CNEP: Science and technology

Our core technology is best explained by the name of the company: the ab initio computational neuroelectrophysiology (CNEP) technology.

Our investigation and modeling of synergetic interactions amongst LGIC (ligand-gated ion channels), VGIC (voltage-gated ion channels), and DGIC (dual-gated ion channels) in the cytomembrane of a cortical neuron are conducted at two spatial and temporal scales. Take glutamatergic cortical pyramidal cells as example, we first model and simulate the coupling of electrical and chemical signals in a local area of cytomembrane with characteristic spatial scale of a membrane patch consisting of a few non-gated and/or gated ionic channels and temporal scale of a few milliseconds, and then model and simulate the whole-cell signal propagation, integration, and transformation involving multiple patches with typical spatial scale of microns and temporal scale up to hundreds of milliseconds. In brief, single-cell CNEP investigation is conducted at intra-patch and inter-patch levels, respectively.  

With CNEP, investigation into these topics is done in an interdisciplinary manner. While individual structures and functions of LGIC, VGIC, and DGIC have been intensively investigated thanks to advances in structural biology and in new imaging and recording tools and methods, the interrelation and interaction amongst these three types of ionic channels are much less known and is a central interest of CNEP. At intra-patch level, these inter-channel interactions are modeled and numerically simulated for each of all possible patch configurations. Then, these simulation results are compared and verified with established neurophysiological data. This intra-patch investigation provides a basic information on how gated channels initiate, regulate, facilitate, and relay membrane potentials and currents within a patch and how they provide transformation mechanisms between electric and chemical messengers.

At inter-patch level, these verified patch models are plugged into various dendrograms (dendritic structures) and simulated with structured stimulus inputs at synaptic contacts that a neuron has with other neurons. Given that the number of all possible combinations of patch configurations, dendrograms, and input patterns can be huge, it is formalized as an ‘acquired evolution’ problem where we employ certain AI algorithms to explore the entire space of possible combinations, identifying those that make physiological sense and would likely be chosen by biological selection and others which are likely the losers of competition. Those in the winner set are further compared with database of experiments in search of their biological counterparts. With the assistance of certain numerical simulation tools and AI-powered algorithms we aim to conduct an exhaustive coverage of the configuration space. Again, the simulated results are to be verified and compared with neurophysiological discoveries. Inter-patch interaction models are also developed for soma, axon, and presynaptic terminals.

With CNEP, synaptic efficacies are described in terms of efficiencies of inter-patch interaction between a group of presynaptic patches (patches that pave the presynaptic terminal) and a group of postsynaptic patches (patches that pave the postsynaptic density or spine). This extension of inter-patch interaction model from within a single neuron to across two neurons unifies the description of neural activity and neural connectivity. Representing synaptic efficacy in terms of cross-cell inter-patch interactions allows a systematic investigation of various presynaptic and postsynaptic ionic channel mechanisms that contribute to the efficacy. 

With CNEP, synaptic plasticity is defined in an activity-dependent manner. If C is the efficacy of synaptic connection from a presynaptic neuron to a postsynaptic neuron, then the plasticity (LTP or LTD) of C is considered to be determined by activities in the group of presynaptic patches and the group of postsynaptic patches which participate in the construction of C, and corresponds to reconfigurations of one or a plurality of presynaptic and / or postsynaptic patches. Synergetic interactions among ionic channels remain to be the basic language for CNEP to investigate the dynamics of synaptic efficacy and plasticity. 

From single neuron to cortical circuits

We start with pyramidal cells in mammalian visual cortex. This involves models with glutamatergic intra-patch and inter-patch interactions for most of component segments including PSD and spines, dendrites, soma, Hillock and axon. Then, the modeling work will extend to include other excitatory and inhibitory neurotransmitters and their associated receptors. Accordingly, new patch models concerning new messengers and gated channels will be added and the whole-cell, end-to-end signal reconstruction models will be developed for other types of neurons that are important members to form meaningful functional circuits in visual cortex. This is also the stage where cross-cell and multi-cell inter-patch interactions shall be considered for those local neuronal circuits (known as functional columns) typically consisting of tens of thousands of neurons of different types. With the established knowledge regarding anatomy, physiology, and function of mammalian visual cortex, an ab initio reconstruction of visual cortex functions from ionic channel activities should have no fundamental difficulties.

Some mid- to long-term steps can be projected with a natural extension to apply the modeling to ‘what’ and ‘where’ pathways so that a full visual perception system can be reconstructed from the neural activities at levels of ionic channels. Similar approach applies to other sensory perceptions such as auditory and olfactory systems.