Layer, cell-type and pathway-specific thalamocortical input to mouse somatosensory cortex

In the mouse whisker system, sensory information is relayed to the whisker somatosensory cortex by two major thalamic nuclei, the ventral posterior medial nucleus (VPM) and the posterior medial nucleus (POM). While the cortical axonal innervation pattern of these two nuclei has been studied anatomically in some detail, their synaptic input to distinct cell-types across different layers in barrel cortex is incompletely understood. I used the specificity of optogenetics to selectively stimulate axons from VPM or POM, and I measured the evoked excitatory postsynaptic potentials in vitro with whole-cell patch-clamp recordings in primary whisker somatosensory cortex (wS1). VPM or POM was infected in vivo with an adenoassociated virus (AAV) encoding the light-gated cation channel channelrhodopsin (ChR2). Synaptic input onto individual neurons of the barrel cortex was recorded in brain slices in vitro by activating the ChR2-expressing thalamic axons with blue light. I measured thalamic inputs onto excitatory and three distinct classes of GABAergic neurons, expressing Parvalbumin (PV), somatostatin (SST) or vasoactive intestinal peptide (VIP) neurons across all layers of the barrel cortex. In excitatory and PV neurons, I found that the biggest inputs appeared to largely colocalize with the anatomical innervation pattern Anatomically, VPM preferentially innervates layer4 (L4), deep L3 and the L5B/6A border, and, functionally, we found that the biggest input was observed in L4, followed by L3. Anatomically, POM innervates L5A and L1, and, functionally, I found the biggest input in L5A. SST neurons received very weak input from both thalamic nuclei. VIP neurons on the other hand received larger inputs than SST neurons, however, they were weaker than excitatory and PV neurons. POM is considered to have both first-order and higher-order properties and I, therefore, began to investigate connectivity within sub-parts of POM. Taking advantage of a recently developed anterograde transsynaptic AAV injected into somatosensory brainstem nuclei, I defined first-order and higher-order sub-nuclei of POM and investigated their functional connectivity with somatosensory cortex. Anatomically, the first-order POM preferentially innervates L4 of the secondary somatosensory cortex (wS2) and functionally, I found that neurons across many layers in wS2 received direct synaptic input from first-order POM. Higher-order POM innervates both wS2 and wS1, and I found that neurons in wS2 also receive direct synaptic input. The data suggest that first-order POM relays sensory information to wS2, apparently in a parallel signaling pathway to the classical VPM to wS1 sensory pathway. In contrast, higher order POM does not appear to receive direct sensory input from the brainstem Our results begin to provide a more complete understanding of the distribution of thalamic input to specific cell-types across the layers of the mouse somatosensory cortex

In vivo measurement of excitatory synaptic transmission between identified neurons in layer 2/3 mouse barrel cortex

Aurélie Pala

The neocortex is the most distinctive feature of the mammalian brain and it is considered to be the substrate of high-order cognitive functions. The nature and the arrangement of the diverse neuronal elements constituting its multiple areas have received longstanding attention. Progressively such anatomical and functional investigations are undertaken in the context of an intact living being. It offers the possibility of examining the role of particular areas, networks or even individual neurons during various behavioral states. Synaptic connectivity and synaptic transmission have been traditionally investigated in reduced preparations. Typically, electrophysiological and optical techniques have been used to control and record the propagation of electrical activity between two or more neurons in acute brain slices in vitro. The purpose of this thesis is the investigation of synaptic connectivity and synaptic transmission within the intact neocortex of the living mouse. Here I took advantage of the recent development of optogenetics, in combination with electrophysiology and two-photon microscopy to systematically and directly record synaptic transmission between a single excitatory neuron and two main types of GABAergic neurons in layer 2/3 of the mouse barrel cortex in vivo. Overall, I discovered stronger excitatory connections onto GABAergic neurons than onto excitatory neurons, irrespective of the absolute or relative locations of the pre- and postsynaptic neurons somas. I further revealed that parvalbumin-expressing (PV) and somatostatin-expressing (Sst) GABAergic neurons received excitatory inputs that were similar in magnitude, but were more reliable and faster in PV neurons than in Sst neurons. Exploring postsynaptic responses to multiple presynaptic action potentials elicited at high frequency, I found a strong short-term facilitation accompanied by significant input summation in Sst neurons, but little short-term dynamics with no summation in PV neurons. Lastly, I compared the amplitude of single action potential-evoked postsynaptic responses as a function of neocortical activity level and found that it was unchanged in both neuron types. Overall, the results of this thesis provide new insights into the functioning of microcircuits in vivo while confirming many findings from reduced preparations. In the future, it will be interesting to extend these initial in vivo measurements to other neuron and synapse types, particularly in awake animals engaged in different behavioral states.

