Neuromorphic computing: Mimicking neural plasticity; protein nanowires; artificial synapses.
Researchers from Korea Advanced Institute of Science and Technology (KAIST) developed a frequency switching neuristor device that mimics the intrinsic plasticity of neurons. The device can autonomously adjust the frequency of its signals, similar to the way the brain becomes less startled by repeated stimuli or becomes increasingly sensitive through training.
The team combined a volatile Mott memristor, which reacts momentarily before returning to its original state, with a non-volatile memristor, which remembers input signals for long periods of time. This resulted in a device in which neuronal spike signals and memristor resistance changes influence each other, automatically adjusting responses to control how often a neuron fires, also referred to as its spiking frequency.
In simulations conducted with a sparse neural network, the system achieved the same performance with 27.7% less energy consumption compared to conventional neural networks. Additionally, intrinsic plasticity allowed the network to reorganize itself and restore performance even if some neurons were damaged.
“This study implemented intrinsic plasticity, a core function of the brain, in a single semiconductor device, thereby advancing the energy efficiency and stability of AI hardware to a new level,” said Kyung Min Kim, a professor in the Department of Materials Science and Engineering at KAIST, in a statement. “This technology, which enables devices to remember their own state and adapt or recover even from damage, can serve as a key component in systems requiring long-term stability, such as edge computing and autonomous driving.” [1]
Researchers at the University of Massachusetts Amherst created an artificial neuron with electrical functions that closely mirror those of biological ones and could be capable of interfacing directly with living cells.
The artificial neuron uses a protein nanowire synthesized from the bacteria Geobacter sulfurreducens, which can also produce electricity and has been used to create a variety of sensor and energy harvesting devices.
“We currently have all kinds of wearable electronic sensing systems, but they are comparatively clunky and inefficient. Every time they sense a signal from our body, they have to electrically amplify it so that a computer can analyze it. That intermediate step of amplification increases both power consumption and the circuit’s complexity, but sensors built with our low-voltage neurons could do without any amplification at all,” said Jun Yao, associate professor of electrical and computer engineering at UMass Amherst, in a release. “Previous versions of artificial neurons used 10 times more voltage—and 100 times more power—than the one we have created. Ours register only 0.1 volts, which about the same as the neurons in our bodies.” [2]
Researchers from Seoul National University of Science and Technology (SeoulTech), Gyeongsang National University, Pohang University of Science and Technology, and Inha University designed organic semiconductors with glycol side chains for neuromorphic computing that can improve artificial synapses by enabling more efficient, bulk-mediated ion transport.
“Our research shows a simple way to make the next wave of AI hardware more efficient by improving electrolyte-based organic transistors, which are soft, low-voltage devices that process signals with ions as well as electrons,” said Eunho Lee, an assistant professor of chemical and biomolecular engineering at SeoulTech, in a statement. “A long-standing bottleneck has been doping efficiency: how effectively ions can enter and leave the polymer channel to switch the device. We addressed this by engineering the polymer’s side chains so they actively attract and guide ions, like molecular ‘handles’ and ‘lanes,’ leading to faster and deeper ion uptake.”
The researchers say the electrolyte-based organic transistors with diffusion-driven doping can function as analog synapses for ultra-low-power co-processors in wearables, cameras, and IoT nodes. The design also supports hybrid integration with CMOS for compact analog memory arrays that reduce data movement and latency. Additionally, it is well-suited to skin and tissue environments, potentially enabling bioelectronic interfaces for closed-loop therapies and electrochemical biosensors. [3]
[1] W. Park, H. Song, E. Y. Kim, et al. Frequency Switching Neuristor for Realizing Intrinsic Plasticity and Enabling Robust Neuromorphic Computing. Adv. Mater. (2025): e02255. https://doi.org/10.1002/adma.202502255
[2] S. Fu, H. Gao, S. Wang, et al. Constructing artificial neurons with functional parameters comprehensively matching biological values. Nat Commun 16, 8599 (2025). https://doi.org/10.1038/s41467-025-63640-7
[3] J. Sung, H. J. Cheon, D. Lee, et al. Improving ion uptake in artificial synapses through facilitated diffusion mechanisms. Mater. Horiz., 2025,12, 5225-5235 https://doi.org/10.1039/D5MH00005J
 |
 |
 |
 |
 |  |
 |  |
 |
Leave a Reply