A Sparse Reformulation of the Green's Function Formalism Allows Efficient Simulations of Morphological Neuron Models

Area of Science:

• Computational Neuroscience
• Mathematical Biology
• Applied Mathematics

Background:

• Partial differential equations (PDEs) are crucial for modeling biological systems.
• The cable equation and its generalizations model phenomena like signal propagation in neurons.
• Existing computational methods for these PDEs on tree graphs often exhibit quadratic scaling.

Purpose of the Study:

• To develop a computationally efficient method for simulating PDEs on tree graphs.
• To reduce the computational complexity of the Green's function (GF) formalism for specific PDE classes.
• To enable faster and more scalable simulations of biological models, such as neuronal signal propagation.

Main Methods:

• Reformulating the Green's function (GF) formalism for PDEs on tree graphs with discrete input locations.
• Achieving a linear O(n) scaling with the number of input locations (n).
• Combining linear scaling with an exponential sum expansion for simulation kernels.

Main Results:

• Demonstrated a shift from O(n^2) to O(n) scaling for the GF formalism.
• Validated the simulation paradigm on models of nerve cells.
• Showcased potential for significant computational performance gains compared to traditional methods like finite differences.

Conclusions:

• The developed O(n) GF method offers a significant computational advantage for simulating a class of PDEs on tree graphs.
• This approach is particularly relevant for accelerating simulations in computational neuroscience and related fields.
• The method provides a scalable and efficient alternative to existing numerical techniques.