Dye-enhanced visualization of rat whiskers for behavioral studies

Version of Record

Accepted for publication after peer review and revision.

Download
Cite
Share
CommentOpen annotations (there are currently 0 annotations on this page).

Version of Record updated: March 13, 2018 (This version)
Version of Record updated: November 3, 2017 (Go to version)
Version of Record published: July 14, 2017 (Go to version)
Accepted Manuscript published: June 14, 2017 (Go to version)
Accepted: June 13, 2017
Received: January 20, 2017

1. Of interest
Using light and X-ray scattering to untangle complex neuronal orientations and validate diffusion MRI

Miriam Menzel, David Gräßel ... Marios Georgiadis

Research Article Updated Jun 6, 2023
Further reading

Article
Figures and data
Abstract
Introduction
Results and discussion
Materials and methods
Data availability
References
Decision letter
Author response
Article and author information
Metrics

Abstract

Visualization and tracking of the facial whiskers is required in an increasing number of rodent studies. Although many approaches have been employed, only high-speed videography has proven adequate for measuring whisker motion and deformation during interaction with an object. However, whisker visualization and tracking is challenging for multiple reasons, primary among them the low contrast of the whisker against its background. Here, we demonstrate a fluorescent dye method suitable for visualization of one or more rat whiskers. The process makes the dyed whisker(s) easily visible against a dark background. The coloring does not influence the behavioral performance of rats trained on a vibrissal vibrotactile discrimination task, nor does it affect the whiskers’ mechanical properties.

https://doi.org/10.7554/eLife.25290.001

Introduction

Many nocturnal or crepuscular rodents, including rats and mice, have long vibrissae that enable tactile exploration of the nearby environment (Grant et al., 2009; Carvell and Simons, 1990; Gustafson and Felbain-Keramidas, 1977; Vincent, 1912). Rats contact and palpate objects by projecting their facial vibrissae in front of and around the snout, giving rise to neuronal representations of what is being contacted and where it is located (Diamond et al., 2008; Diamond and Arabzadeh, 2013).

A rich variety of whisker follicle mechanoreceptors (Ebara et al., 2002) encode whisker deflections with remarkable sensitivity and temporal resolution (Bale et al., 2015; Whiteley et al., 2015; Ikeda et al., 2014). Signals follow afferent pathways to reach the primary somatosensory, secondary somatosensory and the primary motor cortex (Ahissar et al., 2000; Kleinfeld et al., 2006; Yu et al., 2015). Quantifying the relationship between animal’s sensing of the environment and neuronal firing in these subcortical and cortical stages requires precise monitoring of whisker shape and motion during behavioral tasks.

The literature puts forward various strategies to estimate whisker movement. The electrical activity of facial muscles provides information about motor output commands, and correlates with whisking activity (Hill et al., 2008), but the technique includes an invasive implant and does not directly specify whisker position and deflection. Another technique is to image a high-contrast particle fixed to the whisker (Harvey et al., 2001; Venkatraman et al., 2009); however, the technique fails to track whisker shape and it likely perturbs whisker dynamics due to the particle mass and air resistance. Whisker position can be estimated by its intersection with a laser sheet (Jadhav et al., 2009), but this necessitates trimming the unwanted whiskers; moreover, whisker shape is lost. High-speed videography holds the greatest potential for noninvasive measurement of whisker movements, deflections, and shape (O’'Connor et al., 2010; Voigts et al., 2008; Ritt et al., 2008; Knutsen et al., 2005; Perkon et al., 2011; Clack et al., 2012). Videography requires heavy computations to distinguish the target whiskers from the background, and sometimes requires trimming, which may limit behavior. Image processing for whisker tracking must be tailored to the specific lighting conditions. Unless the whisker is uniformly illuminated and reflects the light homogeneously (to avoid optical aberration), image quality will be degraded.

A higher degree of luminance contrast could solve many of the technical challenges inherent to videography. Identified single whiskers could be followed across sessions, and whisker-cross imaging issues could be abated. Here, we describe the application of a fluorescent dye to highlight specific sets of whiskers. The advantages of fluorescence include the conversion of each whisker into a homogeneous source of light, as well as the facilitation of multi-whisker tracking by color segmentation. However, the application of a fluorescent dye may affect the mass and the stiffness of the whisker, hence the rat’s tactile perception during a behavioral task. In order to assess fluorescence as a potential tool for whisker visualization, we test performance in a vibrissal vibrotactile discrimination task (Fassihi et al., 2014). After finding performance unhindered, we confirm that the conserved tactile perception after application of the fluorescent dye arises from the fully conserved dynamic response of the whiskers.

Results and discussion

Dye-enhanced visualization

We applied two fluorescent dyes on two whiskers of a rat during brief sedation (green and red dyes - StarGazer). With a macro lens, under white light and with no optical filter, all whiskers were visible (Figure 1A). Despite its high quality, tracking multiple whiskers from this type of image would be challenging. Under blue light, with a long-pass filter (FEL0500 Long-pass filter cut-on wavelength 500 nm - Thorlabs) in front of the lens to remove blue reflections, the whiskers were visible in two colors and with enhanced contrast (Figure 1B). They would be readily separable by masking specific color ranges. Contrast can be further increased by a more red-shifted long-pass filter (red plexiglass, cut-on wavelength 625 nm measured with Implen NanoPhotometer), selecting only one dye (Figure 1C). Video recordings confirm the increased visibility of the dyed whiskers in the mobile rat (Video 1), facilitating tracking (Supplementary file 1 and Video 2). We performed whisker tracking using JFilament (Smith et al., 2010), a plugin of ImageJ (Schneider et al., 2012), a public domain Java-based image processing program. JFilament can segment and track 2D/3D filaments and has been widely used for fluorescence microscopy images. It allows tracking of these elements through a sequence of frames and is thus suitable for whisker tracking.

Figure 1

Download asset Open asset

Dye-enhanced whisker visualization and effect on behavior.

(A) Vibrissal pad with two whiskers dyed, under white light. (B) Same as (A) but under blue light and long-pass optical filter. (C) Same as (B) but with red plexiglass. (D) Effect of vibrissal pad coloring (pink, olive) and sham treatment (violet, jade) on daily performance of a vibrissal vibration discrimination task. Chance performance would be 50%. (E) Violin plots of difference in performance before and after treatment. The distributions are generated by permuting every data point before with every data point after.

https://doi.org/10.7554/eLife.25290.002

Video 1

Download asset

posterframe for video — Proof of concept of the method.
https://doi.org/10.7554/eLife.25290.003

Video 2

Download asset

Behavioral test

We examined four rats, well-trained in a whisker-mediated vibrotactile delayed comparison task (Fassihi et al., 2014), for behavioral effects of dye application. Under sedation, two underwent the application of fluorescent dye while two were sham controls (saline applied). To increase the likelihood of detecting any adverse effect of dyeing, the entire vibrissal pad on the side used in the vibrotactile task was colored. The next day the rats were returned to the behavioral apparatus for testing. Figure 1D shows daily performance and reveals that the application of fluorescent dye, like the sham procedure, had no discernible effect on behavior. Figure 1E summarizes the performance, showing no difference between before versus after the dye application.

Quantification of the frequency dynamic response of the whiskers

The behavioral tests support the notion that dye application did not affect the dynamic mechanical properties of the whiskers, but an alternative must be entertained: rats might have successfully compensated for altered whisker properties, just as they compensate for interventions as drastic as whisker clipping (Zuo et al., 2011). To exclude this possibility, we explored whiskers by theory – exploiting Euler-Bernoulli beam theory to perform computations by Finite Element Method (FEM) – and by experiment – measuring whisker linear dynamic response.

