By monitoring brain neural signals, neural recorders allow for the study of neurological mechanisms underlying specific behavioural and cognitive states. However, the large brain volumes of non-human primates and their extensive range of uncontrolled movements and inherent wildness make it difficult to carry out covert and long-term recording and analysis of deep-brain neural signals. Here we report the development and performance of a stealthy neural recorder for the study of naturalistic behaviours in non-human primates. The neural recorder includes a fully implantable wireless and battery-free module for the recording of local field potentials and accelerometry data in real time, a flexible 32-electrode neural probe with a resorbable insertion shuttle, and a repeater coil-based wireless-power-transfer system operating at the body scale. We used the device to record neurobehavioural data for over 1 month in a freely moving monkey and leveraged the recorded data to train an artificial intelligence model for the classification of the animals’ eating behaviours.

通过监测大脑神经信号，神经记录器可以研究特定行为和认知状态背后的神经机制。然而，非人类灵长类动物的大量脑容量及其广泛的不受控制的运动和固有的野性使得难以对深部脑神经信号进行隐蔽和长期的记录和分析。在这里，我们报告了用于研究非人类灵长类动物自然行为的隐蔽神经记录器的开发和性能。神经记录仪包括一个完全植入式无线和无电池模块，用于实时记录局部场电位和加速度测量数据，一个灵活的 32 电极神经探针，带有可吸收插入穿梭车，以及一个在身体尺度上运行的基于中继线圈的无线电力传输系统。我们使用该设备记录了一只自由移动的猴子 1 个多月的神经行为数据，并利用记录的数据训练了一个人工智能模型，用于对动物的饮食行为进行分类。