Membrane potential correlates of sensory perception in mouse barrel cortex

(a) The probability of licking during the reward window in trials with stimuli of different intensities (black, hit rate) and trials without stimuli (red, catch trials for assessing false alarm rate) for well-trained mice (n = 5 mice). The maximum stimulus strength was used for all training and recording sessions. (b) When the iron filings were removed from the whisker, the mouse did not lick above the false alarm rate in response to magnetic stimuli (n = 3 mice). The detection task is therefore dependent upon whisker stimulation. Data shown as mean ± sem.

(a) A brief 1 ms magnetic pulse evoked a rapid whisker deflection measured with an optical displacement sensor. Averaged traces for the evoked magnetic field, whisker angle and whisker angular velocity. (b) Highly reproducible magnetic fields were generated by the magnetic coil for each individual stimulus. (c) The magnetic field reliably evoked whisker movements of similar angular deflection and velocity for each stimulus. (d) Values of the evoked magnetic field, whisker angle and whisker angular velocity for 20 consecutive trials show stability over repeated stimuli. (e) Mean and SD of the evoked magnetic field, whisker angle and whisker angular velocity from 3 mice (above), with corresponding coefficient of variation (CV) values (below). (f) Magnetic field strength measurements (blue) show that the peak amplitude of the generated field changes very little over the spatial scale explored by whisker movements. Over the centre of the coil we measured a peak magnetic field of 41.1 mT. Measurements of the magnetic field across a horizontal grid (5 × 5 mm away from the centre) show only small differences in the peak magnetic field strength. Angular whisker deflection amplitude (green) and peak angular whisker velocity (brown) evoked at different whisker orientations (centred around the middle of the coil) are similar, indicating that the magnetic field evoked similar whisker deflections regardless of whisker position.

(a) Averaged LFP responses recorded in the C2 barrel column and evoked by C2 whisker stimulus, before and after TTX injection into the C2 barrel column. (b) Averaged LFP responses recorded in the C2 barrel column and evoked by C2 whisker stimulus, before and after TTX injection into forepaw S1 representation, which is located ~1.5 mm anterior-medial to the C2 barrel column.

(a) Mice were trained to lick in response to C2 whisker stimulation and, after achieving stable, good behavioral performance, the lick response to optogenetic stimulation of S1 whisker barrel cortex was tested at different light durations (n = 6 mice; 5 ms data same as shown in Fig. 1e). (b) To control for non-specific effects of the blue light stimulus, the optical fiber was moved to a part of the cortex not expressing ChR2. Light flashes delivered to cortical regions not expressing ChR2 did not evoke licking (n = 6 mice). (c) ChR2-YFP was expressed in the S1 forepaw representation and mice were trained to detect C2 whisker stimuli. Coronal section through the center of the injection site in S1 forepaw cortex (left, 0.02 mm anterior Bregma) and another section from the same mouse showing no expression in whisker barrel cortex (right, 1.7 mm posterior to Bregma). On the transfer test day, mice failed to respond with licking to optogenetic stimulation of the S1 forepaw area (n = 3 mice). (d) Mice were trained to lick in response to optogenetic stimulation of S1 whisker barrel cortex. On the transfer test day, the ability of a C2 whisker stimulus, of differing pulse durations, to drive licking responses was checked (n = 3 mice; 1 ms data same as shown in Fig. 1f). (d) In subsequent ‘no whisker stimulus’ blocks of control trials, the iron filings were removed from the whisker. Magnetic pulses during these ‘No stim’ control trials did not evoke licking above the false alarm rate (n = 3 mice). Data shown as mean ± sem.

Evoked whisker movements and detection. a, Example hit and miss trials. Rhythmic whisker protraction was observed following detected whisker stimuli in hit trials prior to licking. b, Average traces for whisker position (above) and probability histogram of first lick time (below) (n = 24 trained mice; n = 4 naive mice). Note the large whisker movements in hit trials compared to miss trials and naive mice. c, Example traces of whisker position and licking evoked by C2 whisker stimuli before (left) and after (right) CNQX+APV injection into the C2 barrel column of the same mouse. d, The average whisker angle and first-lick histogram evoked by C2 whisker stimuli before and after CNQX+APV injection into the C2 barrel column (n = 4 mice). e, Mice expressing ChR2 in S1 whisker barrel cortex and trained on C2 whisker stimuli detection, displayed brief whisker retraction22 followed by rhythmic whisker protraction evoked by the 5 ms optogenetic stimulation of S1 whisker barrel cortex on the transfer test day. f, Mice expressing ChR2 in S1 whisker barrel cortex, trained on optogenetic stimuli detection, displayed rhythmic whisker protraction evoked by C2 whisker stimuli on the transfer test day. g, Whisker retraction is evoked by ChR2 stimulation of S1 barrel cortex (left)22. Whisker movement is absent following transection of the facial nerve (FN, the motor nerve controlling whisker movements) (center). Summary graph for evoked whisker movements, before and after FN transection (n = 3 mice) (right). h, FN transection had no significant effect on lick probability after whisker stimulus or during catch trials in trained mice (left). Following FN transection, ChR2 stimulation of S1 whisker barrel cortex remained capable of substituting for the whisker stimulus to evoke a licking response (n = 3 mice) (right). Data shown as mean ± sem.

(a) Four example hit trials from the recording of mouse 85 (same cell as shown in Fig. 2a). Mouse 85 displayed slow, correlated, large-amplitude fluctuations in prestimulus Vm and LFP, which were consistently observed across different trials. (b) Four example hit trials from the recording of mouse 82 (same cell as shown in Fig. 2b). Across trials mouse 82 consistently displayed a more desynchronized cortical state, with poorly correlated prestimulus Vm and LFP fluctuations. Action potential amplitude is truncated to expand the Vm scale. LFP is shown inverted.