We simulated the frequency response function (FRF) $H$ that relates the imposed vertical displacement at the clamped base of a D2 whisker with the output displacement at any location x along its longitudinal axis (Figure 2A, Materials and methods). We also used high-speed videography (Figure 2B) to measure an extracted D2 whisker’s dynamic response to white noise vibration and plotted the magnitude of the FRF (mFRF) $| H |$ along the longitudinal axis of the whisker (Figure 2C). The real and simulated mFRF are similar (Figure 2A vs 2C), confirming previous claims that a real rat whisker behaves as a tapered beam. The mFRF constitutes a fingerprint of a whisker and thus it is a simple measure of the similarity between pairs of whiskers and the similarity of one whisker over time.

Figure 2

Download asset Open asset

Whisker modeling and testing.

(A) Log mFRF of the whisker approximated as a tapered beam by FEM model. Three arrowheads indicate the first three modal frequencies. (B) Experimental setup. The blue device at the top is the custom-made orientable lamp mounting. Below is the whisker mounted on a custom-made holder attached to the loudspeaker. To the left is a sample of the shaker input signal, which was estimated by tracking the base of the whisker and has been used as input for the computation of the mFRFs (vertical line corresponds to ± 1 standard deviation, horizontal line corresponds to 200 ms of signal, which has been also zoomed). To the right, image processing is summarized. The image is depicted from the camera’s point of view, with the angle selected so as to maximize the visible motion-induced displacement. From the thresholded raw image, points belonging to the whisker are identified in the region of interest and used to compute n equidistant samples along the whisker deflection using a B-spline model. (C) Log mFRF observed for the whisker modeled in (A). Three arrowheads indicate the first three modal frequencies, perfectly aligned with those of the simulation.

https://doi.org/10.7554/eLife.25290.005

To outline the similarity index, mFRFs of whiskers designated W₁ and W₂ (similar in length, 53 and 56 mm) and W₃ (much shorter, 23 mm) are illustrated (Figure 3A,B,C, respectively). The overall similarity between any two whiskers’ dynamic responses is summarized as a scatter plot (Figure 3D), where each point corresponds to the two whiskers’ spectral values for a given [frequency, position] coordinate. In Figure 3D, we compare two matrices (the mFRFs) in order to test their similarity. In particular, we unwrapped these 2D matrices into two 1D vectors (i.e. we vectorized the matrices) and we computed the coefficient of determination R² of a linear regression between them. These latter values are summarized in Figure 4 for each pair of whiskers.

Figure 3

Download asset Open asset

Comparison of whisker dynamic responses.

(A) Log mFRF of whisker W₁. (B) Log mFRF of whisker W₂. (C) Log mFRF of whisker W₃. (D) Each scatter plot compares a pair of transfer functions, where each dot gives mFRF for a given frequency, position of both whiskers. On left panel W₁ versus W₂; on right panel W₁ versus W₃. The dot density has been encoded by coloring them from blue (single dot) to red.

https://doi.org/10.7554/eLife.25290.006

Figure 4

Download asset Open asset

mFRF comparisons across whiskers and time.

(A) This and the next four panels depict matrices of similarity between mFRFs in pairwise (row against column) combinations. This first matrix illustrates similarities before dye application. (B) Matrix comparing the dyed and undyed conditions. (C) Matrix comparing whiskers after dye application. (D) Matrix computed for FEM-simulated (Sim) whiskers. (E) Matrix comparing one whisker over time. Time measures begin with ‘cut’ (immediately after whisker extraction from the snout), ‘dye’ (immediately after dye application) and ‘pol’ (immediately after the application and drying of nail polish). (F) In the upper panel, the diagonal values of the matrix shown in (B) are plotted (each whisker against itself, before and after dyeing). The set of values obtained by taking the i-th sub- and super-diagonal values of the matrix shown in (B) are averaged and represented with horizontal lines labeled $θ$ _i. The further from the main diagonal, the smaller is $θ$ _i, indicating that mFRF similarity decreases with increasing disparity in whisker length. The values of the diagonal are always high, highlighting the absence of dye effect on all samples: despite the application of the dye, each whisker is always more similar to itself rather than to other whiskers. On the lower panel, the supra-diagonal values of the comparison matrix shown in (E), which highlights stability up to the application of polish. In particular, by reference to the horizontal lines shown in the upper panel of (F), the change of the whisker caused by polish can be compared to the difference between the two whiskers of length 18 mm and 56 mm.

https://doi.org/10.7554/eLife.25290.007

If two mFRFs were identical, all points would lie along the unity line and the coefficient of determination of the linear regression (R²) would be 1. For W₁ vs W₂R² = 0.91, and for W₁ vs W₃R² = 0.43. Thus, as predicted by beam theory, a difference in whisker length corresponds to a difference in dynamic response.

Dye effects on mechanical properties

Confident that the mFRF constitutes a robust measure of whisker dynamics, we measured a large set of extracted whiskers and quantified the similarity between any two mFRFs by the R² measure. First, we compared the dynamic responses of all whiskers, undyed, in all pairwise combinations (Figure 4A). The whiskers are labeled and ordered by length. Next, we compared the same set of whiskers, immediately before and after dye application (Figure 4B). Then, we compared the dynamic responses of all dyed whiskers, in all pairwise combinations (Figure 4C). The plots all attest to the stability of mFRFs in spite of dye application.

As predicted by beam theory, the mFRF is sensitive to geometry (length, cross-section area along the shaft) and to material parameters (Young’s modulus, damping coefficients, mass density). We modeled the set of whiskers and computed the similarity between simulated mFRFs (Figure 4D). The FEM simulations (Materials and methods) uncover the same three groups of elevated mFRF similarity (orange-red blocks near the diagonal, against yellow-green-blue off diagonal) found in empirical data (Figure 4A–C), confirming the validity of the beam model in describing whisker dynamic response.

As a final test of stability, we examined a single whisker at many time points before and after dyeing (Figure 4E). Finally, to verify the sensitivity of the mFRF method to a manipulation that changes mass and stiffness, we tested the whisker after application of nail polish. The similarity matrix depicts all pairwise mFRF comparisons. Neither the passage of time nor the application of the fluorescent dye had a significant effect on dynamic response (R² > 0.8 for all comparisons), whereas polish severely altered the response (R² < 0.3).

Figure 4F summarizes the effects of time and dye application. The upper plot gives R² values of the diagonal of Figure 4B, showing that for the full set of tested whiskers the similarity before/after dye application was >0.8. The lower plot gives R² values of the superdiagonal of Figure 4E, showing that a single whisker maintained a near-constant degree of similarity from the moment of cutting, through the dyed stage, until the application of nail polish.

In vision and audition, the external stimulus provides an adequate descriptor of the information to be processed by the sensory system. For instance, although the eyes move continuously (Ahissar et al., 2016), it is common to discount this motion and to consider the incoming sensory signal to be the image presented on the screen. The equivalence between external stimulus and input to the sensory processing pathway is sufficient because, to a first approximation, the eye does not change form or function as it absorbs the stimulus. The same simplification does not hold for whisker-mediated touch: the mere definition of the stimulus – a texture, a shape, a vibration – provides little information about the signal that actually engages the sensory receptors and enters the nervous system. Instead, there is a stage of pre-processing by the whisker and follicle (Bagdasarian et al., 2013). While the discriminandum itself can be easily quantified, the pre-processing afforded by the whisker can be difficult to measure, especially in freely moving animals.

