Synaptic mechanisms underlying functional dichotomy between intrinsic-bursting and regular-spiking neurons in auditory cortical layer 5

Abstract

Corticofugal projections from the primary auditory cortex (A1) have been shown to play a role in modulating subcortical processing. However, functional properties of the corticofugal neurons and their synaptic circuitry mechanisms remain unclear. In this study, we performed in vivo whole-cell recordings from layer 5 (L5) pyramidal neurons in the rat A1 and found two distinct neuronal classes according to their functional properties. Intrinsic-bursting (IB) neurons, the L5 corticofugal neurons, exhibited early and rather unselective spike responses to a wide range of frequencies. The exceptionally broad spectral tuning of IB neurons was attributable to their broad excitatory inputs with long temporal durations and inhibitory inputs being more narrowly tuned than excitatory inputs. This uncommon pattern of excitatory-inhibitory interplay was attributed initially to a broad thalamocortical convergence onto IB neurons, which also receive temporally prolonged intracortical excitatory input as well as feedforward inhibitory input at least partially from more narrowly tuned fast-spiking inhibitory neurons. In contrast, regular-spiking neurons, which are mainly corticocortical, exhibited sharp frequency tuning similar to L4 pyramidal cells, underlying which are well-matched purely intracortical excitation and inhibition. The functional dichotomy among L5 pyramidal neurons suggests two distinct processing streams. The spectrally and temporally broad synaptic integration in IB neurons may ensure robust feedback signals to facilitate subcortical function and plasticity in a general manner.

Figure 1.

Response properties of RS and IB neurons in L5 of the rat A1. A, Spike TRF of an example RS neuron. Left, Array of PSTHs for responses to pure tones of various frequencies and intensities. Each PSTH trace depicts the spike response evoked by a 50 ms tone, sampled over five trials. Calibration: 1 spike per 10 ms bin, 50 ms. Right, Color map depicts the average evoked spike number in the frequency–intensity space. Inset, Three example response traces to the CF tone (at 70 dB), with the red line denoting the stimulus duration. Example spontaneous spikes outside the evoked response window are also shown on the right. The PSTH at the bottom was generated from responses to all the tone stimuli (bin size of 1 ms). B, An example IB neuron. Data are presented in the same way as in A. Calibration: 2 spikes per 10 ms bin, 50 ms. Note that CF tone evoked bursts of spikes and that spontaneous bursts were also observed in this neuron. Within the burst, the spike amplitude reduced over time. C, Color maps for TRFs of another three RS neurons. D, Color maps for TRFs of another three IB neurons. E, Average bandwidth at 10 dB above the intensity threshold (BW10) of the spike TRF for RS, IB, and L4 pyramidal neurons. One-way ANOVA (F = 83.2, p = 2.2 × 10⁻¹⁶) and multiple comparison Scheffé's test showed significant differences between RS–IB and IB–L4 groups (***p < 0.001). The cell number is indicated in the bars. F, Average spontaneous spike rate and evoked spike rate for RS and IB neurons. **p < 0.01, two-sample t test with unequal variance (p = 0.0051 for spontaneous, p = 0.0087 for evoked). G, Average spike latency, defined as the interval between the onset of the tone and the time point at which the spike rate in the PSTH becomes higher than the average baseline level by 3 SDs of the baseline fluctuation. One-way ANOVA (F = 42.5, p = 1.9 × 10⁻⁹) and Scheffé's test showed significant differences between RS–IB and RS–L4 groups (***p < 0.001). H, Plot of BW10 versus spike latency. RS (black) and IB (red) neurons distinguished by the absence/presence of bursting firing patterns were partitioned into two separate clusters by the K-means clustering analysis. Whiskers show mean ± SD.

Figure 2.

