Temporal divergences emerge between barrel and septa domains upon repeated whisker stimulation in Layer 4 of mouse wS1.

(A) Either principal whiskers (SWS, or single whisker) or most of the principal whiskers (MWS, or multi whisker) were stimulated for 2s at 10Hz (20 pulses) for 20 times (20 trials). (B) Acute simultaneous in-vivo silicon probe (8×8) recordings were performed from barrel and septa domains after the start of active whisking (>P21) (C) t-SNE analysis of the average firing dynamics from three domains after averaging 20 single trials. (D) Left: Average multi-unit firing profiles recorded at Layer 4 of barrel (black), unstimulated neighbor (blue) and septa (pink) columns upon 2s-long 10Hz repeated single whisker stimulation. Middle: Matrix representation of the data. Right: Pair-wise statistical comparisons of first 50ms of the responses (yellow line represents the cut off) of three conditions for all pulses (Mann-Whitney-U-Test). (E) Same with E but response profiles upon MW stimulation. (F) Top: Temporal ratio dynamics (average AUC MW responses divided by average AUC SW responses) for barrel and septa. Bottom: statistical comparison of average AUC ratios of all pulses (1-20), first six (1-6) or last 14 pulses (7-20), respectively. (p values are color coded in D, stars represents *: p<0.05; **: p<0.01; ***: p<0.001 in E and F.)

SST+ and VIP+ neuron densities differ at Barrel and Septa Domains in Layer 4.

(A) A passive CLARITY-based tissue clearing protocol was performed on the collected brains which were subsequently imaged in their entirety using a custom-built light-sheet microscope (mesoSPIM). 561nm wavelength was used to image tdTomato, 488nm wavelength was used to image autofluorescence. (B) 3D reconstruction of cleared SST-Ai14 (left) and VIP-Ai14 (right) barrel cortex. X-Y, X-Z and Y-Z views are presented focusing with a single barrel max projection. The brain was rotated to achieve the orientation that the barrel columns are perpendicular to the X-Y view. The scale bars are set for 500μm. (C) Max projection examples of SST (red) and VIP (blue) neurons in wS1 L2-3, L4 and L5. Quantified barrels (yellow lines) and septa (green lines) are circled out which are perpendicular to X-Y (top-down) view. The scale bar is set for 200μm. (D) SST+ neuron density distribution over laminae for barrel and septal domains (N=4 mice, n=54 Barrels and surrounding septa per barrel). Normalized cell densities plotted as a function of depth. 0µm represents the top of layer 4. SST+ interneuron density is significantly higher in septa (p=0.0015, F=30.64, repeated-measures ANOVA). (E) VIP+ neuron density distribution over laminae for barrel and septal domains. (N=4 mice, n=53 Barrels and surrounding septa per barrel). Normalized cell densities plotted as a function of depth. 0µm represents the top of layer 4. VIP+ interneuron density is also significantly higher in septa (p=0.021, F=9.63, repeated-measures ANOVA).

Single- vs multi- whisker responses of SST+ and VIP+ Interneurons.

(A) Interneurons are labeled with tdTomato using reporter mouse lines and Ca2+ imaging has been performed after bulk loading of OGB-1. scale bar 35μm. (B) Schematic representation of the single- and multi- whisker stimulation protocol upon 10Hz stimulation. (C) ΔF/F traces of evoked activity of SST (on top) and VIP (on bottom) interneurons (N = 3 animals per group, VIP P21+: 138 cells, SST P21+: 51 cells). (D) Area under the curve (AUC) of the first 2s of the ΔF/F traces of all SST and VIP cells upon single- (SW: 2s-long principal whisker evoked activity) and multi- whisker (MW: 2s-long multi whisker stimulation evoked activity) stimulation (Wilcoxon signed-rank test, Base: 2s-long baseline activity).

Loss of Elfn1 abolishes barrel-septa response divergences upon single-whisker stimulation.

(A) t-SNE analysis of the average firing dynamics from three domains. (B) Left: Average multi-unit firing profiles recorded at Layer 4 of barrel (gray), unstimulated neighbor (light blue) and septa (purple) columns (11-barrel columns, 7 septal columns, 11 unstimulated neighbors from N=4 mice) upon 2s-long 10Hz repeated single whisker stimulation. Middle: Matrix representation of the same data by representing each stimulation pulse response on one row. Bottom: Pair-wise statistical comparisons of three conditions for all pulses (Mann-Whitney-U-Test). (C) Same with but response profiles upon MW stimulation. (D) Top: Ratio of MW/SW dynamics for barrel and septa. Bottom: statistical comparison of average ratios of 20 average of all pulses, first six or last 14 pulses, respectively. (p values are color coded in B, stars represents *: p<0.05; **: p<0.01; ***: p<0.001 in C and D.)

Accumulative decoding analysis shows an alteration of columnar domain identity in Elfn1 KO animals.