Over the last 10 years, neuroscientists have begun to focus on sensorimotor processing in behaving rodents (Maravall and Diamond, 2014; Grion et al., 2016; Zuo et al., 2015). A critical step is to visualize the whiskers with clarity. The ideal methodology should be (i) simple and easy to execute, (ii) dependent on materials that are low-cost and readily available, (iii) innocuous for the animal and uninfluential on behavior, (iv) long-lasting, and (v) uninfluential on the dynamic mechanical properties of the whiskers. The method described in the present work satisfies these constraints. Its limitation is the reduced number of vibrissae that can be simultaneously imaged and tracked. To face this challenge, future work could focus on further wavelength separation.

Materials and methods

Sedation for the application of the dye

Request a detailed protocol

Rats that underwent the application of the dye (N = 3 two were used in the behavioral study and one used to shoot the Video 2) and sham-control rats (N = 2) were sedated for about 2 to 2.5 hr by injection of Domitor (Medetomidine hydrochloride, dose: 0.5 mg/kg, i.p.). With the procedure terminated, an injection of Antisedan (Atipamezole hydrochloride, standard dose 0.2 ml) was used to reverse the sedated state.

Behavioral test subjects and setup

Request a detailed protocol

Four male Wistar rats (Harlan Laboratories) were housed individually or with one cage mate and kept on a 14/10 light/dark cycle and reached a weight of 450–550 g. They were examined weekly by a veterinarian. At the beginning of the experiment, they had already been trained to perform a vibrotactile delayed comparison task (Fassihi et al., 2014): they compared the mean speed of two vibrotactile stimuli by placing their whiskers on a 20 × 30 mm plate attached to the diaphragm of a shaker motor (type 4808; Bruel and Kjaer). Performance is defined as the percent of trials in which the rat correctly selected the reward side (left or right), a selection determined by the relative mean speeds of two stochastic vibrations. Double-sided adhesive was fixed to the plate to maintain the contact between the touching whiskers and the plate. Each stimulus was built by sampling a normal distribution in position centered at 0 and then applying a Butterworth low-pass filter with 150 Hz cutoff frequency. Protocols conformed to international norms and were approved by the Italian Health Ministry and the Ethics Committee of the International School for Advanced Studies.

Collection of whisker samples

Request a detailed protocol

One male naive Wistar rat (Harlan Laboratories) was killed by CO₂ overdose in order to extract whisker samples from both vibrissal pads. To collect samples, firstly the vibrissal pad was surgically removed and then the skin was precisely cut around the whisker follicle. In this way, each sample included the follicle that mechanically wedges the hair shaft into the skin.

Application of the fluorescent dye

View detailed protocol

We used the ‘Hair Color Rinse Semi Permanent’ line, in particular ‘UV Red’ and ‘UV Green’ in Figure 1A–C. We used ‘UV Red’ for the measurements of the dynamic responses (Figures 2–4). The ingredients, as detailed by the supplier, consist of a common base mixture to which are added specific cosmetic colorants. The common ingredients are: purified water, cetearyl alcohol, distearoylethyl-hydroxyethylmonium, methosulfate, cetearyl alcohol, ceteareth-20, citric acid, and methylparaben. The specific cosmetic colorants for ‘UV Red’ are: CI 42520, Basic Orange 31, Basic Violet 16, Basic Yellow 40, HC Yellow 4, and CI 16255. The specific cosmetic colorants for ‘UV Green’ are: Basic Yellow 40, CI 42090, CI 61570, HC Yellow 4, and CI 62045.

Semi-permanent hair rinse is designed to diffuse into and bind to the hair (Robbins, 2012). In order to bleach the whiskers before dye application, we applied a gel of 3% hydrogen peroxide mixed with a booster powder (Robbins, 2012, Table 4-3). The chemical bleaching process would be expected to follow a diffusion law, where the diffusivity for the human hair has been estimated as 1.8 × 10⁻⁹ cm²/min (Robbins, 2012, Figure 4-3). Afterwards, we washed the whiskers with distilled water, dried them, and then submerged them for 30 min in the hair color gel (Supplementary file 2). The fluorescent dye method is described more in detail at Bio-protocol (Rigosa et al., 2018).

Quantification of the deposited dye and of the photobleaching

Request a detailed protocol

As an empirical verification, we measured the mass of a set of 34 whiskers (a multiwhisker sample reduces the likelihood of measurement error) by means of a scale (RADWAG AS 220.R2) with an accuracy of 0.1 mg. Prior to processing, the pooled mass was 14.8 mg. No difference in mass was detected after the chemical bleaching, while the coloring produced an increment of 0.1 mg (from 14.8 to 14.9 mg; within the measurement error of the scale), less than 0.7%.

We measured the photobleaching of the dye for 7 hr with all seven of the lamp’s LEDs pointing at the whiskers at a distance of about 15 cm. Bleaching was quantified by integrating pixel by pixel the grayscale values of an area containing colored whiskers (foreground) and another empty area of the same dimension (background). We subtracted the background from the foreground, and we normalized to the initial value in order to obtain a percentage decay (Supplementary file 3). We fit the data with an exponential law, the standard decay function for photobleaching. The measured time constant was about 14 hr (about 50,000 s), equivalent to about 25,000 trials of 2 s.

Computing a mechanical fingerprint of whisker samples

Request a detailed protocol

As the geometry of animal whiskers resembles that of a thin, flexible filament, their dynamic response to mechanical stimuli can be suitably explored by means of beam theory (Birdwell et al., 2007). In particular, one can exploit the mechanical analogy between a whisker and a tapered beam (Quist and Hartmann, 2012; Williams and Kramer, 2010; Hires et al., 2013; Bush et al., 2016). A truncated conical beam model can predict rat whiskers’ natural resonant frequencies (Yan et al., 2013), the so-called normal modes. In this study, the linear beam model of Euler-Bernoulli was employed while accounting for whisker taper. Let $v (x, t)$ denote the beam deflection at position $x$ and time $t$ . Then, by assuming a viscous damping model of the Rayleigh type, the balance laws of linear and angular momentum yield

A (x) (\frac{\partial^{2} v (x, t)}{\partial t^{2}} + α \frac{\partial v (x, t)}{\partial t}) + \frac{E}{ρ} \frac{\partial^{2}}{\partial x^{2}} (J (x) \frac{\partial^{2} v (x, t)}{\partial x^{2}} + β J (x) \frac{\partial^{3} v (x, t)}{\partial t \partial x^{2}}) = 0.

In the equation above, $ρ$ denotes the mass density, $A$ the cross-section area, $J$ the moment of inertia of the cross-section and $E$ the Young’s modulus, whereas $α$ and $β$ are the mass- and stiffness-proportional damping coefficients. The governing Equation (1) is complemented by the boundary conditions that correspond to the experimental setting (Figure 2B), namely

v (0, t) = δ (t), \frac{\partial v (0, t)}{\partial x} = 0, \frac{\partial^{2} v (L, t)}{\partial x^{2}} = 0, \frac{\partial^{3} v (L, t)}{\partial x^{3}} = 0

for $t \in [0, \infty]$ , and by the initial conditions

v (x, 0) = 0, \frac{\partial v (x, 0)}{\partial t} = 0

for $x \in [0, L]$ . According to the boundary conditions of Equation (2), a beam of length $L$ is clamped at its base and subject to the vertical, time-varying displacement $δ (t)$ . As for the whisker taper, we assumed the radius $r$ of the circular cross-section to linearly decrease along the whisker shaft, such that

r (x) = r_{b} - (r_{b} - r_{t}) \frac{x}{L}, A (x) = π r {(x)}^{2}, J (x) = π \frac{r {(x)}^{4}}{4},

where $r_{b}$ and $r_{t}$ denote the base and tip radius, respectively.

