Laminar specificity of functional input to distinct types of inhibitory cortical neurons

Abstract

Despite the presence of numerous inhibitory cell types, laminar excitatory input has only been characterized for limited identified types, and it is unknown whether there are differences between cell types in their laminar sources of inhibitory input. In the present study, we characterized sources of local input to nine distinct types of layer 2/3 inhibitory neurons in living slices of mouse somatosensory cortex. Whole-cell recordings from identified cell types, facilitated by use of transgenic mice expressing green fluorescent protein in limited inhibitory neuron populations, were combined with laser scanning photostimulation. We found that each inhibitory cell type received distinct excitatory and inhibitory laminar input patterns. Excitatory inputs could be grouped into three categories. All inhibitory cell types received strong excitation from layer 2/3, and for calretinin (CR)-positive Martinotti cells and burst-spiking interneurons, this was their dominant source of excitatory input. Three other cell types, including fast-spiking basket cells, CR-negative Martinotti cells, and bipolar interneurons, also received strong excitatory input from layer 4. The remaining four inhibitory cell types, including chandelier cells, neurogliaform cells, irregular spiking basket cells, and regular spiking presumptive basket cells, received strong excitatory input from layer 5A and not layer 4. Laminar sources of inhibitory input varied between cell types and could not be predicted from the sources of excitatory input. Thus, there are cell-type specific differences in laminar sources of both excitation and inhibition, and complementary input patterns from layer 4 versus layer 5A suggest cell type differences in their relationships to lemniscal versus paralemniscal pathways.

Figure 1.

Laminar patterns of excitatory and inhibitory input to pyramidal cells and fast-spiking cells. A, An example map of excitatory input to one pyramidal neuron. A reconstruction of the neuron's cell body and major dendrites is shown in black over the color-coded excitatory input map. The color scale codes evoked input in units of picoamperes, calculated by subtracting the mean spontaneous currents from the mean currents measured after photostimulation for each stimulation site during the 150 ms analysis window (see Materials and Methods). Values for each position are calculated from a linear interpolation of values at discrete stimulation sites. Red indicates strong excitatory input, and blue range indicates no or weak excitatory input as indicated by the color scale beneath the map. The thin black lines delineate borders between layers which are labeled to the right. B, An example map of inhibitory input to a different pyramidal cell in layer 2/3. The cell drawing is shown in white. The inhibitory input map uses a different color scale, where the yellow/white range indicates strong inhibitory input, and the dark red range indicates no or weak inhibitory input, as indicated by the color scale beneath the map. The thin white lines delineate laminar boundaries. In both A and B, small gray squares indicate stimulation sites omitted from quantitative analysis because of interference of large direct currents during photostimulation (see Materials and Methods and supplemental Fig. S3, available at www.jneurosci.org as supplemental material). Examples of excitatory (sites i and ii in A) and inhibitory (sites iii and iv; B) postsynaptic current responses of the pyramidal cells are shown in A. Only the 200 ms duration of the 400 ms recorded traces are shown, with short red marks beneath the traces indicating the 10 ms of the laser on time. C, The summary data of laminar excitatory input strength (%EI) for 10 pyramidal cells. D, The summary data of laminar inhibitory input strength (%EI) pooled from seven pyramidal cells. Bars in C and D represent mean ± SE values. Note that negative values indicate that spontaneous activity during control trials was greater than the level of activity during stimulation trials, but it was not statistically significant (see supplemental Table S1, available at www.jneurosci.org as supplemental material). E, F, Examples of excitatory and inhibitory input maps from two different fast-spiking cells, respectively (for morphological and electrophysiological features of fast-spiking cells, see supplemental Fig. S1A–C, available at www.jneurosci.org as supplemental material). Examples of excitatory (sites v and vi; E) and inhibitory (sites vii and viii; F) postsynaptic current responses of the fast-spiking cells are shown by E. G, H, Summary data of laminar input strength (mean ± SE; %EI) based on 12 excitatory input maps and 11 inhibitory input maps, respectively. Scale bars, 100 μm.

Figure 2.

Laminar patterns of excitatory and inhibitory input to two subtypes of Martinotti cells. A, B, Example excitatory input maps from a SOM+/CR− Martinotti cell and a SOM+/CR+ Martinotti cell, respectively (for morphological, immunochemical, and electrophysiological features of the SOM+/CR− and SOM+/CR+ cells illustrated in A and B, see supplemental Fig. S1G–L, available at www.jneurosci.org as supplemental material). C, D, Example inhibitory input maps from a different SOM+/CR− Martinotti cell and a different SOM+/CR+ Martinotti cell, respectively. Scales bars: A–D, 100 μm. E, Average laminar strength of excitatory input (%EI) for SOM+/CR− cells (N = 18; solid red bars) versus SOM+/CR+ cells (N = 25; open bars). These two subtypes differed significantly both in their input strength (% EI) from layer 2/3 and from layer 4 (p < 0.005). (Asterisks denote statistical significance.) F, Average laminar strength of inhibitory input (%EI) for SOM+/CR− cells (N = 6; solid blue bars) versus SOM+/CR+ cells (N = 9; open bars). Generally, the two subtypes have similar inhibitory input patterns. Conventions as in Figure 1.

Figure 3.

