fig4

Neurochemical memory diversification towards liquid reinforcement learning

Figure 4. (A) Schematic of a rodent navigation task; (B) Training process of the navigation; (C) Device responses under 50 continuous light pulses after the treatment of H2S (red curve) and H2O2 (blue curve) (2 s, 15.17 mW·cm-2 light pulses); (D) Exploration result with increased training epochs with the assistance of a reward; (E) Synaptic weights matrix after training with the assistance of a reward; (F) Exploration result with increased training epochs without the assistance of a reward; (G) Synaptic weights matrix after training without the assistance of a reward. H2S: Hydrogen sulfide; H2O2: hydrogen peroxide; ΔG: ΔConductance.