Emulating neurons/synapses in the brain is an important step to realizing highly efficient computers. This fact makes neuromorphic devices important emerging solutions to the limitations imposed by the current computing architecture. To mimic synaptic functions in the brain, it is critical to replicate ionic movements in the nervous system. It is therefore important to note that ions move easily in liquids. In this study, we demonstrate a liquid-based neuromorphic device that is capable of mimicking the movement of ions in the nervous system by controlling Na⁺ movement in an aqueous solution. The concentration of Na⁺ in the solution can control the ionic conductivity of the device. The device shows short-term and long-term plasticity such as excitatory postsynaptic current, paired-pulse facilitation, potentiation, and depression, which are key properties for memorization and computation in the brain. This device has the potential to overcome the limitations of current von Neumann architecture-based computing systems and substantially advance the technology of neuromorphic computing.

在大脑中模拟神经元/突触是实现高效计算机的重要一步。这一事实使神经形态设备成为解决当前计算体系结构局限性的重要新兴解决方案。为了模仿大脑中的突触功能，复制神经系统中的离子运动至关重要。因此，必须注意离子在液体中容易移动。在这项研究中，我们演示了一种基于液体的神经形态设备，它能够通过控制水溶液中的Na ⁺运动来模仿神经系统中离子的运动。Na ⁺的浓度在溶液中可以控制设备的离子电导率。该设备显示出短期和长期可塑性，例如兴奋性突触后电流，成对脉冲促进，增强和抑制，这是大脑中记忆和计算的关键特性。该设备有可能克服当前基于冯·诺依曼体系结构的计算系统的局限性，并极大地促进神经形态计算技术的发展。