For the computation of the beam FRF, it is now expedient to Laplace-transform the set of Equation 1 and 2. By taking into account the initial conditions (3), the governing Equation (1) simplifies into the following linear ODE in the space variable $x$

A (x) (s^{2} + s α) V (x, s) + \frac{E}{ρ} \frac{\partial^{2}}{\partial x^{2}} (J (x) \frac{\partial^{2} V (x, s)}{\partial x^{2}} + s β J (x) \frac{\partial^{2} V (x, s)}{\partial x^{2}}) = 0,

whereas the boundary conditions of Equation (2) yield

V (0, s) = Δ (s), \frac{\partial V (0, s)}{\partial x} = 0, \frac{\partial^{2} V (L, s)}{\partial x^{2}} = 0, \frac{\partial^{3} V (L, s)}{\partial x^{3}} = 0,

where $V (x, s)$ and $Δ (s)$ denote the Laplace transform of $v (x, t)$ and $δ (t)$ , respectively. The whisker model of Equation 5 and 6 was implemented in the commercial, finite element software COMSOL Multiphysics (version 5.2a, RRID:SCR_014767). Whiskers were accurately discretized into a finite element mesh comprising 200 elements. Cubic and quadratic shape functions were employed for the approximation of the displacement and rotation field, respectively. Since the numerical analyses were carried out in the frequency domain, the whisker FRF $H$ was readily computed by taking the ratio of $V (x, s) / Δ (s)$ evaluated at $s = i ω$ , where $i$ is the imaginary unit and $ω$ is the angular frequency. As for the geometric parameters of the biological samples, these were measured with the aid of custom code, which was used to process images from high-speed videos for the whisker length and bright-field micrographs for the base and tip radii. Reference material parameters were taken from the literature (Hartmann et al., 2003) and tuned to fit the experimental results with the numerical computations (Supplementary file 4).

As becomes evident from the mathematical formulation, the dynamic response of rat whiskers is sensitive to their geometric and constitutive parameters. For example, the log mFRF of a simulated D2 whisker is shown in Figure 2A. This reveals the characteristic peaks at the modal resonant frequencies (the first three modes are highlighted by black arrows). Modal shapes can also be inferred from the figure. Given a whisker geometry, these are distributed along the frequency axis depending upon the material parameters. Thus, a change in the modal frequencies distribution can only be explained by a change in the whisker constitutive parameters. Hence, in this work, we propose the dynamic responses comparison as a meaningful approach to determine whether mechanical changes occurred or not as a consequence of the chemical processes needed to apply the fluorescent dye on the whisker.

Ex vivo experimental setup and data processing

Request a detailed protocol

The setup (Figure 2B) included a mounting structure, designed to fix the sample without gluing it, a condition which would have affected the measurements. The whole system was installed on a base that was shaken by a loudspeaker (8 Ohm, 30 W - Visaton) controlled by using a custom-made LabVIEW code (National Instruments, RRID:SCR_014325). The set-point of the control input to the loudspeaker was computed to be a white noise sampled at 2 kHz and low-pass filtered with a Gaussian filter at 1 kHz (Figure 2B). The vibrational apparatus was inserted in a wooden stage completely covered with neoprene to avoid light reflections. A custom-made lamp was placed on the ceiling: it had seven bulbs (ILH-OO01-DEBL-SC211-WIR200, Wavelength 455 nm, Flux @700mA 1400 mW, Radiance angle ±60°; concentrator lens FA11205_Tina-D-OSL FWHM angle ±6°) that could be oriented so that the whisker sample could be homogeneously illuminated. In front of the stage, a high-speed camera (CamRecord 450, Optronis) recorded the sample vibrations at 2000 frames per second. Each frame was analyzed to compute the deflection of each whisker as a function of time (Figure 2B) and consequently estimate the mFRF for each trial of 2.25 s each. The mFRFs were averaged over 10 repetitions of different trials using different white noise signals, because this number of trials was sufficient in order to lower the statistic error of the measurement below the image pixel resolution.

Image processing was performed using a Python (PSF, http://www.python.org) custom code that makes use of the OpenCV library (Bradski, 2008). Firstly, for each frame the part of the image which contained the whisker was identified, then the points belonging to the whisker were revealed by thresholding. Hence, the whisker shape was readily estimated by means of a B-spline modeling of these points. The B-spline returns N equidistant samples along the whisker direction, which model a specific whisker segment at a specific time (Figure 2B). One trial thus provided N time series with a length of 4500 (2.25 s times 2000 frames per second) that portray whisker dynamics as a function of its one-dimensional space. In the linear regime, each segment of the beam is a linear time invariant system (LTI) that has an input defined as the movement of the whisker base (i.e. the segment which is closest to the loudspeaker). Hence, we estimated the mFRF for each whisker’s segment using the Wiener–Khinchin theorem, which states that the autocorrelation function of a wide-sense-stationary random process has a spectral decomposition given by the power spectrum of that process and thus links the cross spectral density functions of the input and the output with the energy transfer function.

Two mFRFs correlate only if they have the same mode shape in the same modal frequencies. We define the coefficient of determination (R²) of their linear regression as non-negative normalized similarity metric: R² equal to 1 corresponds to the identical responses while R² equal to 0 corresponds to completely uncorrelated responses.

Data and code availability

Request a detailed protocol

The code (Rigosa, 2017) includes (1) the image processing from raw video and (2) the postprocessing of preprocessed data from video. A copy is archived at https://github.com/elifesciences-publications/whisker-dynamic-response. Data (Rigosa et al., 2017) includes one raw video to run a demo (~1 GB), the whole preprocessed dataset to reproduce results (~9 GB), a copy of the code for data acquisition and processing.

Data availability

The following data sets were generated

1. Rigosa J
2. Lucantonio A
3. Noselli G
4. Fassihi A
5. Zorzin E
6. Manzino F
7. Pulecchi F
8. Diamond ME
(2017) Data from: Dye-enhanced visualization of rat whiskers for behavioral studies
Available at Dryad Digital Repository under a CC0 Public Domain Dedication.

https://dx.doi.org/10.5061/dryad.9k9n0

References

1. Ahissar E
2. Arieli A
3. Fried M
4. Bonneh Y
(2016) On the possible roles of microsaccades and drifts in visual perception
Vision Research 118:25–30.
https://doi.org/10.1016/j.visres.2014.12.004
- PubMed
- Google Scholar
(2000) Transformation from temporal to rate coding in a somatosensory thalamocortical pathway
Nature 406:302–306.
https://doi.org/10.1038/35018568
- PubMed
- Google Scholar
1. Bagdasarian K
2. Szwed M
3. Knutsen PM
4. Deutsch D
5. Derdikman D
6. Pietr M
7. Simony E
8. Ahissar E
(2013) Pre-neuronal morphological processing of object location by individual whiskers
Nature Neuroscience 16:622–631.
https://doi.org/10.1038/nn.3378
- PubMed
- Google Scholar
(2015) Microsecond-scale timing precision in rodent trigeminal primary afferents
Journal of Neuroscience 35:5935–5940.
https://doi.org/10.1523/JNEUROSCI.3876-14.2015
- PubMed
- Google Scholar
(2007) Biomechanical models for radial distance determination by the rat vibrissal system
Journal of Neurophysiology 98:2439–2455.
https://doi.org/10.1152/jn.00707.2006
- PubMed
- Google Scholar
Book
1. Bradski G
(2008)
The OpenCV Library

Dr. Dobb's Journal of Software Tools.
- Google Scholar
1. Bush NE
2. Schroeder CL
3. Hobbs JA
4. Yang AE
5. Huet LA
6. Solla SA
7. Hartmann MJ
(2016) Decoupling kinematics and mechanics reveals coding properties of trigeminal ganglion neurons in the rat vibrissal system
eLife 5:e13969.
https://doi.org/10.7554/eLife.13969
- PubMed
- Google Scholar
1. Carvell GE
2. Simons DJ
(1990)
Biometric analyses of vibrissal tactile discrimination in the rat

