Fig. 1. (Color online) Schematic structure of this review's specific contents^[30−39].

Fig. 2. (Color online) Schematic diagram of sensory nerve systems in the human brain^[40].

Fig. 3. (Color online) Emerging silicon neurons based on latch-up effect. (a) A 1T neuron fabricated by silicon nanowire; (b) energy band diagrams of the 1T neuron device for the excitatory function (left and middle) and the inhibitory function (right); (c) real time spiking characteristics modulated by the gate voltage (V_G) pulse in which the 1T neuron is significantly suppressed when the V_G is 1 V^[33].

Fig. 4. (Color online) (a) 3D device structure of sub-10 nm vertical ITO transistor^[47]; (b) 3D device structure of sub-10 nm vertical coplanar-multiterminal flexible transient ITO phototransistor network^[48].

Fig. 5. (Color online) (a) Typical I−V curve of a NbO_x-based TS memristor; (b) schematic diagram of artificial spiking neuron circuit based on NbO_x-based memristor; (c) oscillation and output spiking characteristics of memristive neuron under constant voltage stimuli^[54]; (d) summary of effect of TS device performance on neuron characteristics^[66]; (e) TS characteristics before and after the RESET operation with a cutoff voltage of −3 V; (f) V_th variation tendency upon different RESET voltages^[36].

Fig. 6. (Color online) (a) LIF neurons with frequency-adaptive firing functions^[69]; (b) the volatile typical transfer curves (up) of an antiferroelectric field effect transistor (AFeFET) and the continuous firing events (down) of an AFeFET neuron under voltage pulses; (c) the dynamic of leaky and integration process of AFeFET neuron under gate pulse with different amplitudes^[30]; (d) schematic of an antiferromagnetic spintronic device; (e) schematic of a polar magneto-optic Kerr effect microscope setup for in-situ magneto-electrical transport probing (up) and the measured domain wall position of hall bar under current stimuli (down); (f) domain wall position signal (up), neural threshold signal (middle) and the output voltage spike dynamics (down) of the antiferromagnetic spintronic device under current stimuli, the inset presents dynamics of domain wall motion^[38].

Fig. 7. (Color online) (a) Schematic diagram of human visual system (up) and an artificial visual neuron composed of a UV sensor and an oscillation neuron (down); (b) the output oscillatory spikes of the artificial visual neuron under simultaneous UV irradiation of 254 and 365 nm; (c) the segmentation result of UV image based on the artificial visual system^[89]; (d) an artificial visual perception system based on a filament-based TS neuron; (e) evolution of parameters of the artificial visual perception system during the car passing process^[35]; (f) schematic illustration of the eye self-adaptation to near and distant vision; (g) circuit scheme of an artificial visual neuron for near and distant vision stimulation; (h) the relationship between firing rate and distance of the artificial visual neuron^[90].

Fig. 8. (Color online) Optoelectronic neuron devices. (a) TEM image of an optoelectronic neuron based on silicon transistor; (b) spiking characteristics of the optoelectronic neuron based on silicon transistor under red light stimuli with different intensities; (c) spiking frequency change of the silicon neuron under various intensities of red, blue, and green light stimuli^[95]; (d) an artificial eye based on FLBP/CsPbBr₃-based TS memristor; (e) a collision detection system based on the artificial eye; (f) decision-making for a robot car with optic signal processing ability^[97]; (g) schematic structure of a MoS₂-based optoelectronic graded neuron; (h) photocurrent changes of the optoelectronic graded neuron with time under single light pulse stimuli; (i) motion direction recognition with high-efficiency based on the optoelectronic graded neuron^[99]; (j) schematic diagram of polarization-sensitive photodetection system based on ReS₂ phototransistor; (k) transmission characteristics of ReS₂ phototransistor at different polarization angles^[100].

Fig. 9. (Color online) (a) Illustration of an artificial tactile neuron consisting of a NbO_x-based memristor and a piezoelectric sensor; (b) the neural firing dynamics of the artificial tactile neuron under pressure; (c) the self-protective behavior of the artificial tactile neuron under high pressure^[102]; (d) an artificial tactile neuron composed of a resistive pressure sensor with micro-pyramid and NbO_x-based memristor; (e) the spiking frequency mapping of the artificial tactile neuron array under handprint pressure; (f) the output grayscale image of handprint after the PCNN processing^[103]; (g) the output currents at a different input voltage amplitudes in log scale (up) and in linear scale (down) and emulation of allodynia and hyperalgesia phenomena of nociceptor; (h) the input voltage curves (Ch1) and the output voltage curves (Ch2) of an artificial thermal nociceptor^[104].

Fig. 10. (Color online) An artificial auditory perception system based on artificial neuron. (a) Schematic of biological auditory system; (b) schematic of an artificial auditory perception system based on silicon biristor neuron; (c) circuit of an artificial auditory perception system for pitch classification; (d) the synaptic currents of the artificial auditory perception system after applying various G3 type music^[111].

Fig. 11. (Color online) Artificial gustatory perception systems based on artificial neurons. (a) Schematic of biological gustatory system; (b) schematic of artificial gustatory system based on a 1T neuron and a gustatory sensor; (c) the principle of the pH sensor response to hydrogen ions; (d) output characteristics of the 1T neuron under different pH levels; (e) spiking characteristics of the fabricated artificial gustatory neuron under different pH levels; (f) schematic of an artificial gustatory perception system composed of hydron/sodion sensitive sensors and 1T neurons; (g) synapse currents measured at output layer A (I_syn,_A) when vinegar is applied to the gustatory system; (h) synapse currents of output layer B (I_syn,B) when brine is applied to the gustatory system^[112].

Fig. 12. (Color online) An artificial olfactory perception system based on artificial neuron. (a) Schematic of biological olfactory system; (b) an artificial olfactory neuron composed of a 1T neuron and a gas sensor based on semiconductor metal oxide; (c) SEM image of the 1T neuron; (d) measurement scheme for the 1T neuron operation; (e) spiking characteristics of the 1T-neuron under 100 and 400 nA stimuli; (f) dynamic responses of the SnO₂ and WO₃ gas sensors to NH₃ gas; (g) spiking characteristics of the artificial olfactory neuron composed of SnO₂ gas sensor and 1T-neuron to NH₃ with various parts per million (ppm); (h) synapse currents collected at the output layer of the artificial olfactory perception system after applying Merlot wine; (i) synapse currents of the artificial olfactory perception system after applying Shiraz wine^[32].

Fig. 13. (Color online) Tactile pressure and temperature multi-modal fusion perception. (a) Schematic of multi-modal perception neurons based on the MFSN array; (b) I−V curve of the NbO_x threshold resistive memristor; (c) schematic circuit diagram of the MFSN; (d) characteristics of the output spike under different pressures; (e) output spike characteristics of the MFSN with a constant input voltage of 4 V at temperatures of 20, 40, and 60 °C^[120].

Fig. 14. (Color online) (a) Schematic of the bimodal artificial sensory neuron with visual-haptic fusion. Sub-figures ⅰ to ⅳ: photodetector, pressure sensor, hydrogel (dyed by 0.04% methylene blue), and synaptic transistor, respectively; (b) schematic of the relative change of ΔEPSC% over time interval (ΔT); (c) schematic of visual-haptic fusion for muscle actuation; (d) different responses of skeletal muscle tube to average velocity at different ΔEPSC% changes^[119]; (e) schematic of a biologically inspired multi-sensory neural network; (f) schematic diagram of visual-tactile fusion perception, which illustrates the human ability to recognize and visualize tactile input; (g) visual−auditory fusion in which the number of PSC pulses is used to simulate how close the car is to the person^[37].