Fig. 5: Emulation of the brain’s sound localization function based on dual-gate V-EGT. | Nature Communications

Fig. 5: Emulation of the brain’s sound localization function based on dual-gate V-EGT.

From: A low-power vertical dual-gate neurotransistor with short-term memory for high energy-efficient neuromorphic computing

Fig. 5: Emulation of the brain’s sound localization function based on dual-gate V-EGT.The alternative text for this image may have been generated using AI.

a Principle of the brain recognizing sound azimuth, that is, based on the trigonometric relationship between interaural time difference (ITD) and sound azimuth θ, the brain can identify the sound azimuth through ITD. b Relationship between ITD and θ. c Principle of sound source distance computation by the brain. The farther the sound source is from the ear, the weaker the intensity of the sound heard by the ear. Specifically, the sound intensity is inversely proportional to the square of the distance. d Relationship between sound intensity and distance. e Schematic of the neural network for sound azimuth and distance identification. Blue spheres, blue lines, blue nodes, green or yellow lines, and green or yellow spheres represent (sound) sensing neurons, axons, synapses, dendrites, and post sound information processing neurons, respectively. The synaptic weight matrix between the two green post neurons and pre-neurons is diagonal. The yellow post-neuron has the same synaptic connection weights as two pre-neurons. f Hardware neural network for sound azimuth and distance recognition based on dual-gate V-EGTs. g Equivalent circuit diagram of (f). h, i An example of evaluating sound azimuth (ITD = 2 s). The intensity and sequence of sound signal spikes received by each dendrite of post neurons 1 and 3 are shown, as well as the outputs. j Recognition for all sound azimuths by our hardware neural network. k Emulation of the brain’s sound source distance computation by our hardware neural network.

Back to article page