由加州大学圣巴巴拉分校(UCSB)、苏黎世联邦理工学院(ETH Zurich)及米格尔·埃尔南德斯大学(MHU)组成的联合研究团队,开发出一项基于人工智能的仿生视觉技术,相关成果已发表于《Neuron》期刊 1。该研究利用深度学习系统为临时植入人类视觉皮层的电极开发电刺激模式,旨在预测盲人感知的闪光和形状 1。论文标题为“Deep learning-based control of electrically evoked activity in human visual cortex” 1。
在位于西班牙Elche IMED医院进行的实验中,研究人员对一名27岁因创伤性脑损伤失明的男性患者植入了96个电极 1。与传统方法不同,AI模型不仅考虑刺激设置,还分析了刺激前大脑的静息活动状态,从而更准确地预测神经反应 1。结果显示,相比对比方法,经AI选择的新刺激模式能以更少的电流更准确地产生预期的光幻视感知及神经活动图案 1。这项发表于《Neuron》的研究标志着个性化视觉假体开发的重要进展,有望帮助盲人恢复部分视力 1。
此次计算研究的推进得益于美国国立卫生研究院(NIH)院长创新奖的支持,该奖项于2022年授予了加州大学圣巴巴拉分校副教授Michael Beyeler 1。文章发布日期为2026年11月12日 1。
A collaborative research team from the University of California, Santa Barbara (UCSB), ETH Zurich, and Miguel Hernández University (MHU) has developed an artificial intelligence-based visual prosthetic technology that utilizes deep learning to predict and precisely control electrical stimulation responses in the visual cortex. 1 Published on November 12, 2026, the study marks a significant advancement in the development of personalized visual implants aimed at restoring partial sight to blind individuals. 1
The breakthrough was demonstrated in a successful experiment conducted in Spain with a 27-year-old male patient who lost his vision due to traumatic brain injury. 1 At IMED Hospital in Elche, researchers implanted 96 electrodes into the patient's visual cortex to test the system's ability to generate expected phosphenes—perceptions of flashes and shapes. 1 Unlike traditional methods, the AI model not only considered specific stimulation settings but also analyzed the brain's resting activity state prior to stimulation, allowing for more accurate predictions of neural responses. 1 The results showed that the AI-selected new stimulation patterns could produce expected neural activity with greater precision using less current compared to baseline approaches. 1
The findings have been published in the journal Neuron under the title "Deep learning-based control of electrically evoked activity in human visual cortex." 1 Michael Beyeler, an associate professor at UCSB whose computational research was supported by a 2022 NIH Director's New Innovator Award, is part of the multi-institutional collaboration behind the study. 1
评论
还没有评论,欢迎留下第一条。