Subthreshold membrane potential responses of RS and IB neurons. A, Sequential cell-attached and current-clamp recordings from an example RS neuron. Top, Spike TRF (from 1 sample trial) recorded in the cell-attached mode. Color map depicts the average spike number over eight trials. Below the color map are example response traces to the CF tone and a tone outside the TRF (NR tone), with the red line denoting the 50 ms tonal stimulation. Boxed are 50 superimposed spike waveforms (3.5 ms trace), with vertical lines and arrows denoting the TPI (0.8 ms in this cell). Bottom, Subthreshold membrane potential responses recorded in the current-clamp mode (spikes were filtered out). Color map depicts the average peak depolarization voltage (millivolts) over four trials. Below the color map is the reconstructed morphology of this RS neuron. Note that the apical dendrite ended in L2/L3. B, Sequential cell-attached and current-clamp recordings from an example IB neuron. Data are presented in the same way as in A. The TPI was 0.7 ms. Note that this IB neuron had a tufted apical dendrite reaching L1. C, Top, Average bandwidths of spike (supra) and subthreshold (sub) response at 20 dB above the intensity threshold of the subthreshold TRF (BW20). The bandwidth of spike response in this measurement is consistent with BW10 in Figure 1E, because the threshold for spike response is usually 10 dB higher than that for subthreshold response. Bottom, Average onset latencies of spike and subthreshold depolarization response. **p < 0.01, ***p < 0.001, two-sample t test with equal variance. D, Traces of evoked subthreshold responses of sample RS and IB neurons. Traces are averaged responses (normalized) of four to six repetitions to tones at and near CF (±0.2 octave) at 70 dB. Red line denotes the tone duration. Black line denotes the level of the resting potential (V_R). Dash line marks the half-peak duration of the depolarization response. Arrow marks the time point of 100 ms after the tone onset. E, Average duration of the rising phase of the depolarization response (white), and half-peak duration (gray). ***p < 0.001, two-sample t test with unequal variance. F, Dendritic morphologies and laminar locations of the reconstructed cells. G, Plot of cells in three dimensions. The three axes represent BW20, half-peak duration, and onset latency of depolarization response. Data points were segregated into two clusters based on the K-means clustering analysis, which were consistent with the grouping based on spike patterns. The RS and IB groups are outlined by the pink and blue oval, respectively.

Figure 3.

Synaptic inputs to RS and IB neurons. A, Excitatory (top) and inhibitory (bottom) TRFs (average of 3 trials) of a putative RS neuron recorded under clamping voltage of −80 and 0 mV, respectively. Color maps depict the peak amplitude of synaptic currents (nanoamperes). Below the color map is an example response trace evoked by a best-frequency tone at 70 dB, with the red line denoting the stimulus duration. B, Excitatory and inhibitory TRFs of a putative IB neuron. Data are presented in the same way as in A. C, Plot of cells in three dimensions. The three axes represent latency and half-peak duration of excitatory responses and half-maximum bandwidth of the excitatory tuning curve at 20 dB above the intensity threshold (50% BW20). Cells were partitioned into two separate clusters based on the K-means clustering analysis, which were categorized as the RS (outlined by pink) and IB (outlined by blue) group, respectively. D, Average 50% BW20 (top) and BW20 (bottom) of excitatory (exc) and inhibitory (inh) TRFs for the three types of neurons. For excitatory 50% BW20, one-way ANOVA was significant (F = 48.7, p = 3.6 × 10⁻⁹), and Scheffé's test showed significant differences between RS–IB and IB–L4 groups (***p < 0.001). For inhibitory 50% BW20, there was no significant difference among the cell types (one-way ANOVA, F = 0.17, p = 0.84). For excitatory BW20, one-way ANOVA was significant (F = 20.8, p = 1.0 × 10⁻⁵), and Scheffé's test showed significant differences between RS–IB (*p < 0.05), RS–L4 (**p < 0.01), and IB–L4 (***p < 0.001) groups. For inhibitory BW20, there was no significant difference among the cell types (one-way ANOVA, F = 1.34, p = 0.28). Paired t test was applied within each cell type to compare excitation and inhibition (^#p < 0.05, ^##p < 0.01, ^###p < 0.001). E, Neuron modeling of membrane potential (V_m) response tuning under co-tuned inhibition (top) and narrower inhibition (bottom). Left, Boxed are the temporal profiles of the simulated excitatory (red) and inhibitory (black) conductance G (top) and that of the derived V_m response (bottom). Middle, The frequency tuning curves of excitatory and inhibitory inputs are modeled with Gaussian functions. In the co-tuned inhibition, excitatory and inhibitory tuning curves have the same bandwidths. In the narrower inhibition, the inhibitory tuning is narrower than the excitatory tuning. Right, The derived V_m response tuning in the absence (magenta) or presence (cyan) of inhibition. Dash line denotes the level of spike threshold (20 mV above the resting potential).

Figure 4.

Temporal properties of synaptic inputs to different types of neurons. A, Top, Normalized synaptic conductances (red for excitation, reversed in polarity) evoked by CF tones for three RS and three IB neurons. Dash line marks the half-maximum level. Bottom, The V_m response generated in the neuron model by integrating the synaptic conductances shown above. Dash line labels the level of the resting potential. Red line indicates the tone duration. B, Average half-peak durations of evoked excitatory (red) and inhibitory (black) conductances. For excitation, one-way ANOVA (F = 21.5, p = 8.1 × 10⁻⁶) and Scheffé's test showed a significant difference between RS–IB and IB–L4 groups (***p < 0.0001). For inhibition, there was no significant difference among cell groups (one-way ANOVA, F = 0.28, p = 0.76). ^##p < 0.01, paired t test within cell type. C, Average onset latencies of CF-tone-evoked excitation and inhibition. One-way ANOVA (F = 7.6, p = 0.0032 for excitation; F = 3.54, p = 0.0475 for inhibition) and Scheffé's test (*p < 0.05) showed significant differences between RS–IB and RS–L4 groups. ^###p < 0.001, paired t test within cell type. D, Average E/I ratio measured by peak response amplitude (white) or integrated charge (gray). *p < 0.05, two sample t test with unequal variance. E, The excitatory response of an IB neuron was artificially scaled down at different levels as to change E/I ratio (the E/I ratio value is given on top). The resulting V_m response from integrating the excitation (exc) and inhibition (inh) shown in the top panel consistently exhibited prolonged depolarization.