Journal of Neuroscience 10:2638–2648.
- PubMed
- Google Scholar
1. Clack NG
2. O'Connor DH
3. Huber D
4. Petreanu L
5. Hires A
6. Peron S
7. Svoboda K
8. Myers EW
(2012) Automated tracking of whiskers in videos of head fixed rodents
PLoS Computational Biology 8:e1002591.
https://doi.org/10.1371/journal.pcbi.1002591
- PubMed
- Google Scholar
1. Diamond ME
2. Arabzadeh E
(2013) Whisker sensory system - from receptor to decision
Progress in Neurobiology 103:28–40.
https://doi.org/10.1016/j.pneurobio.2012.05.013
- PubMed
- Google Scholar
(2008) 'Where' and 'what' in the whisker sensorimotor system
Nature Reviews Neuroscience 9:601–612.
https://doi.org/10.1038/nrn2411
- PubMed
- Google Scholar
(2002) Similarities and differences in the innervation of mystacial vibrissal follicle-sinus complexes in the rat and cat: a confocal microscopic study
The Journal of Comparative Neurology 449:103–119.
https://doi.org/10.1002/cne.10277
- PubMed
- Google Scholar
(2014) Tactile perception and working memory in rats and humans
PNAS 111:2331–2336.
https://doi.org/10.1073/pnas.1315171111
- PubMed
- Google Scholar
(2009) Active touch sensing in the rat: anticipatory and regulatory control of whisker movements during surface exploration
Journal of Neurophysiology 101:862–874.
https://doi.org/10.1152/jn.90783.2008
- PubMed
- Google Scholar
1. Grion N
2. Akrami A
3. Zuo Y
4. Stella F
5. Diamond ME
(2016) Coherence between rat Sensorimotor System and Hippocampus is enhanced during Tactile Discrimination
PLOS Biology 14:e1002384.
https://doi.org/10.1371/journal.pbio.1002384
- PubMed
- Google Scholar
1. Gustafson JW
2. Felbain-Keramidas SL
(1977) Behavioral and neural approaches to the function of the mystacial vibrissae
Psychological Bulletin 84:477–488.
https://doi.org/10.1037/0033-2909.84.3.477
- PubMed
- Google Scholar
(2003)
Mechanical characteristics of rat vibrissae: resonant frequencies and damping in isolated whiskers and in the awake behaving animal

Journal of Neuroscience 23:6510–6519.
- PubMed
- Google Scholar
(2001) Discriminative whisking in the head-fixed rat: optoelectronic monitoring during tactile detection and discrimination tasks
Somatosensory & Motor Research 18:211–222.
https://doi.org/10.1080/01421590120072204
- PubMed
- Google Scholar
(2008) Biomechanics of the vibrissa motor plant in rat: rhythmic whisking consists of triphasic neuromuscular activity
Journal of Neuroscience 28:3438–3455.
https://doi.org/10.1523/JNEUROSCI.5008-07.2008
- PubMed
- Google Scholar
1. Hires SA
2. Pammer L
3. Svoboda K
4. Golomb D
(2013) Tapered whiskers are required for active tactile sensation
eLife 2:e01350.
https://doi.org/10.7554/eLife.01350
- PubMed
- Google Scholar
1. Ikeda R
2. Cha M
3. Ling J
4. Jia Z
5. Coyle D
6. Gu JG
(2014) Merkel cells transduce and encode tactile stimuli to drive Aβ-afferent impulses
Cell 157:664–675.
https://doi.org/10.1016/j.cell.2014.02.026
- PubMed
- Google Scholar
(2009) Sparse temporal coding of elementary tactile features during active whisker sensation
Nature Neuroscience 12:792–800.
https://doi.org/10.1038/nn.2328
- PubMed
- Google Scholar
(2006) Active sensation: insights from the rodent vibrissa sensorimotor system
Current Opinion in Neurobiology 16:435–444.
https://doi.org/10.1016/j.conb.2006.06.009
- PubMed
- Google Scholar
(2005) Tracking whisker and head movements in unrestrained behaving rodents
Journal of Neurophysiology 93:2294–2301.
https://doi.org/10.1152/jn.00718.2004
- PubMed
- Google Scholar
1. Maravall M
2. Diamond ME
(2014) Algorithms of whisker-mediated touch perception
Current Opinion in Neurobiology 25:176–186.
https://doi.org/10.1016/j.conb.2014.01.014
- PubMed
- Google Scholar
1. O'Connor DH
2. Clack NG
3. Huber D
4. Komiyama T
5. Myers EW
6. Svoboda K
(2010) Vibrissa-based object localization in head-fixed mice
Journal of Neuroscience 30:1947–1967.
https://doi.org/10.1523/JNEUROSCI.3762-09.2010
- PubMed
- Google Scholar
1. Perkon I
2. Kosir A
3. Itskov PM
4. Tasic J
5. Diamond ME
(2011) Unsupervised quantification of whisking and head movement in freely moving rodents
Journal of Neurophysiology 105:1950–1962.
https://doi.org/10.1152/jn.00764.2010
- PubMed
- Google Scholar
1. Quist BW
2. Hartmann MJ
(2012) Mechanical signals at the base of a rat vibrissa: the effect of intrinsic vibrissa curvature and implications for tactile exploration
Journal of Neurophysiology 107:2298–2312.
https://doi.org/10.1152/jn.00372.2011
- PubMed
- Google Scholar
Data
1. Rigosa J
2. Lucantonio A
3. Noselli G
4. Fassihi A
5. Zorzin E
6. Manzino F
7. Pulecchi F
8. Diamond M
(authors) (2017) Date from: Dye-enhanced visualization of rat whiskers for behavioral studies
Dryad Digital Repository.
https://doi.org/10.5061/dryad.9k9n0
1. Rigosa J
2. Lucantonio A
3. Noselli G
4. Fassihi A
5. Zorzin E
6. Manzino F
7. Pulecchi F
8. Diamond M
(2018) A Fluorescent Dye Method Suitable for Visualization of One or More Rat Whiskers
BIO-PROTOCOL 8:e2749.
https://doi.org/10.21769/BioProtoc.2749
- Google Scholar
Software
1. Rigosa J
(2017) whisker-dynamic-response, version a2e8d94
GitHub.

https://github.com/pgallir/whisker-dynamic-response
(2008) Embodied information processing: vibrissa mechanics and texture features shape micromotions in actively sensing rats
Neuron 57:599–613.
https://doi.org/10.1016/j.neuron.2007.12.024
- PubMed
- Google Scholar
Book
1. Robbins C
(2012)
Chemical and Physical Behavior of Human Hair

Berlin: Springer Berlin Heidelberg.
- Google Scholar
(2012) NIH image to ImageJ: 25 years of image analysis
Nature Methods 9:671–675.
https://doi.org/10.1038/nmeth.2089
- PubMed
- Google Scholar
1. Smith MB
2. Li H
3. Shen T
4. Huang X
5. Yusuf E
6. Vavylonis D
(2010) Segmentation and tracking of cytoskeletal filaments using open active contours
Cytoskeleton 67:693–705.
https://doi.org/10.1002/cm.20481
- PubMed
- Google Scholar
1. Venkatraman S
2. Elkabany K
3. Long JD
4. Yao Y
5. Carmena JM
(2009) A system for neural recording and closed-loop intracortical microstimulation in awake rodents
IEEE Transactions on Biomedical Engineering 56:15–22.
https://doi.org/10.1109/TBME.2008.2005944
- PubMed
- Google Scholar
1. Vincent SB
(1912)
The Function of Vibrissae in the Behavior of the White Rat