(a) Prestimulus Vm FFT (over a 2 s period immediately preceding whisker stimulus) for hit trials, miss trials and naive mice (trained, n = 19 cells; naive, n = 10 cells). (b) Prestimulus Vm FFT 1–5 Hz integral in hit trials, miss trials and naive mice (trained, n = 19 cells; naive, n = 10 cells, P = 0.008, hit vs naive and P = 0.01, miss vs naive, Wilcoxon-Mann-Whitney test). Each grey line connects hit and miss data points from the same recording. The light grey circles show individual data points for naive mice. Average data shown as mean ± sem. Statistical significance is indicated by * for P < 0.05, ** for P < 0.01. (c) As for panel b, but computed for prestimulus Vm FFT 30–100 Hz integral. (d) There was no correlation between prestimulus Vm and prestimulus Vm FFT 1–5 Hz integral (r = −0.31, P = 0.20; n = 19 cells). (e) There was no correlation between the sensitivity index d' and prestimulus Vm FFT 1–5 Hz integral (r = 0.14, P = 0.58, n = 19 cells). The sensitivity index, d', from signal detection theory was computed as d' = z(hit rate) – z(false alarm rate), with the z scores computed in Excel (Microsoft) using the function NORMSINV. f, Vrev was significantly correlated with prestimulus Vm FFT 1–5 Hz integral (r = 0.59, P = 0.008; n = 19 cells).

(a) Grand average Vm trace (upper traces) for trained (n = 19 cells, hit trials only) and naive mice (n = 10 cells), with corresponding AP PSTH (lower panel, trained, n = 31 cells; naive, n = 17 cells). (b) Prestimulus Vm (left) and prestimulus AP rate (right) in hit trials of trained mice and naive mice (Prestimulus Vm, trained, n = 19 cells; naive, n = 10 cells; AP rate, trained, n = 31 cells; naive, n = 17 cells). (c) PSP amplitude (left) and AP rate (right) of the early sensory response in hit trials of trained mice and naive mice quantified over the first 50 ms post whisker stimulus (PSP amplitude, trained, n = 19 cells; naive, n = 10 cells, P = 0.03, Student's unpaired t test; AP rate, trained, n = 31 cells; naive, n = 17 cells, P = 0.002, Wilcoxon-Mann-Whitney test). (d) Absolute mean Vm (left) and AP rate (right) of the secondary late sensory response in hit trials of trained mice and naive mice, quantified as the average value for each cell between 50 ms and 400 ms after the whisker stimulus (Vm, trained, n = 19 cells; naive, n = 10 cells, P = 0.007, Student's unpaired t test; AP rate, trained, n = 31 cells; naive, n = 17 cells). Each light grey circle represents the data from a single recording. Statistical significance is indicated by * for P < 0.05, and ** for P < 0.01. Average data shown as mean ± sem.

(a) Prestimulus Vm was not different comparing hit trials and miss trials (n = 19 cells). (b) Prestimulus AP rate was not different comparing hit trials and miss trials (n = 31 cells). (c) PSP amplitude of the early sensory response was not different comparing hit trials and miss trials (n = 19 cells). Vrev was also not different in hit and miss trials either (P = 0.30, Vrev hit, −48.5 ± 1.2 mV, n = 15 cells; Vrev miss, −47.7 ± 2.4 mV, n = 13 cells, Student's unpaired t test, data not shown). (d) APs during the early sensory response quantified within 50 ms following whisker stimulation was not different comparing hit trials and miss trials (n = 31 cells). (e) Absolute mean Vm of the secondary late response was significantly depolarized in hit trials compared to miss trials, quantified as the average Vm between 50 ms and 400 ms after the whisker stimulus (n = 19 cells, P = 0.03, Student's paired t test). (f) AP rate during the secondary late response was significantly enhanced in hit trials compared to miss trials, quantified over the same duration as in e (n = 31 cells, P = 0.04, Wilcoxon Signed Rank paired test). Each grey line connects hit and miss data points from the same recording. Statistical significance is indicated by * for P < 0.05. Average data shown as mean ± sem.

(a) Example hit trials before and after transection of the facial nerve (FN, the motor nerve controlling whisker movements). Note the absence of whisker movements following FN transection, (b) FN transection had no significant effect on lick probability after whisker stimulus or during catch trials in trained mice (n = 4 mice). (c) Average Vm traces of hit and miss trials from an example whole-cell recording carried out after FN transection. The late depolarization is larger on hit trials compared to miss trials. (d) Grand average Vm traces of hit and miss trials (n = 6 cells). Similar to control mice, in mice with transected facial nerves, the late depolarization is larger on hit trials compared to miss trials. (e) Absolute mean Vm of the secondary late response in hit trials and miss trials quantified as the average value between 50 ms and 400 ms after the whisker stimulus (n = 6 cells, P = 0.02, Student's paired t test). Statistical significance is indicated by * for P < 0.05. Each grey line connects hit and miss data points from the same recording. Average data shown as mean ± sem.

(a) ChR2-YFP (green) was expressed in PV expressing neurons by injecting a Cre-dependent AAV vector into the C2 barrel column of S1 barrel cortex in PV-Cre mice (left), which were then trained to detect C2 whisker stimuli. Coronal section through the center of the ChR2-expression site (green) indicating localized expression (right). Red fluorescence shows tdTomato expression in PV neurons (PV-Cre mice were crossed to LoxP-Stop-LoxP-tdTomato reporter mice). (a) Activation of ChR2 in PV-expressing neurons in forepaw S1 during the early or the late phase of the sensory response did not affect detection performance in trained mice (n = 3 mice).