Action effect anticipation in the frontoparietal sensorimotor network

为了达到或操纵物体而产生精确、灵活和情境敏感的目标定向手臂运动的能力是人类和非人类灵长类动物的标志。对一个动作的感官结果的预测——可能是后来的奖励信号，可能是一旦计划的运动目标实现后预期的身体结构，或者是由正在进行的运动命令产生的即时感官反馈——对于目标导向行动的计划和执行来说是至关重要的。灵长类动物前额-顶叶网络的大脑皮层区域，特别是背侧运动前皮层和顶叶到达区，支持前臂运动计划的情境依赖的感觉运动转换，并且是行动计划过程中假定的感觉预期信号的皮质计算和编码的强有力候选区域。在第一个资助期的主要项目中，我们用恒河猴在执行伸手任务时的两个区域的神经记录来测试这一假设，在这个任务中，视觉反馈和实际的身体运动由于视觉输入的反向而分离。我们的研究结果为我们的假设提供了支持，因为观察到的神经编码可以形成视觉动作效果预期的基础。为了提供更直接的证据并排除混淆的解释，我们现在想利用脑机接口（BCI）范式来更好地控制相关的行为和感觉反馈参数。为此，我们开发了一个具有实时行为控制和神经解码的三维虚拟现实设置。在一只动物体内植入3个感觉运动区192个通道的大规模慢性单神经元记录的初步数据证实了这种方法的可行性。如果成功的话，这个项目的结果有可能澄清心理学和认知神经科学中一个长期存在的问题，即灵长类动物在认知处理水平上的运动计划的本质。

The capability to generate precise, flexible, and context-sensitive goal-directed arm movements in order to reach or manipulate objects is a hallmark of humans and non-human primates. Predictions of the sensory consequences of an action - be it later reward signals, be it the anticipated body configuration once a planned motor goal will be achieved, or be it the immediate sensory feedback resulting from an ongoing motor command - appear central for the planning and execution of goal-directed actions. Cerebral cortical areas of the fronto-parietal network in primates, particularly dorsal premotor cortex and parietal reach region, support context-dependent sensorimotor transformations for planning forearm movements, and are strong candidate areas for a cortical computation and encoding of putative sensory anticipation signals during action planning. In the main project of the first funding period, we tested this hypothesis with neural recordings from both areas in rhesus monkeys while they performed a reach task where visual feedback and actual physical movement were dissociated due to reversed visual input. Our results provided support for our hypothesis since the observed neural encoding could form the basis of visual action effect anticipation. To provide more direct evidence and rule out confounding interpretations, we now want to utilize a brain-computer interface (BCI) paradigm to better control the relevant behavioural and sensory feedback parameters. For this we developed a 3-D virtual reality setting with real-time behavioural control and neural decoding. Preliminary data from one animal implanted with large-scale chronic single neuron recordings from 192 channels in 3 sensorimotor areas confirm the feasibility of this approach. If successful, the results from the proposed project have the potential to clarify a longstanding question in psychology and cognitive neuroscience about the nature of motor planning at the level of cognitive processing in primate.