Neurogrid simulates cortical cell-types, active dendrites, and top-down attention

Abstract

A central challenge for systems neuroscience and artificial intelligence is to understand how cognitive behaviors arise from large, highly interconnected networks of neurons. Digital simulation is linking cognitive behavior to neural activity to bridge this gap in our understanding at great expense in time and electricity. A hybrid analog–digital approach, whereby slow analog circuits, operating in parallel, emulate graded integration of synaptic currents by dendrites while a fast digital bus, operating serially, emulates all-or-none transmission of action potentials by axons, may improve simulation efficacy. Due to the latter’s serial operation, this approach has not scaled beyond millions of synaptic connections (per bus). This limit was broken by following design principles the neocortex uses to minimize its wiring. The resulting hybrid analog–digital platform, Neurogrid, scales to billions of synaptic connections, between up to a million neurons, and simulates cortical models in real-time using a few watts of electricity. Here, we demonstrate that Neurogrid simulates cortical models spanning five levels of experimental investigation: biophysical, dendritic, neuronal, columnar, and area. Bridging these five levels with Neurogrid revealed a novel way active dendrites could mediate top-down attention.

DYNAP-SE2: a scalable multi-core dynamic neuromorphic asynchronous spiking neural network processor

“…The ability to configure synapses as diffusive gap junctions [63] with 2D nearest neighbor connections supports the configuration of networks with local spatially distributed connectivity kernels, as originally proposed in [36,64]. In addition, excitatory synapse circuits can be configured to emulate both slow voltage-gated NMDA receptor dynamics [31] as well as fast AMPA dynamics [35].…”

“…Figure 2. Neuronal compartments 64. synapses with 4 bit weights and conditional delay and short-term plasticity (STP) convert pre-synaptic spikes to pulses.…”

Bio-realistic neural network implementation on Loihi 2 with Izhikevich neurons

“…Notably, some architectures draw inspiration from biological neural networks, including those mimicking the brain's navigational system [25,26] and a canonical columnar cortical circuit, emphasizing their potential impact on functions like perceptual organization, motion detection, and attention [27]. Bio-realistic neural network implementation on neuromorphic hardware: There are studies that simulate bio-realistic neural networks on neuromorphic chips, such as the simulations of top-down and bottom-up interactions between visual cortical area (V4) and frontal cortical area [28]; and architectures inspired by the cerebellum for efficient supervised learning [29]. In [30], the simple network with 1000 neurons given in [20] is realized on neuromorphic compute node by using Izhikevich neurons.…”