An Organic Platform with Neurotransmitter Closed-Loop Control for Actuation and Reinforcement Learning Tasks

In the brain information is transferred between neurons at synapses, which are sites where chemical interactions between neurotransmitters and receptors generate an electrical potential. Furthermore, neurotransmitter availability can alter the electrochemical information transfer, influencing neuronal plasticity, which is the foundation of the brain's capacity for learning. Nowadays, brain-computer interfaces (BCIs) can be used to treat neuronal communication impairment, but bidirectional device-nervous system connectivity remains a barrier.<br/><br/>The most promising method for this is to exploit the parallel computing paradigm, which is inspired by the brain architecture. Organic electrochemical transistors (OECTs) based on PEDOT:PSS are good candidates for BCIs, because of their ionic-to-electronic signal transduction and biocompatibility^1,2.<br/><br/>In this work, we implemented an integrated organic platform that can work in tandem with mainstream silicon technologies to realize brain-inspired computing. After conditioning due to neurotransmitter oxidation³, hydrogen peroxide allowed for the partial recovery of the PEDOT:PSS doping level. The artificial "synaptic weights", which were based on the PEDOT:PSS conductance, were adjusted to alter the electrical dynamics of the circuit, demonstrating the computational power of the closed-loop system implemented. Additionally, the same neurotransmitter-mediated OECT was used to connect the motors of a robotic hand to regulate the hand's opening and closing. This platform was included reinforcement learning for robotic hand gripping of balls of varying diameters. When the hand was not in contact with the ball, the punishment was expressed by injecting additional neurotransmitter; when the hand established a firm grasp, the reward was maintaining the motors in standby. The robotic hand had a sensor pressure applied to it, and as soon as the ball contacts its fingers, an Arduino board provided electrical feedback⁴.<br/><br/>The same system could be adopted for other computational tasks, such as the stimulation of a biological neural network, in response to a specific range of neurotransmitters concentration and specific electrical activities.<br/><br/>By implementing local adaptive computing in brain computer interfaces, this closed-loop technology may one day be utilized to reestablish synaptic transmission.<br/><br/><br/>References<br/>1. Gkoupidenis, P. et al. Organic mixed conductors for bioinspired electronics. Nat Rev Mater 1–16 (2023).<br/>2. Bernards, D. A. & Malliaras, G. G. Steady-State and Transient Behavior of Organic Electrochemical Transistors. Adv Funct Materials 17, 3538–3544 (2007).<br/>3. Keene, S. T. et al. A biohybrid synapse with neurotransmitter-mediated plasticity. Nat. Mater. 19, 969–973 (2020).<br/>4. Bruno, U. et al. An organic brain-inspired platform with neurotransmitter closed-loop control, actuation and reinforcement learning. Mater. Horiz. 11, 2865–2874 (2024).