脑机接口中文解码实现临床突破 失语患者有望“借机”重获“新声”

Core Insights - The collaboration between Shanghai Yansi Brain-like AI Research Institute, Huashan Hospital, and other institutions has achieved a clinical breakthrough in brain-computer interface (BCI) Chinese decoding, potentially improving treatment for patients with speech impairments due to diseases like ALS and stroke [1][2][3] Group 1: Clinical Breakthrough - The implanted brain electroencephalogram (EEG) model allows patients to convert thoughts into audible speech, marking a significant advancement in BCI technology [2][3] - A patient underwent training for 100 minutes, during which the model was able to output coherent sentences after reading 54 Chinese characters, demonstrating the model's efficiency [2][3] Group 2: Technical Achievements - The model can decode Chinese, which has a higher complexity than English, achieving a decoding accuracy of over 83% for 10 consonants and over 84% for 15 vowels [3] - The model's extrapolation capability is 1:36, allowing for high accuracy and strong decoding performance [3] Group 3: Future Applications - The technology aims to enable patients with speech impairments to communicate naturally, significantly enhancing their quality of life [4][5] - The establishment of the iBRAIN EEG data alliance aims to create a standardized, shared invasive intracranial EEG database, facilitating data sharing and improving BCI technology [5][6] Group 4: Broader Implications - The open sharing platform will support multi-domain research, potentially aiding in understanding brain activity related to motor signals and emotional signals, with implications for various languages [6]