Figure 5.

IB neurons receive direct thalamocortical input. A, Color maps of example multiunit spike TRFs in cortical L4–L6 and in the ventral division of the medial geniculate nucleus (MGBv) before and after cortical injection of the muscimol mixture. Color represents evoked spike number. Color scale (from left to right): 17, 18, 8, 11 for maximum. B, Remaining number of evoked spikes after cortical silencing (in percentage of the initial spike number) at different laminar locations plotted against the horizontal distance away from the injection site. n = 3. Error bar indicates SD. C, Percentage change in evoked spike number and bandwidth of multiunit spike TRF in the MGBv. n = 3. Error bar indicates SD. D, Excitatory TRFs of three example cells recorded in the silenced A1. Note that, in the “L5 silenced” neuron, spontaneous excitatory currents were observed in some trials. Color map depicts the average peak amplitude of excitatory current over three trials. Color scale (from left to right): 40, 50, 40 pA for maximum. Enlarged traces (to the best-frequency tone at 70 dB) in the insets are to show response temporal profiles. E, Summary of excitatory onset latency for IB neurons in the normal A1, L5 active neurons in the silenced A1, L4 neurons in the normal and silenced A1. F, Summary of BW20 of excitatory TRF. ***p < 0.001, two-sample t test with unequal variance. G, Summary of half-peak duration of CF-tone-evoked excitatory response. One-way ANOVA (F = 21.8, p = 2.2 × 10⁻⁷) and Scheffé's test (*p < 0.001). H, Tone-evoked excitatory responses (at a level of 20 dB above the intensity threshold) in the same neuron before and after cortical silencing. Color map depicts the peak response amplitude. Color scale: 100, 60, 40 pA for maximum. I, Summary of onset latencies of CF-tone-evoked excitatory responses before and after cortical silencing for active (putative IB; act IB) neurons and silenced (putative RS; sil RS) neurons. One-way ANOVA (F = 8.8, p = 0.07) and Scheffé's test (*p < 0.05). J, Summary of BW20 of excitatory TRF. One-way ANOVA (F = 18.8, p = 6.2 × 10⁻⁴) and Scheffé's test (**p < 0.01). K, Summary of half-peak duration of CF-tone-evoked excitatory response. One-way ANOVA (F = 18.8, p = 6.2 × 10⁻⁴) and Scheffé's test (**p < 0.01).

Figure 6.

Properties of FS neurons in L5 and potential circuits. A, Sequential cell-attached and current-clamp recordings from an FS neuron. Top, Spike TRF (one trial) recorded in the cell-attached mode. Color map is the averaged spike TRF over seven trials. Inset, Example spike response traces to the CF tone and a tone outside the TRF. Note that the spike amplitude did not reduce within a burst. Boxed are 50 superimposed spike waveforms. Note that the spike shape is much narrower than that of pyramidal cells. The TPI was 0.3 ms in this FS neuron. Bottom, Its subthreshold TRF recorded in the current-clamp mode. B, Summary of TPI of spike waveform for RS, IB, and FS neurons. One-way ANOVA (F = 18.8, p = 6.2 × 10⁻⁴) and Scheffé's test (***p < 0.001). C, Dendritic morphologies of three reconstructed L5 FS neurons. D, Average bandwidths of spike and subthreshold TRFs for IB and FS neurons in the sequential cell-attached and current-clamp recordings. *p < 0.05, **p < 0.01, ***p < 0.001, two-sample t test with unequal variance. E, Comparison of onset latencies of spike and subthreshold depolarization responses among cell classes. One-way ANOVA (F = 26.5, p = 8.6 × 10⁻¹¹ for spike latency; F = 9.4, p = 5.3 × 10⁻⁵ for subthreshold latency) and Scheffé's test (^##p < 0.01; ^###p < 0.001). *p < 0.05, two-sample t test with equal variance between IB and FS spike latency. F, A model for L5 circuits. IB neurons receive direct thalamic input and feedforward inhibitory input from L5 FS inhibitory neurons, as well as polysynaptic excitatory input from upper layers and possibly within L5. RS neurons receive polysynaptic excitation from upper layers and possibly within L5. Their inhibitory input is likely from interneurons (IN) other than FS neurons, which are not directly driven by thalamic input. The size of arrow depicts the tuning broadness of the corresponding input. Red arrow represents excitatory input, and blue arrow represents inhibitory input.