Mougkogiannis Panagiotis, Nikolaidou Anna, Adamatzky Andrew
Unconventional Computing Laboratory, UWE, Bristol, BS16 1QY, U.K.
ACS Omega. 2024 Nov 5;9(46):45789-45810. doi: 10.1021/acsomega.4c03546. eCollection 2024 Nov 19.
The integration of proteinoid-polyaniline (PANI) nanofibers with neuromorphic architectures shows potential for developing computer systems that are adaptable, energy-efficient, and have the capacity of tolerating faults. This work examines the capacity of proteinoid-PANI nanofibers to imitate different spiking patterns in stimulated Izhikevich neurons. The proteinoid-PANI nanofibers exhibit diverse spiking behaviors on different substrates, showcasing a broad range of control and programmability, as confirmed by experimental characterization and computational modeling. K-means clustering technique measures the extent and selectivity of the proteinoid-PANI spiking behavior in response to various stimuli and spiking patterns. The presence of strong positive correlations between membrane potential and time suggests that the system is capable of producing reliable and consistent electrical activity patterns. Proteinoid-PANI samples demonstrate enhanced stability and consistency in numerous spiking modes when compared to simulated input neurons. The results emphasize the capability of proteinoid-PANI nanofibers as a bioinspired substance for neuromorphic computing and open up possibilities for their incorporation into neuromorphic structures and bioinspired computer models.
类蛋白-聚苯胺(PANI)纳米纤维与神经形态架构的整合显示出开发具有适应性、高能效且具备容错能力的计算机系统的潜力。这项工作研究了类蛋白-PANI纳米纤维模仿受刺激的艾克米维奇神经元中不同脉冲模式的能力。通过实验表征和计算建模证实,类蛋白-PANI纳米纤维在不同底物上表现出多样的脉冲行为,展现出广泛的可控性和可编程性。K均值聚类技术测量了类蛋白-PANI脉冲行为对各种刺激和脉冲模式的响应程度及选择性。膜电位与时间之间存在强正相关,这表明该系统能够产生可靠且一致的电活动模式。与模拟输入神经元相比,类蛋白-PANI样本在多种脉冲模式下表现出更高的稳定性和一致性。这些结果强调了类蛋白-PANI纳米纤维作为用于神经形态计算的生物启发物质的能力,并为将其纳入神经形态结构和生物启发计算机模型开辟了可能性。