Magnetic Resonance Elastography (MRE) quantifies soft tissue stiffness by measuring induced shear waves. MRE inversion techniques for parameter reconstruction are often affected by noise and compression waves. Neural network-based inversions have emerged as a possible solution to address these challenges. However, current approaches lack generalizability and do not provide uncertainty estimates. Therefore, we propose ElastoNet, a novel neural network-based approach for MRE wave inversion that analyzes multiple wave components independently of resolution and vibration frequency and provides uncertainty quantification maps. ElastoNet was trained on synthetically generated wave patches of 5 × 5 pixels. Uncertainty quantification was implemented using evidential deep learning. ElastoNet was evaluated on synthetically generated plane waves, finite element simulations of abdominal MRE, phantom MRE data, and a prospective wideband multifrequency abdominal MRE study (excitation frequencies of 20 to 80 Hz) in 14 healthy volunteers. ElastoNet was compared with established inversion methods LFE and k-MDEV, as well as neural network-based TWENN. ElastoNet generated shear wave speed maps as a proxy of stiffness with comparable or better accuracy than established methods and did not require retraining for different resolutions and vibration frequencies. ElastoNet achieved a lower root mean square error relative to ground truth values in finite element simulations and phantom data than other inversion methods and provided uncertainty maps. ElastoNet is a promising method for universal neural network-based inversion in MRE, effectively overcoming current challenges and expanding the potential use of neural networks in diagnostic MRE applications.

A fundamental function of cortical circuits is the integration of information from different sources to form a reliable basis for behavior. While animals behave as if they optimally integrate information according to Bayesian probability theory, the implementation of the required computations in the biological substrate remains unclear. We propose a novel, Bayesian view on the dynamics of conductance-based neurons and synapses which suggests that they are naturally equipped to optimally perform information integration. In our approach apical dendrites represent prior expectations over somatic potentials, while basal dendrites represent likelihoods of somatic potentials. These are parametrized by local quantities, the effective reversal potentials and membrane conductances. We formally demonstrate that under these assumptions the somatic compartment naturally computes the corresponding posterior. We derive a gradient-based plasticity rule, allowing neurons to learn desired target distributions and weight synaptic inputs by their relative reliabilities. Our theory explains various experimental findings on the system and single-cell level related to multi-sensory integration, which we illustrate with simulations. Furthermore, we make experimentally testable predictions on Bayesian dendritic integration and synaptic plasticity.

Simulation is a third pillar next to experiment and theory in the study of complex dynamic systems such as biological neural networks. Contemporary brain-scale networks correspond to directed random graphs of a few million nodes, each with an in-degree and out-degree of several thousands of edges, where nodes and edges correspond to the fundamental biological units, neurons and synapses, respectively. The activity in neuronal networks is also sparse. Each neuron occasionally transmits a brief signal, called spike, via its outgoing synapses to the corresponding target neurons. In distributed computing these targets are scattered across thousands of parallel processes. The spatial and temporal sparsity represents an inherent bottleneck for simulations on conventional computers: irregular memory-access patterns cause poor cache utilization. Using an established neuronal network simulation code as a reference implementation, we investigate how common techniques to recover cache performance such as software-induced prefetching and software pipelining can benefit a real-world application. The algorithmic changes reduce simulation time by up to 50%. The study exemplifies that many-core systems assigned with an intrinsically parallel computational problem can alleviate the von Neumann bottleneck of conventional computer architectures.