Mice whisked against surfaces of jointly varying stickiness, coarseness, and position, with simultaneous stereo videography and trigeminal ganglion electrophysiology.

A) Experimental apparatus, viewed from the perspective of the camera. Some structural elements are excluded from the schematic for ease of visualization. B) Texture presentation sequence schematic. Vertical gray bars indicate trials; blue and red drops indicate times of water and sucrose solution dispensation, respectively. C) Unfiltered voltage trace showing spikes from an isolated whisker touch responsive unit in the trigeminal ganglion. D, E) Examples of top-down (D) and rear (E) views of whiskers and surface. F) Complete set of textured tiles used, grouped by material and ordered within material groups from coarse to fine spatial frequency (left to right, top to bottom). G) Grating waveforms used for 3D printer fabricated (resin and silicone, left) and CNC milled (aluminum and Delvin, right) tiles. H) Example tile positions in a single session (left and top right) and distributions of tile positions presented to all C2 whiskers in the dataset (bottom right).

Whisker shape, strain, and kinematics were extracted from 3D reconstructions.

A) 3D reconstruction of whiskers C1-C4 and the textured surface of a resin tile (gray mesh). Whisker centerlines (gray and black curves), estimated base positions (blue dots), mid-shaft reference points (orange dots), and tips or whisker–surface contact points (yellow dots) are shown. Whisker centerlines end at their intersections with the face silhouette masks (not displayed). B, C) Reprojection of whiskers and points in A into original top-down (B) and rear (C) views. D) Curvature and torsion evaluated along the full whisker shaft for the example frame in A. The measured curvature and torsion and the estimated noncontact curvature and torsion given the whisker’s position in this frame are used to compute the curvature and torsion deformations at each point. Gray shading indicates the region over which the change in torsion Δτ is integrated to obtain the total twisting Δτtotal; vertical lines indicate the arclengths of the estimated whisker base (zero), the whisker’s intersection with the facemask, the fixed mid-shaft reference point, and the contact point. E) Computation of whisker tip kinematics. Whisker C2 is displayed at times t and t + dt; for visualization, dt is 80 ms rather than the true frame period of 0.25 ms. The angle between the base-to-tip vector rc(t) (solid purple vector) and rc(t+dt) (dashed purple vector) is θc.

Whisking against silicone sampled a broader region of the whisker strain space than whisking against less sticky surfaces, reaching greater magnitudes of strain.

A) Trajectories through bending, twisting, and roll space of five example whisks (colors) against each material. Opaque lines are smoothed data; transparent lines are unsmoothed. B) Example whisker deformation distributions for whisking against each material. C) Scatter plots of the 50% highest density region volumes of all whisker deformation distributions, grouped by material. Each point corresponds to a single whisker instance observed under a single experimental condition (a unique combination of surface material, grating frequency, and position). D) Mean whisker deformation values for all whisker instances in C, color coded by material.

Comparison of the whisker deformation distribution’s HDR volume and mean between material groups.

Whisking against silicone evoked higher-amplitude stick-slips than whisking against less sticky materials, but lowered stick-slip rates.

A-D) Example slips against all materials presented. Upper panels: reconstructions of the whisker and surface at slip onset (black curve), over the duration of the slip (white to blue curves), and over the 2 ms preceding and following the slip (gray curves). Colored meshes are textured tile surfaces. View is zoomed in to the portion of the whisker from tip to reference point. Reference point (orange dot) and contact point (yellow dot) are shown for each frame. Lower panels: whisker tip acceleration (black trace) and speed (gray trace) profiles for the period from 2 ms before to 2 ms after the slip onset. Cyan shading indicates the region in which the slip peak acceleration is sought; red triangles mark the identified peak acceleration. E, F) Median acceleration (E) and speed (F) profiles of slips by the example C2 whisker in A-D, grouped by material. G, H) Scatter plots of mean slip amplitude (G) and rate (H), grouped by material. As in Figure 3C, each point corresponds to a single whisker instance observed under a single experimental condition.

Comparison of mean slip amplitude and rate between material groups.

Some whisker follicle innervating units responded preferentially to whisking against silicone.

