Layer 6 ensembles can selectively regulate the behavioral impact and layer-specific representation of sensory deviants

Abstract

Predictive models can enhance the salience of unanticipated input, and the neocortical laminar architecture is believed to be central to this computation. Here, we examined the role of a key potential node in model formation, layer (L) 6, using behavioral, electrophysiological and imaging methods in mouse somatosensory cortex. To test the contribution of L6, we applied weak optogenetic drive that changed which L6 neurons were sensory-responsive, without affecting overall firing rates in L6 or L2/3. This stimulation suppressed L2/3 deviance encoding, but maintained other stimulus encoding. The stimulation also selectively suppressed behavioral sensitivity to deviant stimuli without impacting baseline performance. In contrast, stronger L6 drive inhibited firing and suppressed overall sensory function. These findings indicate that, despite their sparse activity, specific ensembles of stimulus-driven L6 neurons are required to form neocortical predictions, and for their behavioral benefit. 45 50 55 60 65 70 75 2-Photon Imaging. A two-photon microscope (Bruker/Prairie Technologies) using an 8 kHz resonant galvanometer (CRS) for fast x-axis scanning, and a non-resonant galvanometer (Cambridge 6215) for y-axis increments was used. In Voigts, Deister and Moore Page 16/23 590 595 600 605 610 615 620 625 630 635 Code Availability. All custom software used in this study is freely available: Behavioral experiments were controlled using a custom state machine (www.github.com/open-ephys/behavioral_state_machine) written in Matlab via PCI DIO boards (National Instruments). Vibrissae were tracked using an automated tracker (Extended Data Figure 2, www.github.com/jvoigts/whisker_tracking). Spike sorting was performed using a custom manual sorting tool (www.github.com/open-ephys/simpleclust). Voigts, Deister and Moore Page 17/23 645 650 655 660 665 670 675 680 685 690

Auditory Corticothalamic Neurons Are Recruited by Motor Preparatory Inputs

“…Because GCaMP expression is limited to upper L6 in Ntsr1-Cre 3 Ai148 mice and because these CT cells have sparse, slender, and short apical dendritic and axonal fields ( Figure 1A), there is very little out-of-plane fluorescence, permitting two-photon calcium imaging at depths of $0.7 mm without using high excitation laser power ( Figure 4C). 5,22,34,35 In agreement with our phototagged recordings, we found that L6 CT neurons were strongly recruited by orofacial movements, where the activity rates of 74% of cells was significantly changed from baseline, mostly reflecting enhanced activity beginning prior to lick spout contact (n = 739, N = 3; Figure 4D).…”

“…The advent of optogenetic approaches to activate and silence layer 6 corticothalamic (L6 CT) neurons in Ntsr1-Cre transgenic mice reinvigorated research on corticofugal circuits, inspiring new hypotheses about their role in sensory gain control and predictive coding. [2][3][4][5][6][7] Optogenetic stimulation or inactivation of Ntsr1-Cre+ neurons has demonstrated that L6 CTs are not subtle modulators of downstream activity, but instead throttle the excitability of thalamic and cortical circuits through their direct connections onto combinations of excitatory and inhibitory cell types in each brain region. [2][3][4][8][9][10][11][12][13] Conventional optogenetic stimulation induces artificial patterns of spatiotemporal activity that do not resemble natural spiking, but these experiments can identify testable hypotheses about the type of firing patterns that might be naturally employed during particular behavioral states.…”

Auditory Corticothalamic Neurons are Recruited by Motor Preparatory Inputs

“…Because they are spatially intermingled with other L6 cell types, and because of the strong reciprocal connectivity of the thalamus and cortex, traditional methods for in vivo neural recordings, stimulation or inactivation have made it challenging to identify a specific role for L6 CTs in sensory processing. The advent of optogenetic approaches to activate and silence genetically targeted L6 CT neurons in Ntsr1-Cre transgenic mice has reinvigorated research on CT circuits, inspiring new hypotheses about their role in sensory gain control and predictive coding (Crandall et al, 2015;Gong et al, 2007;Guo et al, 2017;Olsen et al, 2012;Vélez-Fort et al, 2014;Voigts et al, 2019;Williamson and Polley, 2019). However, any role for L6 CTs in active sensing has been purely speculative, as targeted recordings from L6 CTs have never been made in behaving animals.…”

“…3B-C (Fig. 1A), there is very little out of plane fluorescence, permitting 2-photon calcium imaging at depths of ~0.7mm without using high excitation laser power (Daigle et al, 2018;Liang et al, 2019;Voigts et al, 2019) (Fig. 4C).…”

“…For 2-photon imaging of L2/3 PyrNs, imaging was performed 175-225 µm below the pial surface. The focal plane for L6 imaging was 600-700 µm below the pial surface, which can be accomplished with relatively low excitation power (107-138 mW) because Ntsr1-Cre neurons have sparse apical dendritic fields that produce minimal out of plane excitation (Daigle et al, 2018;Liang et al, 2019;Voigts et al, 2019). The DeepSee precompensation oscillator (Spectra Physics) was adjusted for each imaging session to improve image quality and reduce laser power for L6 imaging.…”