(A) For the decoding analysis, the entire pulse response, only the first 50ms or only the last 45ms has been used. (B) Decoder analysis of the full pulse response profiles in L4 upon SWS. (C) Decoder analysis of the first part of the pulse response (1-50ms) profiles in L4 upon SWS. (D) Decoder analysis of the second part of the pulse response (51-95ms) profiles in L4 upon SWS. (E) Decoder analysis of the full pulse response profiles in L4 upon MWS. (F) Decoder analysis of the first part of the pulse response (1-50ms) profiles in L4 upon MWS (G) Decoder analysis of the second part of the pulse response (51-95ms) profiles in L4 upon MWS.

wS1 barrel and septa columns differentially project to wS2 and M1.

(A) The schematics of the experimental protocol. The retro-AAV-CAG-GFP was injected in either S2, or M1 at P20-30 and the brains were dissected 4 weeks after the virus injection following the tissue clearing and whole brain imaging. (B) 3D reconstruction of the cleared brain with S2(Left panels) or M1(right panels) injection. X-Y, X-Z and Y-Z views are presented focus with a single barrel max projection. The brain was rotated to achieve the orientation that the septa is perpendicular to the X-Y view (top-down). Scale bars in whole brain view are set for 2000um and the scale bars in X-Y, X-Z and Y-Z views are set for 200um. (C) Max projection examples of S2-projected wS1 neurons and M1-projected wS1 neurons in L2-3, L4 and L5 in barrel vs septa. Quantified barrels (yellow lines) and septa (blue lines) are circled out which are perpendicular to X-Y (top-down) view. Scale bars are set for 200um. (D) Left: Normalized cell density profiles of S2 projection neurons at barrel and septa columns. 0µm represents the top of layer 4. L4 and L5a septa columns send significantly more projections to S2 than barrel column (N=4 mice) (p=0.0139, F=11.79, repeated-measures ANOVA). Right: Normalized cell density profiles of M1 projections neurons at barrel and septa columns. 0µm represents the top of layer 4. L3 and L4 septa columns send significantly more projections to M1 (N=6 mice) (p=0.0032, F=14.82, repeated measures ANOVA). (E) Left: Normalized cell density profiles of S2 and M1 projection neurons at barrel columns. 0µm represents the top of L4. L3 barrel columns send significantly more projections to S2 (N=4 mice with S2 injections, N=6 mice with M1 injections) (p=0.0293, F=7.02, repeated-measures ANOVA). Right: Normalized cell density profiles of S2 and M1 projections neurons at septa columns. 0µm represents the top of L4. L5a septal columns send significantly more projections to M1 N=4 mice with S2 injections, N=6 mice with M1 injections) (p=0.0402, F=5.98, repeated measures ANOVA).

An example of assignment of barrel, septa and Neighbor electrodes using histology and current source density (CSD) analysis.

(A) Shank electrode map of 8×8 silicon probes. (B) Probe locations were labeled (DiI) after each experiment using histology (vGlut2 staining). Matching shank locations are indicated using yellow numbers. Blue dots represent the insertion sites. The red and white text depicts corresponding barrel names (C) All principal whiskers are stimulated one-by-one. C1 to C5 barrel responses upon single whisker stimulation are shown from top to bottom. (D) A closer look on C1, C2 and C3 whisker stimulations and corresponding CSD at shanks 5,6,7 and 8. (E), Barrel (B), adjacent septa (S) and the adjacent neighboring barrel (N) were assigned according to the probe insertion sites and L4 barrel autofluorescence in combination with CSD responses. For example, while shanks 6 and 8 show strong response upon single whisker stimulation (and hence assigned as barrels), shank 7 does not show the expected response (hence assigned as septa). Only the triplets of electrodes were used in the analysis in which principal barrel (B), adjacent septa (S) and the adjacent neighboring barrel (N) were captured by the probe insertion sites for any experiment (Left). 14-barrel columns, 9 septal columns, 18 unstimulated neighboring barrels from N=4 mice. (F) An ideal schematic representation (on the left). Realistically, we cannot always be sure if an entire shank is consistently going through a barrel or septal column due to the curvature of the cortex (on the right) for above and below of layer4.

Representation of the conversion of the two-second-long single trial recordings into 20×100ms matrices.

(A) Schematic representation. (B) An example data. The last 5ms of each pulse have been discarded in further analysis.

(A) Left: Average multi-unit firing profiles recorded at Layer 2/3 of barrel (black), unstimulated neighbor (blue) and septa (pink) columns upon 2s-long 10Hz repeated single whisker stimulation. Right: Another representation of the same data but converted into an image by representing each stimulation pulse response on one row. (B) Left: Average multi-unit firing profiles recorded at Layer 2/3 of barrel (black), unstimulated neighbor (blue) and septa (pink) columns upon 2s-long 10Hz repeated multi whisker stimulation. Right: Another representation of the same data but converted into an image by representing each stimulation pulse response on one row. (C) Left: Temporal ratio dynamics (average AUC MW responses divided by average AUC SW responses) for barrel and septa for layer2/3. Right: statistical comparison of average AUC ratios of all pulses (1-20), first six (1-6) or last 14 pulses (7-20), respectively (stars represents *: p<0.05; **: p<0.01; ***: p<0.001 in F and G. Wilcoxon rank sum test).