The stimulus selectivity and connectivity of layer six principal cells reveals cortical microcircuits underlying visual processing

Abstract

Sensory computations performed in the neocortex involve layer six (L6) cortico-cortical (CC) and cortico-thalamic (CT) signaling pathways. Developing an understanding of the physiological role of these circuits requires dissection of the functional specificity and connectivity of the underlying individual projection neurons. By combining whole-cell recording from identified L6 principal cells in the mouse primary visual cortex (V1) with modified rabies virus-based input mapping, we have determined the sensory response properties and upstream monosynaptic connectivity of cells mediating the CC or CT pathway. We show that CC-projecting cells encompass a broad spectrum of selectivity to stimulus orientation and are predominantly innervated by deep layer V1 neurons. In contrast, CT-projecting cells are ultrasparse firing, exquisitely tuned to orientation and direction information, and receive long-range input from higher cortical areas. This segregation in function and connectivity indicates that L6 microcircuits route specific contextual and stimulus-related information within and outside the cortical network.

Figure 1

Functional Diversity and Morphologies of V1 L6 Projection Cells (A) Tuning polar plots for all regular spiking (RS) L6 cells in this study that fired action potentials in response to moving gratings. The strength and tuning of AP firing is indicated by the radial length and orientation of the filled area. Red arrowheads indicate the cells’ preferred direction. The scale bar represents 0.25 Hz where not indicated. The histogram shows the population orientation tuning and number of cells that failed to spike during the presentation of any grating; the bin sizes are 0.15 (left) and 0.01 (right). (B) 3D density projections of the dendrites and axons of neurons separated according to the absence (CC, n = 6) or presence (CT, n = 10) of an axonal thalamic projection. (C) Schematic of the cortico-cortical and cortico-thalamic pathways morphologically defined by the axonal projection of these two classes of principal neurons.

Figure 2

Orientation-Dependent AP Tuning in CC and CT Neurons (A) Left: examples of four membrane voltage traces of spiking responses to gratings moving in the preferred, orthogonal, and antipreferred (null) directions for a CC cell. Spikes are represented as raster ticks under the traces. The black bar indicates the stimulus motion, and the shaded area indicates the analysis time window. Right: polar plots are shown from four representative CC cells). The polar plot corresponding to the example cell is shown (top left). (B) Average polar plot from all CC cells that fired spikes in response to moving gratings aligned to preferred direction. (C) Left: example of five membrane voltage traces of spiking responses to gratings moving in the preferred, orthogonal, and antipreferred (null) directions for a CT cell. Spikes are represented as raster ticks under the traces. The shaded area indicates the analysis time window. Right: polar plots from four representative CT cells are shown. The polar plot corresponding to the example cell is shown (top left). (D) Average polar plot from all CT cells that fired spikes in response to moving gratings aligned to preferred direction. (E) Box plot of AP orientation selectivity scores for all CC and CT cells exhibiting evoked firing. (F) Bar graphs of evoked mean AP firing rates for CT and CC neurons excluding and including those cells in which no evoked APs were observed. Error bars show SEM.

Figure 3

Orientation-Dependent Synaptic Tuning in CC and CT Neurons (A) Left: example of four membrane voltage recordings of the synaptic response to gratings moving in the preferred, orthogonal, and antipreferred (null) directions for a CC cell (spikes are clipped). The black bar indicates the stimulus motion, and the shaded area indicates the analysis time window. Right: CC population polar plots show the tuning of the mean PSP integral (top) and peak amplitude (bottom) for all orientations. (B) Left: example of five membrane voltage recordings of the synaptic response to gratings moving in the preferred, orthogonal, and antipreferred (null) directions for a CT cell (spikes are clipped). The shaded area indicates the analysis time window. Right: CT population polar plots show the tuning of the mean PSP integral (top) and peak amplitude (bottom) for all orientations. (C) Normalized tuning plot comparing the integral of the PSP depolarization for CC and CT cells for each grating orientation. The shaded line indicates the standard error of the mean of the CC population at the preferred direction. (D) Orientation selectivity index scores of the integral of the evoked PSP (OSI PSP_integral) for CC and CT cells. (E) Normalized tuning plot directly comparing the peak amplitude of the PSP depolarization for CC and CT cells for each grating orientation. The shaded line indicates the standard error of the mean of the CC population at the preferred direction. (F) Orientation selectivity index scores of the evoked PSP peak (OSI PSP_peak) for CC and CT cells. Error bars show SEM.

