SIGNALING OF SALIENCE AND PREDICTION ERRORS BY THE INSULA

点击翻译按钮获取中文摘要

Overview. Human fMRI and LFP recording studies suggest that the anterior insula encodes the salience of stimuli and deviations from expectations. However, BOLD signals and LFPs have insufficient resolution to assess how these codes are distributed across insular cells. Moreover, the origin and functional relevance of insular codes is unclear. We will address these questions with multi-site single unit and LFP recordings as well as optogenetic methods in rats. Critically, our pilot data indicates that presentation of salient stimuli elicits identical LFP responses in rats and humans: bursts of high amplitude beta (13-30 Hz) oscillations. Significance and approach. Because its volume is reduced in most psychiatric disorders, the insula is a promising therapeutic target for transcranial or deep brain stimulation. However, progress in this field has been hindered by a lack of high-resolution data about insula signaling. Thus, this proposal may facilitate the development of novel therapies. To study the insula in rats, we developed a novel reinforcement learning task that features a behavioral readout of the rats’ expectations (and their violations). This task will allow us to assess how anterior insula neurons and their synaptic partners encode multiple variables and probe their role in learning. Aim #1 Characterize the coding properties of neurons in the anterior insula and its synaptic partners. We will perform multi-site unit and LFP recordings in the anterior insula and its synaptic partners. This will allow us to: (1) test whether cells at these various sites encode outcome valence, magnitude, salience, and deviations from expectations; (2) test whether valenced stimuli elicit bursts of beta oscillations in the anterior insula, as seen in humans; (3) compare the entrainment of neurons at these various sites by oscillations of different frequencies; (4) test whether the amplitude of insular beta bursts tracks stimulus salience and deviations from expectations. Aim #2: Determine which inputs drive coding of different variables in anterior insula neurons. The data obtained in Aim 1 will suggest which inputs convey information about different variables to insula neurons. Similar hypotheses will arise regarding the origin and propagation of insular beta bursts. To test these hypotheses, we will infuse AAVs driving the expression of an inhibitory opsin in projection-defined cell types. Then, we will inhibit the candidate cell types and assess how these manipulations affect the coding properties of insula neurons as well as the genesis and propagation of insular beta bursts. Aim #3 Test whether reducing or enhancing insular beta bursts impairs or facilitates learning. We will achieve real-time control over insular beta bursts by combining optogenetics with programmable multi-channel signal processors, giving us unprecedented control over fast neuronal events. We will enhance or dampen insular beta bursts by delivering low-intensity light stimuli that bias spike times in or out-of-phase with respect to beta oscillations, without altering firing rates. If insular beta bursts facilitate learning, up- or down-regulating them should respectively accelerate or slow down learning in the task.