van den Boom Max A, Vansteensel Mariska J, Koppeschaar Melissa I, Raemaekers Matthijs A H, Ramsey Nick F
Department of Neurology and Neurosurgery, UMC Utrecht Brain Center, University Medical Center Utrecht, Utrecht, The Netherlands.
Biomed Phys Eng Express. 2019 Aug 2;5(5). doi: 10.1088/2057-1976/ab302c. eCollection 2019 Aug.
Brain-computer interfaces aim to provide people with paralysis with the possibility to use their neural signals to control devices. For communication, most BCIs are based on the selection of letters from a (digital) letter board to spell words and sentences. Visual mental imagery of letters could offer a new, fast and intuitive way to spell in a BCI-communication solution. Here we provide a proof of concept for the decoding of visually imagined characters from the early visual cortex using 7 Tesla functional MRI. Sixteen healthy participants visually imagined three different characters for 3, 5 and 7 s in a slow event-related design. Using single-trial classification, we were able to decode the characters with an average accuracy of 54%, which is significantly above chance level (33%). Furthermore, the imagined characters were classifiable shortly after cue onset and remained classifiable with prolonged imagery. These properties, combined with the cortical location of the early visual cortex and its decodable activity, encourage further research on intracranial interfacing using surface electrodes to bring us closer to such a visual imagery based BCI communication solution.
脑机接口旨在为瘫痪患者提供利用其神经信号控制设备的可能性。对于通信而言，大多数脑机接口基于从（数字）字母板中选择字母来拼写单词和句子。字母的视觉心理意象可为脑机接口通信解决方案提供一种新的、快速且直观的拼写方式。在此，我们通过7特斯拉功能磁共振成像提供了从早期视觉皮层解码视觉想象字符的概念验证。16名健康参与者在缓慢的事件相关设计中对三个不同字符进行了3秒、5秒和7秒的视觉想象。使用单次试验分类，我们能够以54%的平均准确率解码字符，这显著高于机遇水平（33%）。此外，想象的字符在提示开始后不久即可分类，并且随着想象时间延长仍可分类。这些特性，再加上早期视觉皮层的皮质位置及其可解码活动，鼓励进一步开展使用表面电极进行颅内接口的研究，以使我们更接近这种基于视觉意象的脑机接口通信解决方案。
