Neural circuitry represents sensory input with patterns of activity within populations of individual neurons. These representations are transformed across brain areas into representations that drive adaptive behavior. The circuitry across which transformations occur often extend over many millimeters: the mesoscale level or organization. Imaging at the mesoscale with micron resolution, deep in brain tissue that badly scatters light, with subsecond time resolution is a challenging optical problem. Here, we used new multiphoton imaging technology to measure representations across multiple visual cortical areas simultaneously in mice. By measuring shared variability for pairs of neurons within or across cortical areas, we obtained measurements of correlational structure, which can constrain models of circuit connectivity and mechanisms for transforming neural representations. In the course of this work, we also uncovered new principles for the organization of spatiotemporal and orientation tuning among visual cortex neurons. Together, these findings are illuminating how visual stimuli are represented in primary and higher visual areas in mice, and how these areas can be connected to each other.
Neuronal activity in cerebral cortex shows many attention-related changes that might contribute to improved behavioral performance: neurons have stronger responses to attended stimuli and effectively mask unattended stimuli within their receptive field; nearby neurons have activity that is less correlated; and the gamma power in extracellular potentials increases. About ten years ago, it became widely recognized that attention-related changes in neuronal responses are closely linked to a response-integration mechanism known as normalization. However, the relationship between attention and normalization in determining neuronal responses has not been clear. We have examined this relationship by recording responses from individual neurons in visual cortex of trained, behaving rhesus monkeys. We have taken advantage of different stimulus configurations that produce, or do not produce, normalization. The only neuronal correlate of attention that survives the removal of normalization is a modest increase in the strength of neuronal responses. The masking of unattended stimuli, reductions of correlated activity and increase of gamma power all depend having robust normalization. With robust normalization, those same phenomena can be seen when attention is removed from the picture and replaced by simple changes in stimulus strength. Overall, these results suggest that the immediate effect of attention is a modest modulation of sensory responses, and that the more striking signatures that have been previously attributed to attention are better viewed as signatures of normalization mechanisms that lie downstream.
Microglia are dynamic sensors of their extracellular environment and are intimately associated with synapses within neural circuits. Our previous work has demonstrated that microglia sculpt synaptic connectivity in the developing rodent visual system by phagocytosing a subset of less active synapses. Going forward, we are now using another sensory system, the mouse barrel cortex, to dissect how neural activity and sensory experience modulate microglial phagocytic function and plasticity of neural circuits. Ultimately, we are applying these mechanisms to elucidate how dysregulation of synaptic phagocytosis by microglia can play a pivotal role in synapse pathology in neurological disease.
First, the ideas of motor planning and imitation will be linked through the example of apraxia. Second, the vexed issue of motor skill and how it relates to cognition and memory will be discussed. Finally, the notion of recovery from brain injury as a form of skill learning will be examined.
Animals rely on olfaction to find food, attract mates and avoid predators. To support these behaviors, they must be able to reliably identify a given odor over a large range of odorant concentrations, while nevertheless retaining the ability to discriminate small differences in concentration. Through a combination of theoretical and experimental investigations, we identified complementary coding strategies for generating non-interfering representations of odor identity and odor intensity in mouse piriform cortex. We next examined the neural circuit operations that underlie these representations. We find that intrinsic recurrent circuitry is required for concentration-invariant odor recognition and for stabilizing odor representations over time and across variable stimulus conditions. Our results therefore highlight the specific and crucial roles that intrinsic cortical circuitry play in shaping sensory representations.