Neurons receive information in the form of temporal patterns of neuronal inputs and decode them through a complex and hierarchical network of biochemical signal transduction systems. Each molecular component in such a network processes upstream inputs and regulates downstream effectors. These elementary steps have been proposed to constitute the simplest forms of computation, information processing, and even cellular decision making such as induction of synaptic plasticity that has been considered as cellular correlates of learning behavior. However, how information encoded in firing rates can be converted into biochemical activation patterns of distinct postsynaptic enzymes remains unknown. To address this issue, we developed a dual FRET imaging platform and recorded CaMKIIalpha and calcineurin activities in individual spines of hippocampal neurons, during high- or low-frequency glutamate uncaging. High-frequency stimuli (20 Hz) that induced spine growth activated both CaMKII and calcineurin, with comparable population kinetics. In contrast, low-frequency stimuli (5 Hz) strongly stimulated calcineurin but not CaMKII, with little spine morphological change. Higher temporal resolution imaging in the neuronal soma revealed that CaMKII activation was summed supralinearly and required both higher frequency and input number, thus acting as a dual high-pass filter. In contrast, calcineurin activity summated sublinearly with increasing input number and showed little frequency-dependence, thus functioning as an input number counter. Taken together, CaMKII and calcineurin are fine-tuned to unique bandwidths and process input variables in an asymmetric, rather than opposing, manner.
Visual spatial attention improves perceptual performance at locations for which prior information indicates a higher likelihood of relevance. What are the neural mechanisms that allow visual attention to change visual information processing to achieve better perceptual performance? How can we use experimental observations from psychophysical experiments, measurements of brain activity and computational models to test hypotheses about neural mechanisms that underly improved behavioral performance with spatial attention? We will discuss and review basic findings from behavioral measurements of spatial attentions effects on perception and the concomitant changes in neural responses that have been measured in experimental animals and humans. We will discuss major theoretical hypotheses of neural mechanisms that have been used to explain behavioral enhancement with attention such as sensory enhancement, noise exclusion, uncertainty reduction and efficient selection of appropriate sensory signals. Using behavioral and neural measurements we will discuss how different theoretical hypotheses of neural mechanisms for spatial attention can be disambiguated.
Abstract: A wide range of organisms follow chemical cues to locate food sources. Behavioral strategies employed for chemotaxis have been studied across phyla, but the neural computations underlying odor-guided navigation remain poorly understood. We use a combination of electrophysiology, optogenetics, quantitative behavior and computational modeling to understand how dynamical olfactory signals experienced during chemotaxis are processed by the olfactory system of the Drosophila larva and how this information is converted into navigational decisions. Our work aims to clarify the link between neural computations at the sensory periphery and decision-making processes that direct navigation in a sensory landscape.
Conscious sensory perception is an active process involving direct communication between cortical areas conveying bottom-up and top-down information. However, the precise nature of the information transmitted between brain areas and their influence is poorly understood. Here, we used a multidisciplinary approach to understand the top-down influence of the secondary motor area (M2) on the primary sensorimotor area (S1) in awake and anesthetized mice. We show that axons from M2 project predominantly to layers 1, 5 & 6 of S1 and that neuronal activity in M2 (but not M1) is elevated in the awake state relative to anesthesia. We found that M2 had a profound influence on the late components of activity in S1 across all layers but in particular on dendritic activity in layer 5 (L5) pyramidal neurons. The results indicate a reverberating circuit between S1 and M2 during conscious sensory processing involving a vital feedback influence by M2.
Advances in neural recording are leading to increasingly large, complex neural datasets, and statistical methods for describing and drawing scientific conclusions from these data are essential to understanding how information is represented and processed in the brain. Over the past few years model-based statistical inference techniques have been developed that allow us to infer functional connections between neurons based on spike recordings. These techniques improve the accuracy of spike prediction and promise to provide insight into the structure and function of neural circuits. In this talk I will describe several recent advances in inferring functional connectivity from spikes: how prediction accuracy increases with the number of connections modeled and how plasticity in these connections can be taken into account. Finally, I will discuss new results from intracellular current injection experiments where functional connections can be simulated in a biophysically realistic setting. These experiments illustrate the possibilities and limitations of functional connectivity methods in fully observed networks of neurons.
TBA
