Layer-specific manipulation to test feedforward/feedback cortical circuitry

大脑皮层必须过滤掉大量的感觉输入。过滤不足或过多可能导致 自闭症和精神分裂症等多种疾病。越来越多的证据表明这取决于 α/β（~8-30 Hz）振荡与γ（>35 Hz）相关尖峰。他们无处不在， 皮质和反相关。在感觉输入期间，伽马/尖峰高，而阿尔法/贝塔高（以及 尖峰低）。中心思想是alpha/beta（~8-30 Hz） 深层皮层的节律抑制了表层的活动（尖峰和伽马，>35 Hz）， 前向感官输入。我们的测试平台将是一个预测编码模型，其中alpha/beta携带 来自高级皮层的自上而下的预测，抑制了自下而上的感觉输入的前馈， 预测刺激。我们将通过直接的因果实验来检验这一假设， 猴子的α/β预测节律。 我们通过开发一种新的超快延迟（<10 ms）闭环系统， 大脑的内源性节律和相位匹配的电刺激。我们还将雇用 高密度“层状”电极记录所有皮质层和浅表目标刺激 vs深层。具有注射端口的层状电极也将允许选择性药理学作用。 操作深层与浅层皮质层。猴子将执行局部和全局（模式化） 古怪的检测任务我们的模型对衰减与放大 α/β预测信号将影响局部生理以及前馈与反馈信号 预测和古怪的大脑皮层。了解神经生理学特性， 皮质层之间的功能和相互作用及其不同的动力学可以提供关键的见解 大脑皮层处理的控制以及由其功能障碍引起的紊乱。

The cortex must filter out the bulk of sensory inputs. Not enough or too much filtering may underlie a variety of disorders like autism and schizophrenia. Mounting evidence suggests that this depends on alpha/beta (~8-30 Hz) oscillations vs gamma (>35 Hz) with associated spiking. They are ubiquitous in cortex and anti-correlated. Gamma/spiking is high during sensory inputs while alpha/beta is high (and spiking is low) during conditions requiring top-down control. The central idea is alpha/beta (~8-30 Hz) rhythms in deep cortical layers inhibit the superficial layer activity (spiking and gamma, >35 Hz) that feed forward sensory inputs. Our testbed will be a model of Predictive Coding in which alpha/beta carries the top-down predictions from higher cortex that inhibit the feeding forward of bottom-up sensory inputs of predicted stimuli. We will test this hypothesis via direct cause-and-effect experiments, by manipulating the alpha/beta prediction rhythms in monkeys. We made this possible by developing a new ultra-fast latency (<10 ms) closed-loop system that can read the brain’s endogenous rhythms and phase-match electrical stimulation to them. We will also employ high-density “laminar” electrodes to record from all cortical layers and target stimulation to superficial vs deep layers. Laminar electrodes with injection ports will also allow selective pharmacological manipulation of deep vs superficial cortical layers. Monkeys will perform a local and global (patterned) oddball detection task. Our model makes specific predictions on how attenuating vs amplifying alpha/beta prediction signals will affect local physiological as well as feedforward vs feedback signaling of predictions and oddballs to higher vs lower cortex. Understanding neurophysiological properties, functions and interactions between cortical layers and their different dynamics can provide key insight into the control of cortical processing and the disorders that come from its dysfunction.