1–2, The Function of Vibrissae in the Behavior of the White Rat.
- Google Scholar
(2008) Unsupervised whisker tracking in unrestrained behaving animals
Journal of Neurophysiology 100:504–515.
https://doi.org/10.1152/jn.00012.2008
- PubMed
- Google Scholar
(2015) Deflection of a vibrissa leads to a gradient of strain across mechanoreceptors in a mystacial follicle
Journal of Neurophysiology 114:138–145.
https://doi.org/10.1152/jn.00179.2015
- PubMed
- Google Scholar
1. Williams CM
2. Kramer EM
(2010) The advantages of a tapered whisker
PLoS One 5:e8806.
https://doi.org/10.1371/journal.pone.0008806
- PubMed
- Google Scholar
1. Yan W
2. Kan Q
3. Kergrene K
4. Kang G
5. Feng XQ
6. Rajan R
(2013) A truncated conical beam model for analysis of the vibration of rat whiskers
Journal of Biomechanics 46:1987–1995.
https://doi.org/10.1016/j.jbiomech.2013.06.015
- PubMed
- Google Scholar
1. Yu C
2. Horev G
3. Rubin N
4. Derdikman D
5. Haidarliu S
6. Ahissar E
(2015) Coding of object location in the vibrissal thalamocortical system
Cerebral Cortex 25:563–577.
https://doi.org/10.1093/cercor/bht241
- PubMed
- Google Scholar
(2011) Whisking and whisker kinematics during a texture classification task
Philosophical Transactions of the Royal Society B: Biological Sciences 366:3058–3069.
https://doi.org/10.1098/rstb.2011.0161
- PubMed
- Google Scholar
1. Zuo Y
2. Safaai H
3. Notaro G
4. Mazzoni A
5. Panzeri S
6. Diamond ME
(2015) Complementary contributions of spike timing and spike rate to perceptual decisions in rat S1 and S2 cortex
Current Biology 25:357–363.
https://doi.org/10.1016/j.cub.2014.11.065
- PubMed
- Google Scholar

Decision letter

David Kleinfeld

Reviewing Editor; University of California, San Diego, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Dye-enhanced visualization of rat whiskers for behavioral studies" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by David Kleinfeld as Reviewing Editor and Andrew King as the Senior Editor. The following individual involved in review of your submission has agreed to reveal his identity: Carl CH Petersen (Reviewer #3).

The reviewers have discussed the reviews with one another. The Reviewing Editor feels that this could be a valuable contribution to the field of vibrissa exploratory behavior as the tracking of vibrissae is essential for a large group of experimentalists and your technique offers an alternative strategy to existing bright-field and dark-field techniques. He has drafted this decision letter to help you prepare a revised submission. Since additional data are required we realize that more time may be needed, and given that this is a methods paper we will reassign this to the Tools and Resources category, as it would be inappropriate to be considered as a research Short Report.

Essential changes:

1) Few details are given about the dye and how it is applied. Exactly what dye was used, as "Stargazer" sells multiple green and red colors.

2) You need to demonstrate tracking position and shape of whiskers in the intact animal.

3) Related to point (2), accurate measurements of deflection and bending require video recording at 1 kHz or more. Is there sufficient signal-to-noise for this to occur? How much dye is deposited and what is the weight of the dye required for such a signal-to-noise ratio and how does this compare to the weight of the unpainted vibrissa? What is the intensity of the excitation light?

4) Please comment on the bleaching rate of the dye. Please comment on the ability to re-stain the vibrissae.

Reviewer #1:

In this technical manuscript, the authors devise a new method for whisker visualization and tracking. They use readily available fluorescent dye suitable for visualization of one or more rat whiskers. The process makes the whiskers easily visible against a dark background. The coloring does not influence the behavioral performance of rats trained on a vibrissal vibrotactile discrimination task, nor does it affect the whiskers' mechanical properties. I generally found the manuscript well organized, written in clear language and appropriately discussed in the context of the current literature, all of this making the described findings accessible to a wide readership.

Next to my overall positive impression, the paper has several drawbacks.

The authors examined four rats, well-trained in a whisker-mediated vibrotactile delayed comparison task (Fassihi 2014), for behavioral effects of dye application. However, this behavioral task does not depend on the mechanical properties of the whiskers, since whisker's movement is induced by mechanical deflection. I suggest taking these data out.

Reviewer #2:

Rigosa et al. offer a novel method of fluorescently dying rat whiskers in order to solve limitations of traditional high-speed video based on brightfield imaging. Such a method would, as the authors point out, be useful in that it (a) allows individual whiskers to be tracked over time among a field of many unlabeled whiskers, (b) could in principle allow this tracking to occur for multiple whiskers simultaneously with color-based image segmentation, (c) in principle allow higher luminance contrast compared with brightfield imaging. This is therefore a potentially important method. The authors actually do not directly show (a)-(c), but instead focus on testing the potential "deal-killer" downsides of the method, namely that it would significantly alter whisker mechanics and disrupt whisker behavior. They do a reasonable job of this testing and find that mechanics and behavior are not noticeably disrupted. There is room for improvement with respect to including information necessary for other labs to evaluate and apply the method. Detailed comments follow.

More information about the dye and its application is needed:

The authors list their dyes as simply "green and red dyes – Stargazer", but Stargazer has multiple green and red colors and it's not clear which exactly the authors used. Moreover, I could not find any product information online that would be useful for understanding its uses as a fluorescent marker of whiskers. What are the relevant chemical properties? (If Stargazer stops making these particular colors, or if one wanted to generalize the method, what factors would one use to choose a different dye?)

How much dye must be applied, for instance per unit of surface area or volume? The authors describe a timed application of 30 min. How much dye is deposited in this time? The authors could perhaps do a calibration based on fluorescence from a known volume of dye, or based on changes in weight. The amount of dye required for good signal-to-noise is important, for example in determining whether the method is likely to be useful for the much smaller whiskers of mice.

How quickly does the dye bleach? To what degree is fluorescent bleaching a limiting factor in use of the method?

Related to the efficiency of the fluorescence and how much dye must be applied, one would like to know how intense the illumination must be. Because bright light is a stressor for rodents and the illumination is in visible wavelengths, this is relevant for use in behavioral studies.

In the Introduction the authors mention a limitation of traditional (brightfield) whisker imaging, which they imply their method will overcome: "Unless the whisker is uniformly illuminated and reflects the light homogenously…image quality will be degraded." Uniform illumination (and detection) is also important with their method, however. This may be evident in their supplemental video: the fluorescence intensity appears to vary a lot from one frame to the next (e.g. frames 61 vs. 62). I would assume this could make comparing whisker shape from frame to frame more difficult. It's not clear that their method offers a way around the stringent illumination requirements of traditional approaches. Please discuss.

In the Introduction, the authors then state that a "higher degree of luminance contrast could solve many of the technical challenges inherent to videography." However, the authors do not report any quantification of the luminance contrast they achieve. They should probably add this.

In the Results: "Video recordings confirm the increased visibility of the dyed whiskers in the mobile rat, facilitating tracking (Supplementary video)." The authors don't actually demonstrate facilitated tracking. In fact, although they set up the paper in the Introduction as offering a useful alternative to traditional approaches to whisker imaging/tracking, the authors never actually demonstrate its use for tracking position and shape of whiskers on the intact animal. It seems important to show this.

In Figure 3, it would be helpful to show example maps of the dyed conditions for the same whiskers.

The assumption of a uniformly tapered conical shape for dyed whiskers should be justified with experimental measurements. In general, the authors should characterize the shape of a whisker before and after dye application. Although not as straightforward, I suggest that one way the authors could add significant value would be to estimate Young's modulus before and after dye application.

