Sensory coding in supragranular cells of the vibrissal cortex in anesthetized and awake mice

Abstract

Sensory perception entails reliable representation of the external stimuli as impulse activity of individual neurons (i.e. spikes) and neuronal populations in the sensory area. An ongoing challenge in neuroscience is to identify and characterize the features of the stimuli which are relevant to a specific sensory modality and neuronal strategies to effectively and efficiently encode those features. It is widely hypothesized that the neuronal populations employ “sparse coding” strategies to optimize the stimulus representations with a low energetic cost (i.e. low impulse activity). In the past two decades, a wealth of experimental evidence has supported this hypothesis by showing spatiotemporally sparse activity in sensory area. Despite numerous studies, the extent of sparse coding and its underlying mechanisms are not fully understood, especially in primary vibrissal somatosensory cortex (vS1), which is a key model system in sensory neuroscience. Importantly, it is not clear yet whether sparse activation of supragranular vS1 is due to insufficient synaptic input to the majority of the cells or the absence of effective stimulus features.

In this thesis, first we asked how the choice of stimulus could affect the degree of sparseness and/or the overall fraction of the responsive vS1 neurons. We presented whisker deflections spanning a broad range of intensities, including “standard stimuli” and a high-velocity, “sharp” stimulus, which simulated the fast slip events that occur during whisker mediated object palpation. We used whole-cell and cell-attached recording and calcium imaging to characterize the neuronal responses to these stimuli. Consistent with previous literature, whole-cell recording revealed a sparse response to the standard range of velocities: although all recorded cells showed tuning to velocity in their postsynaptic potentials, only a small fraction produced stimulus-evoked spikes. In contrast, the sharp stimulus evoked reliable spiking in a large fraction of regular spiking neurons in the supragranular vS1. Spiking responses to the sharp stimulus were binary and precisely timed, with minimum trial-to-trial variability. Interestingly, we also observed that the sharp stimulus produced a consistent and significant reduction in action potential threshold.

In the second step we asked whether the stimulus dependent sparse and dense activations we found in anesthetized condition would generalize to the awake condition. We employed cell-attached recordings in head-fixed awake mice to explore the degree of sparseness in awake cortex. Although, stimuli delivered by a piezo-electric actuator evoked significant response in a small fraction of regular spiking supragranular neurons (16%-29%), we observed that a majority of neurons (84%) were driven by manual probing of whiskers. Our results demonstrate that despite sparse activity, the majority of neurons in the superficial layers of vS1 contribute to coding by representing a specific feature of the tactile stimulus.

Thesis outline: Chapter 1 provides a review of the current knowledge on sparse coding and an overview of the whisker-sensory pathway. Chapter 2 represents our published results regarding sparse and dense coding in vS1 of anesthetized mice (Ranjbar-Slamloo and Arabzadeh 2017). Chapter 3 represents our pending manuscript with results obtained with piezo and manual stimulation in awake mice. Finally, in Chapter 4 we discuss and conclude our findings in the context of the literature. The appendix provides unpublished results related to Chapter 2. This section is referenced in the final chapter for further discussion.