A - C) Cumulative distributions of whisk-average spike rates for all units. Whisks included are restricted to those that contain whisking with contact. D) AUC for discrimination of whisks against silicone from whisks against other materials, based on whisk-average spike rate for each unit. Error bars indicate 95% CI. E) Performance for each unit of an ensemble of regression trees fitted to predict single-unit spike rate from smoothed bending, twisting, roll, and their first derivatives, quantified in terms of correlation between predicted and measured spike rate (r) and variance explained (R2). F) Model performance from panel E plotted against AUC for discrimination of whisks against silicone from panel D.

Samples were evenly distributed across experimental conditions, with the exception of Delrin as least-encountered material.

A, B) Samples (frame-wise observations of single whisker instances) observed in each texture state, pooled across all sessions in which the original low-relief gratings (A) and the supplementary high-relief gratings (B) were used. “Included” samples are those passing the three inclusion criteria of surface contact, active whisking, and absence of outlier whisker strain or kinematics.

Information about whisker shape and motion during whisking against surfaces is lost when tracking is limited to a top-down view.

A) Torsion in the shaft of a single example C2 whisker (left) and the full set of C2 whiskers (right), without contact or whisking. B) Torsion in the shaft of a single example C2 whisker over a 25 ms period of contact and whisking (with 5x temporal downsampling for visualization). C) Example 2D and 3D whisker shape signals measured during a 0.1 s period of contact and whisking. D) Median amplitudes of the whisker shape signals for all whisker instances. E) Example 2D and 3D whisker motion signals measured during the same period of contact and whisking as in C. F) Median amplitudes of the whisker motion signals for all whisker instances. G) Uncertainty coefficients between 3D whisker shape and motion signals and their 2D comparators. H) Position of the whisker tip over the same time period as in C and E, at 2x temporal downsampling.

Highest density regions of the whisker deformation distribution.

A) Gray volumes indicate the boundaries of the 50%, 15%, and 5% highest density regions of the example distributions shown in Figure 3B. B) Scatter plots of the 50% highest density region volumes of all whisker instances’ deformation distributions; points are grouped by position, subgrouped by material, and color coded by frequency. Each point corresponds to a single whisker instance that was observed under the conditions indicated.

Stick-slip event boundaries were extracted from the kinematics of the whisker tip and reference point.

Whisker tip acceleration (top), whisker tip jerk (middle), and whisker tip and reference point speed (bottom) traces over a random interval containing multiple slips. Filled and unfilled triangles above each panel indicate the start and end times, respectively, of the detected stick-slip events. Filled (all panels) and unfilled (bottom only) triangles to the right of each panel mark the high and low thresholds, respectively, for the plotted variables. Shading indicates frames classed as high acceleration, high jerk, sliding, or sticking, based on threshold crossings. Inset at right zooms in on two slips.

Relationships of stick-slip amplitude and rate to spatial frequency were weak in both high and low grating relief conditions.

A, B) Scatter plots of mean slip amplitude (A) and rate (B), grouped by frequency, for the low-relief set of gratings. C, D) Scatter plots of mean slip amplitude (C) and rate (D), grouped by frequency, for the high-relief set of gratings. As in Figures 3 and 4, each point corresponds to a single whisker instance observed under a single experimental condition.

Whisks against silicone were discriminable from other materials on the basis of single-unit response.

Heatmap displays, for each unit, the AUC for discrimination of silicone from each other material the unit encountered.

Computation of whisker shape, strain, and orientation.

A) Contact-induced whisker roll was defined as the angle between the instantaneous plane of curvature of the proximal 60% of the whisker shaft (orange), and the estimated noncontact plane of curvature of the proximal 60% of the whisker shaft (purple). B) The whisker’s estimated noncontact plane of curvature defined the directions of the y and z basis vectors of the whisker-centered coordinate frame. C) The whisker curve’s tangent, normal, binormal (TNB) frame at each point along its length was computed from the curve and its first three derivatives. The curve’s shape (curvature and torsion) was the angular velocity of the TNB frame with respect to normalized arclength. D-E) The curvature (D) and torsion (E) of the undeflected whisker, given arclength along the whisker and the position (azimuth and elevation angles) of the whisker base, were estimated as fifth-order polynomial surfaces (cyan meshes) fitted to the average values observed during low-acceleration whisking without contact (red dots). Elevation angle is limited to a single value for visualization. F) The angle between the undeflected proximal whisker segment’s plane of curvature and the world x-y plane, given the whisker’s position, was estimated as a quadratic surface (cyan mesh) fitted to the average values observed during low-acceleration whisking without contact (red dots).