms due to the progression in life expectancy in the ageing society. For example, Alzheimer’s disease affects currently 15 million people in the US with a death rate of 29.5% among 65+ years old and produces a cost around 200 billion dollars per annum . These type of pathologies are caused by neuronal communication failures in multi-scales of the brain, Fig. 1. Current neuroscience methods fail to fully explain neurodegeneration and the multi-scale communication systems inside the brain, and in specific the cortical microcolumns . We showed that from the perspective of the molecular communications community we can i) designing synthetic communication channels using existing biological systems and ii) using information and communication theory to understand biological systems. We hope that the proposed modelling strategies for both natural and synthetic neuronal communications will be used in the near future to novel understanding of brain diseases in organ-on-chip platforms for regenerative medicine and pharmacology purposes.
We defined novel ways to characterise mathematical networks of neurons in a microcolumns with: Pre-defined networks: Pre-defined synthetic network modelling was employed to study the effects of network flow, which open opportunities for the usage of biological substrate for information processing. Probabilistic network modelling: A probabilistic network model was used to analyse the 25 morphological type (m-type) cells, and 14 electrical type (e-type) cells. Networks were built respecting connectivity data extracted from the Blue Brain Project.
We observed that neuronal goal-oriented communication can induce synthetic biocomputing approaches. We explored the hypothesis that these networks can communicate with the objective of outputting logic functions in the form of gates and circuits.
We also proposed a new direction in modelling neuronal using artificial intelligence and non-linear ODEs under dynamical plasticity. We showed that AI can either classify neuronal types tested in more than 10,000 network configuration topologies, as well as configure networks of brain cells to influence their communication towards biocomputing solutions.
Neurological pathologies and neurodegeneration are increasingly affecting people’s lives in many different forms due to the progression in life expectancy in the ageing society. For example, Alzheimer’s disease affects currently 15 million people in the US with a death rate of 29.5% among 65+ years old and produces a cost around 200 billion dollars per annum . These type of pathologies are caused by neuronal communication failures in multi-scales of the brain, Fig. 1. Current neuroscience methods fail to fully explain neurodegeneration and the multi-scale communication systems inside the brain, and in specific the cortical microcolumns . We showed that from the perspective of the molecular communications community we can i) designing synthetic communication channels using existing biological systems and ii) using information and communication theory to understand biological systems. We hope that the proposed modelling strategies for both natural and synthetic neuronal communications will be used in the near future to novel understanding of brain diseases in organ-on-chip platforms for regenerative medicine and pharmacology purposes.