CRCNS : Thalamocortical vs. recurrent connectivity in active sensation of vibrissa touch

University Of California, San Diego, La Jolla CA

Investigators

Abstract

Animals efficiently collect sensory information from the environment and adapt their behavior accordingly. This process predominantly involves active movement of the body and sensory appendages to sample space through purposeful actions. Mammals explore their olfactory environment through intermittent sniffs, they see using smooth pursuit of the eyes to track moving objects, they touch with digits, and, for rodents, through the rhythmic sweep of their vibrissae. Here we propose experiments, models, and theory to address how object position is robustly computed from vibrissa touch. Whisking is an ideal system to study the coding, circuitry, and dynamics of active sensing. The underlying circuit for rhythmic motor control of the vibrissae is part of the greater orofacial sensory system that is coordinated by the oscillator for breathing. The whisking rhythm per se is sufficiently precise so that under behaviorally relevant conditions the representations of whisking and touch are well described in terms of phase in the whisk cycle, as opposed to by kinetic parameters. We focus on a computation in the input layer, L4, to cortex. Our data is at the level of fields of single spines and boutons as well as single neurons, all in behaving mice. Preliminary results show that a major fraction of thalamocortical (TC) boutons encode phase in the whisk cycle so that the TC input to cortex has the geometry of a torus, with phase and neuron identity as the dimensions. L4 excitatory neurons inherit a phase preference for touch in the whisk cycle through the mapping of this input, across a cortical column. The wiring pattern of these projections and its degree of order will constrain the computation performed at the TC stage. Aim 1 involves the construction of a theoretical model that is motivated by past and preliminary data. The model accounts for phase sensitivity in neurons that are subthreshold but poised to invariantly represent a touch input. The model network has a ring architecture, with a thalamic excitatory rhythmic drive balanced by local inhibition, extending continuous attractor models into the realm active sensing. Aim 2 focuses on measurements of the phase tuning of the thalamocortical input, of the activity of L4 inhibitory neurons, and of the predominantly subthreshold response of L4 excitatory neurons during rhythmic whisking. This Aim establishes the balance between an excitatory TC input and the recurrent inhibition. Aim 3 focuses on the transformation of touch into a ‘bump’ of activity, in the space of phase in the whisk cycle, by L4 excitatory neurons. Here we establish the nature of intracellular dynamics and excitatory synaptic architecture and dynamics. We investigate and characterize invariant properties of the bump in L4 activity. Aim 4 completes our theory with the addition of excitatory synapses that are transiently active during the bump. We conjecture that the local excitation is asymmetric in the phase dimension and operates with time-lags; our preliminary data supports this conjecture. The local excitation leads to a bump that exists only for a fraction of the whisking cycle but is invariant in amplitude and lifetime to signify touch. Our project addresses three neuronal computations essential in active sensing. One is the inherence of order in maps of sensory features between two layers, here phase from thalamus to cortex. A second is the interplay between excitatory inputs and inhibitory interactions to maintain a network at the threshold of activation. The third, and major effort, concerns transient recurrent activity to stabilize a touch response tied to rhythmic self-motion. Our work thus provides a solution to a basic problem in active sensing, i.e., the separation of ex-afferent from self-motion sensory signals. Our study will provide new fundamental knowledge that highlights the role of subcortical mechanisms in the neural computations of sensory invariance. Dysfunction in these computations can lead to sensory "overload", in which individuals no longer reject irrelevant details of the sensory stream. This is reminiscent of conditions that induce meltdowns and shutdown in autism spectrum disorders. An outcome of our research will thus be a new model experimental system to potentially address the root cause of this dysfunction. As such, the proposed fundamental research will contribute to reducing the burden of a major neurological disease.