Traditional brain–machine interfaces decode cortical motor commands to control external devices. These commands are the product of higher-level cognitive processes, occurring across a network of brain areas, that integrate sensory information, plan upcoming motor actions, and monitor ongoing movements. We review cognitive signals recently discovered in the human posterior parietal cortex during neuroprosthetic clinical trials. These signals are consistent with small regions of cortex having a diverse role in cognitive aspects of movement control and body monitoring, including sensorimotor integration, planning, trajectory representation, somatosensation, action semantics, learning, and decision making. These variables are encoded within the same population of cells using structured representations that bind related sensory and motor variables, an architecture termed partially mixed selectivity. Diverse cognitive signals provide complementary information to traditional motor commands to enable more natural and intuitive control of external devices.

中文翻译：

---

通过脑机接口探索认知


传统的脑机接口解码皮质运动命令以控制外部设备。这些命令是高级认知过程的产物，发生在大脑区域网络中，整合感官信息，计划即将到来的运动动作，并监控正在进行的运动。我们回顾了最近在神经假体临床试验期间在人类后顶叶皮质中发现的认知信号。这些信号与皮层的小区域一致，在运动控制和身体监测的认知方面发挥着不同的作用，包括感觉运动整合、规划、轨迹表示、体感、动作语义、学习和决策。这些变量在同一细胞群中使用结合相关感觉和运动变量的结构化表示进行编码，这种架构称为部分混合选择性。多样化的认知信号为传统运动命令提供补充信息，从而实现对外部设备更自然、直观的控制。