Laminar patterns of excitatory and inhibitory input to burst-spiking cells. A, B, Example maps of excitatory and inhibitory input from a BS cell (for morphological and electrophysiological features of BS cells, see supplemental Fig. S2G–I, available at www.jneurosci.org as supplemental material). C, D, Summary data of laminar input strength (mean ± SE; %EI), calculated from eight excitatory input maps and nine inhibitory input maps from BS cells. Scale bars, 100 μm. Conventions are as in previous figures.

Figure 4.

Laminar patterns of excitatory and inhibitory input to multipolar irregular-spiking cells and regular-spiking cells. A, B, Example maps of excitatory and inhibitory input from an IS cell (for morphological and electrophysiological features of the example IS cell illustrated in A and B, see supplemental Fig. S2A–C, available at www.jneurosci.org as supplemental material). C, D, Summary data of laminar input strength (mean ± SE; %EI), calculated from 11 excitatory input maps and nine inhibitory input maps from IS cells. E, F, Example maps of excitatory and inhibitory input from an RS cell (for morphological and electrophysiological features of RS cells, see supplemental Fig. S2D–F, available at www.jneurosci.org as supplemental material). G, H, Summary data of laminar input strength (mean ± SE; %EI), calculated from seven excitatory input maps and six inhibitory input maps from RS cells. Conventions are as in previous figures.

Figure 5.

Laminar patterns of excitatory and inhibitory input to chandelier cells and neurogliaform cells. A, B, Excitatory and inhibitory input maps for an example chandelier cell (for morphological and electrophysiological features of the chandelier cell illustrated in A and B, see supplemental Fig. S1D–F, available at www.jneurosci.org as supplemental material). C, D, Summary data of laminar input strength (mean ± SE; %EI) from three excitatory input maps and four inhibitory input maps from chandelier cells. E, F, Example maps of excitatory and inhibitory input, respectively, for a neurogliaform cell (for morphological and electrophysiological features of the neurogliaform cell illustrated in E and F, see supplemental Fig. S2M–O, available at www.jneurosci.org as supplemental material). G, Average laminar strength of excitatory input for four neurogliaform cells, and H shows average inhibitory input strength from these cells. Conventions are as in previous figures.

Figure 6.

Laminar patterns of excitatory and inhibitory input to bipolar cells. A, Excitatory input map from a bipolar cell, while B shows an inhibitory input map from another bipolar cell (for morphological and electrophysiological features of the bipolar cell illustrated in A, see supplemental Fig. S2J–L, available at www.jneurosci.org as supplemental material). C, Average laminar strength for excitatory input from the bipolar cells, and D shows their average inhibitory input strength. Conventions are as in previous figures.

Figure 7.

Summary of excitatory input patterns across all the cell types examined. A, A series of cumulative histograms with average percentages of excitatory input (%EI) from all the layers examined (layers 2/3, 4, 5a, 5b, and 6) for all the cell types examined: BS cells, SOM+/CR+ Martinotti cells (MC+), SOM+/CR− Martinotti cells (MC−), FS basket cells, BC, pyramidal cells (PY), chandelier cells (Chand), neurogliaform cells (NG), RS cells, and IS cells. B, A three-dimensional plot that shows average excitatory %EIs of layers 2/3, 4, and 5a for the cell types examined. C, Average excitatory EI amplitudes from layers 2/3, 4, and 5a for the cell types examined. Bars represent mean ± SE.

Figure 8.

Summary of inhibitory input patterns across all the cell types examined. A, Cumulative histograms of percentages of inhibitory input (%EI) from layers 2/3, 4, 5a, 5b, and 6 for all the cell types examined: chandelier cells (Chand), neurogliaform cells (NG), pyramidal cells (PY), FS basket cells, RS cells, BS cells, IS cells, SOM+/CR− Martinotti cells (MC−), SOM+/CR+ Martinotti cells (MC+), and BC. B, A three-dimensional plot that shows average inhibitory %EIs of layers 2/3, 4, and 5 for the cell types examined. In B (unlike in A), the %EIs of layers 5a and 5b were pooled into the %EIs of layer 5. Also note that the recalculated %EIs for layers 2/3 and 4 in B are slightly different than for A, but the overall patterns are similar. C, Average inhibitory EI amplitudes from layers 2/3, 4, and 5a for the cell types examined. Bars represent mean ± SE.

Figure 9.

Differing excitatory input patterns and their corresponding functional differences in local cortical circuits. A schematic drawing illustrates that (1) the layer 2/3 inhibitory neurons that receive excitatory input from layer 2/3 excitatory pyramidal cells (observed for all cell types examined) have the potential to provide feedback inhibition onto the excitatory pyramidal cells; (2) the layer 2/3 inhibitory neurons such as SOM+/CR− Martinotti cells (MC−), FS and bipolar cells that also receive strong feedforward excitation from layer 4, have the potential to exert both lemniscal feedforward inhibition and feedback inhibition onto layer 2/3 pyramidal cells; (3) the layer 2/3 inhibitory neurons such as chandelier cells, neurogliaform cells, and IS cells receive strong feedforward excitation from layer 5a, and are thus more directly influenced by the paralemniscal pathway and have the potential to exert both paralemniscal feedforward inhibition and feedback inhibition onto layer 2/3 pyramidal cells.