Parvalbumin-Expressing GABAergic Neurons in Mouse Barrel Cortex Contribute to Gating a Goal-Directed Sensorimotor Transformation

Abstract

Sensory processing in neocortex is primarily driven by glutamatergic excitation, which is counterbalanced by GABAergic inhibition, mediated by a diversity of largely local inhibitory interneurons. Here, we trained mice to lick a reward spout in response to whisker deflection, and we recorded from genetically defined GABAergic inhibitory neurons in layer 2/3 of the primary somatosensory barrel cortex. Parvalbumin-expressing (PV), vasoactive intestinal peptide-expressing (VIP), and somatostatin-expressing (SST) neurons displayed distinct action potential firing dynamics during task performance. Whereas SST neurons fired at low rates, both PV and VIP neurons fired at high rates both spontaneously and in response to whisker stimulation. After an initial outcome-invariant early sensory response, PV neurons had lower firing rates in hit trials compared to miss trials. Optogenetic inhibition of PV neurons during this time period enhanced behavioral performance. Hence, PV neuron activity might contribute causally to gating the sensorimotor transformation of a whisker sensory stimulus into licking motor output.

Graphical abstract

Figure 1

Cell-Type-Specific AP Firing of GABAergic Neurons in Hit Trials during a Whisker Detection Task (A) Top left: setup for two-photon (2P) guided targeting of juxtasomal recordings during the head-fixed whisker deflection detection task. Top right: 2P view shows a PV neuron expressing tdTomato (red) targeted for juxtasomal recording with a pipette containing Alexa-488 (green), together with an example spike recorded from the PV neuron. Bottom left: schematic shows trial types and outcomes of the behavioral task. Bottom right: plot of d′ against reaction time for the different recordings indicates no difference in performance among the different genotypes of mice (each point represents an individual recording from a specific genetically labeled neuron, as indicated by color coding). Lines indicate best fits for PV, VIP, and SST data (linear correlation for PV: r = −0.0004, p = 0.99, t test, n = 17 cells; for VIP: r = −0.18, p = 0.46, t test, n = 19 cells; for SST: r = −0.16, p = 0.50, t test, n = 21 cells). (B) Example hit trials and grand average peri-stimulus time histogram (PSTH) of AP discharge of PV, VIP, and SST neurons for hit trials in response to C2 whisker stimulus. Note the difference in scale for the number of APs discharged per 5-ms bin between the different types of GABAergic neurons. (C) Baseline (1 s) and post-whisker stimulus (5–50 ms and 50–100 ms) AP discharge rates are shown. (D) AP latencies of the first spike after whisker stimulus for PV, VIP, and SST neurons are shown. Open circles represent individual cells. Filled circles with error bars represent group averages shown as mean ± SEM. Boxplots represent the median, the 25^th and 75^th percentiles in the boxes, with the side bars representing the 5^th and 95^th percentiles of the distribution. Statistical significance is indicated as follows: ^∗p < 0.05 and ^∗∗∗p < 0.005. See also Figure S1.

Figure 2

PV-Expressing GABAergic Neurons Fire Fewer APs in Hit Trials Compared to Miss Trials (A) Left: example hit and miss trials during a recording from a PV neuron. Right: grand average PSTH shows PV neurons recorded during the detection task, analyzed separately for hit (red) and miss (black) trials, together with a histogram of the first lick reaction time (brown). (B) AP discharge rates of PV neurons before whisker stimulus (−1,000–0 ms), during early sensory processing (5–50 ms post-whisker stimulus), during the late period (50–250 ms), and during licking (250–750 ms) in hit and miss trials are shown. (C) AP discharge rate difference between hit and miss trials of PV neurons during the late period is shown. Lines and open circles represent individual cells. Filled circles with error bars represent group averages shown as mean ± SEM. Boxplots represent the median, the 25^th and 75^th percentiles in the boxes, with the side bars representing the fifth and 95^th percentiles of the distribution. Statistical significance is indicated as follows: ^∗p < 0.05 and ^∗∗p < 0.01. See also Figure S2.

Figure 3

Comparison of Hit and Miss Trials for VIP- and SST-Expressing GABAergic Neurons (A) Left: example hit and miss trials during a recording from a VIP neuron. Right: grand average PSTH shows VIP neurons during the detection task, in hit and miss trials together, with a histogram of the first lick reaction time. (B) AP discharge rates of VIP neurons before whisker stimulus (−1,000–0 ms), during early sensory processing (5–50 ms post-whisker stimulus), during the late period (50–250 ms), and during licking (250–750 ms) in hit and miss trials. (C) AP discharge rate difference between hit and miss trials of VIP neurons during the late period is shown. (D–F) Same as (A)–(C) are shown, but for SST neurons. Lines and open circles represent individual cells. Filled circles with error bars represent group averages shown as mean ± SEM. Boxplots represent the median, the 25^th and 75^th percentiles in the boxes, with the side bars representing the fifth and 95^th percentiles of the distribution. No statistically significant differences were found comparing hit and miss trials.

Figure 4

Optogenetic Inhibition of PV-Expressing GABAergic Neurons in S1 Can Enhance Behavioral Performance (A) Left: a 2P image showing a juxtasomal recording electrode targeted to a PV neuron expressing eNpHR3.0. Center: example traces of a PV neuron expressing NpHR show AP discharge suppression upon application of yellow light (80–180 ms post-whisker stimulus, green shading) coupled with C2 whisker stimulus. Right: group statistics of AP discharge suppression in PV-NpHR neurons quantified 80–180 ms post-whisker stimulus are shown. (B) Left: a 2P image showing a juxtasomal recording electrode targeted to a non-PV (presumed excitatory) neuron. Center: example traces of a non-PV neuron show enhanced AP discharge upon application of yellow light (80–180 ms post-whisker stimulus) coupled with C2 whisker stimulus. Right: group statistics of AP discharge in these non-PV neurons quantified 80–180 ms post-whisker stimulus are shown. (C) Yellow light coupled with C2 whisker stimulus enhanced performance over C2 stimulus alone in PV-NpHR mice. Yellow light delivery alone did not result in an increase in false alarm rates. Gray lines represent individual cells in (A) and (B). Gray lines and gray circles represent individual mice in (C). Color-coded circles with error bars represent group averages shown as mean ± SEM. Statistical significance is indicated as follows: ^∗p < 0.05 and ^∗∗∗p < 0.005. See also Figure S3.