Figure 4

Output Tuning Is Independent of Biophysical Properties (A) Schematic showing the design for experiments performed in (B)–(D). Individual membrane potential traces recorded in response to drifting gratings for the preferred and related cardinal directions recorded in a CC (blue) and a CT (brown) cell. Polar plots show the mean AP tuning for the same cardinal directions in the same two cells. (B) Average and five individual membrane potential traces recorded in a CC cell during injection of the CT PSP waveform (brown). Spikes recorded in response to the injected waveforms are indicated by the raster plot (black). An example of the injected CC PSP waveform (blue) and resultant spikes is recorded in the same CC cell. (C) Left: polar plots for an example CC and CT cell in which the injected waveforms are the same as shown in (A). These polar plots may be directly compared to (A). Right: population polar plots for comparing injections of CC and CT responses into either CC or CT cells are shown. Three different sets of injection waveforms were used. (D) Box plot showing the range of orientation selectivity index scores for all injected cells. Plots are aligned to the preferred orientation of the AP output of the cells from which the injected waveforms were obtained. (E) Top: example polar plots from a CC (blue) and CT (brown) cell showing the tuning of the peak amplitude of the PSP and AP. PSP polar plots are displaying four repetitions of each stimulus. AP polar plots are showing the mean firing rate for each repetition. Bottom: a histogram of the difference in the orientation preference of the PSP_peak versus the AP response in all CC and CT cells. Error bars show SEM.

Figure 5

Mapping Connectivity onto Individual CC Cells (A) During whole-cell recording, the cell was loaded with DNA plasmids to drive expression of the rabies glycoprotein (RVG) and the avian virus receptor (TVA). This was followed by injection of the modified rabies virus (ΔRV) into the local area that results in targeted infection of the recorded neuron and subsequent retrograde spread and expression of RV-RFP. (B) After at least 10 days postrecording, the brain was fixed and placed under a serial two-photon microscope (left) for whole-brain serial imaging. Inset: a coronal postimmunostained confocal image of the recorded (yellow) and local presynaptic cells (red) is shown. (C) Left: membrane-voltage traces recorded at and two times the rheobase. Top left: the instantaneous frequency of AP firing at two times the rheobase is shown. Right: tuning polar plots of the same CC cell recorded during delivery of plasmids for RV targeting and tracing. (D) Top: coronal two-photon whole-brain image stack showing the location of cells labeled with the modified rabies virus following electrophysiological characterization of the recorded cell in (C). Bottom: following imaging, the labeled cells were localized using a standard mouse brain atlas. Regions relevant to this study include the primary visual cortex (V1), the medial and lateral secondary visual cortices (V2M and V2L, respectively), the retrosplenial cortex (RSP), and the thalamus (TH). (E) Example coronal images of the marked location of labeled cells (red spheres) within V1 (local) and outside V1 (long range). (F) Histogram showing the relative distribution of labeled cells (n = 3 mice). Error bars show SEM.

Figure 6

Connectivity Maps of CT and NTSR1-Expressing Cells (A) Left: membrane-voltage traces recorded at and two times the rheobase. Top left: the instantaneous frequency of AP firing at two times the rheobase is shown. Right: tuning polar plots of a CT cell recorded while delivering plasmids for RV targeting and tracing are shown. (B) Coronal projection of a two-photon whole-brain image stack showing the location of cells labeled with the modified rabies virus following electrophysiological characterization of the recorded cell in (A). (C) Example coronal images of the location of labeled cells within V1 (local) and outside V1 (long range). (D) Histogram showing the relative distribution of labeled cells for CT and CC cells (n = 4 and 3 mice, respectively). Inset: the average tuning profile (aligned to the preferred direction) of the recorded host cells in which single-cell rabies tracing was performed is shown. (E) Schematic showing the experimental design whereby a cre-dependent AAV is injected into cre-NTSR1^+ve mice for targeted RV infection of CT cells. (F) Two-photon whole-brain image stack showing the location of labeled presynaptic cells. In this brain, 421 putative host cells (not shown) were all located in L6 within V1. (G) Example two-photon images showing local and long-range connectivity onto the NTSR1^+ve cell population. (H) Histograms showing the fraction of presynaptic cells located within and outside V1 for CC, CT, and cre-NTSR1^+ve cells. Error bars show SEM.

Figure 7

Functional Specificity and Connectivity of CC and CT Pathways Left: schematic showing CC-projecting cells receiving inputs from neurons located primarily within V1. On average, CC cells receive weakly tuned synaptic input and show poorly tuned output firing. Inset: a population histogram of PSP_peak (input, dashed line, n = 17) and AP (output) tuning (n = 15) for all CC cells recorded in this study (bin size = 0.1) is shown. Right: a schematic showing CT-projecting cells receiving comparatively more long-range inputs from neurons located in V2 and RSP is shown. Inset: a population histogram of PSP_peak (input, dashed line, n = 28) and AP (output) tuning (n = 19) for all CT cells recorded in this study (bin size = 0.1) is shown.