The authors claim their dye labeling method is "long-lasting" (subsection “Dye effects on mechanical properties”, last paragraph). Figure 4E-F show for cut whiskers the stability of mechanics across months (assuming "M" in the x-axis labels indicates months; please include in legend). However, I could not find information showing that the fluorescence is long-lasting. Please clarify.

Reviewer #3:

Rigosa et al. apply fluorescent dye to rat whiskers to enable high contrast visualisation of whisker motion. This method might be useful for many researchers.

The authors give very few details about the dye, how it is applied, how it is excited, and how the fluorescence is imaged. For a methods paper, this is rather surprising!

The authors do not appear to track fluorescent whisker movements in behaving rats. This would be an important proof of concept: to track fluorescent whiskers during task performance at 2 kHz.

I am not convinced that the overall advance is of sufficiently general importance for publication in eLife.

https://doi.org/10.7554/eLife.25290.014

Author response

[…] Essential changes:

1) Few details are given about the dye and how it is applied. Exactly what dye was used, as "Stargazer" sells multiple green and red colors.

We have adapted the draft to include the requested information. Below is the relevant section of the revised manuscript in the section entitled “Application of fluorescent dye”.

“We used the “Hair Color Rinse Semi Permanent” line, in particular “UV Red” and “UV Green” in Figure 1 A-C. […] The specific cosmetic colorants for “UV Red” are: CI 42520, Basic Orange 31, Basic Violet 16, Basic Yellow 40, HC Yellow 4, CI 16255. The specific cosmetic colorants for “UV Green” are: Basic Yellow 40, CI 42090, CI 61570, HC Yellow 4, CI 62045.”

Also in the section “Application of fluorescent dye” we have added the following text:

“Semi-permanent hair rinse is designed to diffuse into and bind to the hair (Robbins 2001). In order to bleach the whiskers before dye application, we applied a gel of 3% hydrogen peroxide mixed with a booster powder (Robbins 2001, Table 4-3). The chemical bleaching process would be expected to follow a diffusion law, where the diffusivity for the human hair has been estimated as 1.8 x 10-9 cm2 /min (Robbins 2001, Figure 4-3). Afterwards, we washed the whiskers with distilled water, dried them, and then submerged them for 30 min in the hair color gel.”

The methodology described above is accompanied by a new figure in Supplementary file 2.

2) You need to demonstrate tracking position and shape of whiskers in the intact animal.

To demonstrate tracking position and shape of whiskers in the intact animal, we video recorded naive rats as they explored a water spout (Video 2). Supplementary file 1 is a frame of the video, now shot at 1,000 frame/sec.

For the whisker tracking we used an open source solution, ImageJ. We added the following text in the Results section “Dye-enhanced visualization”:

“We performed whisker tracking using JFilament (Smith et al. 2010), a plugin of ImageJ (Schneider et al. 2012), a public domain Java-based image processing program. JFilament can segment and track 2D/3D filaments and has been widely used for fluorescence microscopy images. It allows tracking of these elements through a sequence of frames and is thus suitable for whisker tracking.”

3) Related to point (2), accurate measurements of deflection and bending require video recording at 1 kHz or more. Is there sufficient signal-to-noise for this to occur?

As mentioned in the reply to point (2), we have performed video recording at 1,000 frame/sec. The signal-to-noise ratio at this frame rate remains sufficient for tracking. The tracked position of the dyed whisker has been added in an overlay to the new video, as described above.

How much dye is deposited and what is the weight of the dye required for such a signal-to-noise ratio and how does this compare to the weight of the unpainted vibrissa? What is the intensity of the excitation light?

Below, we reply first with the theoretical framework and then the precise empirical data required by the Editor. In the original draft, we thoroughly tested the frequency response function (FRF) of the vibrissae before and after the application of the fluorescent dye (Figure 4). The negligible before/after difference implies that the ratio between the Young modulus and the mass density was unchanged. For the FRF to be conserved, either the Young modulus and the mass density were not significantly altered (the simplest explanation) or else they both changed, but changed in a manner that kept their ratio constant (a more complex and thus less likely explanation). Assuming the same diffusion dynamics for the dye and the bleaching molecules, the penetration depth would be of the order of 1 μm, hence no significant change in whiskers’ mass would be expected. Beyond the theoretical framework, we carried out testing as described in Methods section “Quantification of the deposited dye and of the photobleaching.” The relevant text is:

“As an empirical verification, we measured the mass of a set of 34 whiskers (a multiwhisker sample reduces the likelihood of measurement error) by means of a scale (RADWAG AS 220.R2) with an accuracy of 0.1 mg. Prior to processing, the pooled mass was 14.8 mg. No difference in mass was detected after the chemical bleaching, while the coloring produced an increment of 0.1 mg (from 14.8 to 14.9 mg; within the measurement error of the scale), less than 0.7%.”

4) Please comment on the bleaching rate of the dye. Please comment on the ability to re-stain the vibrissae.

We measured the bleaching and modelled it with an exponential decay. The time constant under strong LED illumination is about 14 hours (about 50,000 seconds), equivalent to about 25,000 trials of 2 seconds (Supplementary file 3).

We added following text in Methods section “Quantification of the deposited dye and of the photobleaching”:

“We measured the photobleaching of the dye for seven hours with all 7 of the lamp’s LEDs pointing at the whiskers at a distance of about 15 cm. […] The measured time constant was about 14 hours (about 50,000 seconds), equivalent to about 25,000 trials of 2 seconds.”

Reviewer #1:

[…] The authors examined four rats, well-trained in a whisker-mediated vibrotactile delayed comparison task (Fassihi 2014), for behavioral effects of dye application. However, this behavioral task does not depend on the mechanical properties of the whiskers, since whisker's movement is induced by mechanical deflection. I suggest taking these data out.

The authors feel that the behavioral verification is critical to the work and the manuscript would be denatured by removal of these contents. At the cost of just two panels of one figure, we can give at least a substantial (if not complete) response to a question that many readers would likely pose: “The method improves whisker visibility, but how do rats deal with it?” We have to disagree with the reviewer’s contention that “this behavioral task does not depend on the mechanical properties of the whiskers, since whisker's movement is induced by mechanical deflection.” When mechanical deflection is applied to a guitar string, surely the properties of that string still matter. In the case of the selected behavioral task, transmission of mechanical energy from the contact point on the moving plate to the whisker follicle does depend on mechanical properties, as noted by reviewer #2.

Reviewer #2:

We thank the reviewer for the critique. In the revision, we have focused additional effort on issues (a)-(c), as requested by the reviewer. Some of changes detailed above for reviewer #1 are also relevant for reviewer #2; other changes are specific to reviewer #2 and are detailed below.

Detailed comments follow.

More information about the dye and its application is needed:

Our reply to this point was addressed directly to the Editor, above.

Likewise, our reply to this point was addressed directly to the Editor. In particular, we tested the mass difference and found it to be at most 1% (limited by scale accuracy). For this reason, it would not be a concern to increase the exposure time of the whisker to the dye in order to obtain higher contrast, nor to re-apply at different times.

How quickly does the dye bleach? To what degree is fluorescent bleaching a limiting factor in use of the method?

See reply to the Editor above.

We successfully shot a video at 1,000 fps of a rat while behaving (drinking) and we provide whisker tracking for that video. We agree that in behavioral studies the light source should be positioned and angled in a manner that provides the highest possible illumination of whiskers, together with the lowest possible illumination everywhere else, especially in a manner that avoids direct shining into the eyes. But such configurations exceed the aim of this paper. In our experience, the rat does not avoid the area of the behavioral apparatus containing the light source. They show no aversive behavior or signs of stress (crouching, arching, excessive grooming, etc.). In the session where we shot the video, the rat ate and drank at the spout when it was in the illuminated position just as willingly as when the spout was in a dark corner. Last but not least, our veterinarian carried out an optical exam and found no detrimental effects. However, our paper is about whisker coloring and we believe that issues of illumination are beyond our scope.