EPFL2014

Anatomical and functional characterization of neocortical circuits involved in transforming whisker sensory processing into goal-directed licking

Georgios Foustoukos

The choice of an action upon perception of an external stimulus, arriving at a sensory organ of an animal, depends on previous experiences and outcomes throughout its life. In the rodent brain, the underlying mechanisms involved in simple sensorimotor transformations, such as the detection of a whisker stimulus through goal-directed licking, still remain largely unknown. In this thesis, using as a model the mouse somatosensory system, I explored the anatomical and functional properties of neuronal circuits at different stages of this cortical processing. To start with, using state-of-the-art viral tracing techniques, I investigated the thalamocortical circuits relaying sensory signals to the primary and secondary whisker somatosensory cortices (wS1, wS2). Challenging the "classical" views, the results indicated two streams of information carrying whisker-selective tactile signals. The principal trigeminal nucleus (Pr5) innervates the ventral posterior medial nucleus of the thalamus (VPM) and finally reaching layer 4 of wS1 while the spinal trigeminal nucleus (Sp5) through the rostral part of the posterior medial (POm) thalamus drives the layer 4 of wS2. Finally, a caudal part of the POm, which does not receive brainstem input, innervates layer 1 and layer 5A. Apart from their anatomical differences, those pathways conveyed distinct whisker sensory signals during goal-directed behaviors. Afterwards, I studied the cortical control of jaw and tongue movements during licking for rewards, using multisensory and multimotor whisker detection tasks. The data revealed a frontal tongue-jaw primary motor area (tjM1) which is necessary and encodes for directional licking, independently of the sensory stimulus type, shedding light on how the neocortex orchestrates the main motor output of the animal. Subsequently, I focused on changes in the L2/3 neuronal networks of wS1 after learning of a whisker stimulus. Using as a benchmark a novel "fast" learning and reward-dependent whisker detection task, I carried out inactivations of wS1 during different stages of learning and chronic two-photon (2P) calcium imaging in the L2/3 of the C2 barrel column. The inactivation results indicated that wS1 is indispensable for the acquisition of the novel stimulus and the execution of the task at expert levels. Moreover, the neural data suggested a learning-induced and "long-lasting" enhancement in the whisker sensory responses even when animals were unmotivated to lick. At a network level, a re-organization of the neuronal circuits was observed at different timescales with some of the alterations accompanying the rapid changes in the animal behavior. Additionally, the changes in the whisker sensory responses of neurons in wS1, after learning, were projection-pathway specific with wS2-projecting neurons showing higher whisker responses than whisker primary motor cortex (wM1)-projecting ones. In the final part, acknowledging the importance of a better characterization of the cortical-cortical communication of wS1, I described recent technical advancements in neuronal reconstructions. In vivo single-cell electroporation combined with 2P tomography and registration to a digital atlas, demonstrated the diversity of the projection targets of neurons in the L2/3 of wS1. Overall, I presented different results which contribute to a pre-existing body of research and help to decipher fundamentals and yet highly complex neural computations of the mammalian brain.

EPFL2021

Optogenetic Probing of Cortical Synaptic Circuits Measured Through in Vivo Whole-Cell Recordings in the Mouse Barrel Cortex

Céline Matéo

Although neocortex underlies higher-order brain functions, little is known about the synaptic interactions that drive neocortical microcircuit function in vivo. The neocortex is spontaneously active in vivo so I explored how such spontaneous network activity affects in vivo the interactions between excitatory and inhibitory neurons in the barrel cortex. To this aim, I have focused on the layer 2/3 microcircuit of the C2 barrel column driving sensory processing for the C2 whisker, and combined in vivo whole-cell patch-clamp recordings in anesthetised and awake head-restrained mice with optogenetic stimulation, that allows controlling the activity of genetically defined neurons with light. I expressed channelrhodopsin-2 (ChR2), a light-gated cation channel to evoke action potentials in excitatory neurons with brief light pulses. The impact of this optogenetic stimulus was measured on the membrane potential dynamics of their postsynaptic target neurons. The optogenetic stimulus drove reliable action potentials in ChR2-expressing excitatory neurons, independent of the state of ongoing spontaneous cortical activity. In contrast, I found a marked modulation of light-evoked postsynaptic responses in ChR2-non-expressing excitatory neurons by cortical activity. During spontaneously hyperpolarised periods (DOWN state), optogenetic stimulation resulted in depolarisations of postsynaptic excitatory neurons, while during depolarised periods (UP state), the response was significantly smaller and shorter. The optogenetic stimulus drove the membrane potential of postsynaptic excitatory layer 2/3 neurons towards reversal potentials, which were hyperpolarised relative to action potential threshold, leading to a reduction in firing rates. Targeting whole-cell recordings to GFP-labelled GABAergic layer 2/3 neurons with two-photon microscopy revealed that both fast-spiking and non-fast-spiking inhibitory postsynaptic neurons were rapidly driven to fire action potentials, preferentially from UP states. Such rapid activation of inhibitory neurons in the UP state is likely to be responsible for decreasing firing rate in excitatory neurons. As a result, optogenetic stimulation of a given set of excitatory neurons results in a state-dependent recruitment of inhibitory neurons that limit the spread of excitation within the layer 2/3 microcircuit. The data presented here indicates that a powerful inhibitory loop within the cortical microcircuit rapidly balances excitation with inhibition and enforces sparse coding of layer 2/3 excitatory neurons.

EPFL2011