Uniform illumination is crucial for our method as well. Motivated by this and the other reviewers, we put additional effort into the collection of high quality, well-illuminated video, which is now presented in the revision.

We did not quantify the illumination because there is no fixed, agreed-upon level of illumination that will satisfy every investigator for every study. The statement is general and uniformly true: in every condition, a higher degree of luminance contrast is always better. We added a new video shot at 1,000 fps with appropriate light conditions and optical filtering. We provided information about the light conditions used for that video. We expect that investigators will compare the whisker to background contrast in that video, and in Figure 1, to their current methods and will appreciate the improved contrast.

In the Results: "Video recordings confirm the increased visibility of the dyed whiskers in the mobile rat, facilitating tracking (supplementary video)." The authors don't actually demonstrate facilitated tracking. In fact, although they set up the paper in the Introduction as offering a useful alternative to traditional approaches to whisker imaging/tracking, the authors never actually demonstrate its use for tracking position and shape of whiskers on the intact animal. It seems important to show this.

The revised manuscript demonstrates whisker tracking on a 1,000 fps video.

In Figure 3, it would be helpful to show example maps of the dyed conditions for the same whiskers.

Figure 3 only explains how to obtain the data that is shown in the matrices of Figure 4. Figure 4C fulfils the reviewer’s request.

We report no significant change in mass after the coloration (Methods section “Quantification of the deposited dye and of the photobleaching”) and no difference of mechanical dynamics, hence the direct conclusion is that Young modulus was not changed.

We added a figure (Supplementary file 3) to detail photobleaching (i.e., how long the fluorophore continues to fluorescence after exposure to strong light). The lifetime, in terms of light exposure in an experimental apparatus, is longer than what would be needed in most studies.

Reviewer #3:

Rigosa et al. apply fluorescent dye to rat whiskers to enable high contrast visualisation of whisker motion. This method might be useful for many researchers.

The authors give very few details about the dye, how it is applied, how it is excited, and how the fluorescence is imaged. For a methods paper, this is rather surprising!

We have added this information to the revised manuscript.

The authors do not appear to track fluorescent whisker movements in behaving rats. This would be an important proof of concept: to track fluorescent whiskers during task performance at 2 kHz.

We added a video recorded at 1,000 fps. We demonstrated whisker tracking on that video. In the literature, 1 kHz is the most common frame rate used for tracking since whisker dynamics have been found not to exceed 500 Hz. In the same session we recorded data up to 2,000 fps but it did not seem to add much to the significance of paper.

I am not convinced that the overall advance is of sufficiently general importance for publication in eLife.

We regret hearing this, but we believe that this method would be helpful for many laboratories that study tactile behavior in rodents because it allows easy single-whisker visualization without trimming.

https://doi.org/10.7554/eLife.25290.015

Article and author information

Author details

Jacopo Rigosa

International School for Advanced Studies, Trieste, Italy

Contribution
JR, Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, designed the study, set up the ex vivo isolated whisker vibration apparatus, wrote data acquisition code, conducted the ex vivo whisker experiments, designed the analyses, executed the analyses and created figures, wrote the manuscript first draft, revised the manuscript

For correspondence
jacopo.rigosa@gmail.com

Competing interests
The authors declare that no competing interests exist.

"This ORCID iD identifies the author of this article:" 0000-0003-3184-994X
Alessandro Lucantonio

International School for Advanced Studies, Trieste, Italy

Contribution
AL, Conceptualization, Resources, Software, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing—review and editing, designed the study, modeled whisker mechanical responses, designed the analyses, executed the analyses and created figures, revised the manuscript

Contributed equally with
Giovanni Noselli

Competing interests
The authors declare that no competing interests exist.
Giovanni Noselli

International School for Advanced Studies, Trieste, Italy

Contribution
GN, Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, designed the study, modeled whisker mechanical responses, designed the analyses, executed the analyses and created figures, wrote the manuscript first draft, revised the manuscript

Contributed equally with
Alessandro Lucantonio

Competing interests
The authors declare that no competing interests exist.
Arash Fassihi

International School for Advanced Studies, Trieste, Italy

Contribution
AF, Conceptualization, Resources, Supervision, Visualization, Methodology, Writing—review and editing, designed the study, set up the in vivo behavioral apparatus, conducted the behavioral experiments, designed the analyses, executed the analyses and created figures, revised the manuscript

Competing interests
The authors declare that no competing interests exist.

"This ORCID iD identifies the author of this article:" 0000-0001-7477-7080
Erik Zorzin

International School for Advanced Studies, Trieste, Italy

Contribution
EZ, Conceptualization, Resources, Software, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing—review and editing, designed the study, set up the ex vivo isolated whisker vibration apparatus, wrote data acquisition code, conducted the ex vivo whisker experiments, designed the analyses, executed the analyses and created figures, revised the manuscript

Competing interests
The authors declare that no competing interests exist.

"This ORCID iD identifies the author of this article:" 0000-0002-4083-7847
Fabrizio Manzino

International School for Advanced Studies, Trieste, Italy

Contribution
FM, Resources, Software, Methodology, Writing—review and editing, set up the in vivo behavioral apparatus, wrote data acquisition code, revised the manuscript

Competing interests
The authors declare that no competing interests exist.
Francesca Pulecchi

International School for Advanced Studies, Trieste, Italy

Contribution
FP, Resources, Supervision, Methodology, Writing—review and editing, conducted the behavioral experiments, revised the manuscript

Competing interests
The authors declare that no competing interests exist.
Mathew E Diamond

International School for Advanced Studies, Trieste, Italy

Contribution
MED, Conceptualization, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, designed the study, designed the analyses, executed the analyses and created figures, wrote the manuscript first draft, revised the manuscript

For correspondence
diamond@sissa.it

Competing interests
The authors declare that no competing interests exist.

"This ORCID iD identifies the author of this article:" 0000-0003-2286-4566

Funding

Ministero dell'Istruzione, dell'Università e della Ricerca (GA 280778)

Mathew E Diamond

International School for Advanced Studies (NOFYSAS 2012)

Giovanni Noselli

Human Frontier Science Program (RG0015/2013)

Mathew E Diamond

European Commission (Project 294498)

Mathew E Diamond

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We acknowledge the financial support of the Human Frontier Science Program (http://www. hfsp.org; project RG0015/2013), the European Research Council Advanced grants CONCEPT (http://erc.europa.eu; project 294498) and MicroMotility (project 340685), and Italian MIUR grant HANDBOT (http://hubmiur.pubblica.istruzione.it/web/ricerca/home; project GA 280778). GN gratefully acknowledges support by SISSA through the excellence program NOFYSAS 2012. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We also acknowledge Natalia Grion and Ben Mitchinson for their valuable consultation on whisker tracking software.

Ethics

Animal experimentation: The rats were under the care of a consulting veterinarian. Protocols followed the guidelines of EU Directive 2010/63/EU, established as Italian decree 26/2014, and were approved by the SISSA Ethics Committee and the Italian Ministry of Health license numbers 569/2015-PR and 570/2015-PR.

Reviewing Editor

David Kleinfeld, University of California, San Diego, United States

Publication history

Received: January 20, 2017
Accepted: June 13, 2017
Accepted Manuscript published: June 14, 2017 (version 1)
Version of Record published: July 14, 2017 (version 2)
Version of Record updated: November 3, 2017 (version 3)
Version of Record updated: March 13, 2018 (version 4)

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.