McGovern Institute for Brain Research and Department of Brain and Cognitive Science, Massachusetts Institute of Technology,Cambridge, MA, USA*Correspondence: jritt@mit.edu (J.T.R.), cim@mit.edu (C.I.M.)DOI 10.1016/j.neuron.2008.11.022Sensorysignalsgeneratedduringfreebe-havior areshaped by objects in the world,the animal’s active sensing choices, andthe innate construction, or embodiment,of the sensory system. Together theseconstitute the ‘‘natural scene’’ of sensoryinputs. Our study (Ritt et al., 2008) wasthe first to measure the fine-scale signalstransmitted through vibrissae whileanimals freely explored textured surfaces.Specifically, we quantified ‘‘micromo-tions,’’ small-amplitude, high-velocity sig-nals thought by most researchers in thefield (Carvell and Simons, 1990, 1995;Mehta and Kleinfeld, 2004; Neimarket al., 2003), including Diamond and col-leagues (Arabzadeh et al., 2005; Hippet al., 2006; von Heimendahl et al., 2007),to play a key role in texture perception.In their letter, Diamond et al. suggestthat our characterization of an embodiedproperty of perception—resonant tuningof vibrissae—is inappropriate. They pre-dict that resonance will not shape trans-ductioninaninitialcontactwindow,whichthey claim is the relevant period in texturetasks, the only time when the brain is ‘‘lis-tening.’’ We directly address this factualconcern below by reanalyzing our data,and observe resonance tuning within theinitial contact window. We also addressconceptual issues raised by their letter,in the hopes that this discussion will helpmove the field forward.We first provide brief historical contextanddescribewhatwebelievetobeamis-understanding about resonance. Despitewidespread study of the vibrissa sensorysystem as a model, consideration of thepossibleroleofitsembodimenthasarisenonlyrecently.Aninitialobservationinvitrowas that vibrissae express mechanicalfrequency tuning related to their length,much as harp string length determinespitch (Andermann et al., 2004; Hartmannet al., 2003; Mehta and Kleinfeld, 2004;Moore and Andermann, 2005; Neimarket al., 2003), leading to the ‘‘resonancehypothesis.’’ The central idea was thatvibrissae might filter sensory informationrather than veridically transmit surfaceprofile. A suggestive further point wasthat stereotyped differences in vibrissalengths across the face, providing differ-ences in frequency tuning, might lead toa ‘‘preneural’’ organization and process-ing of sensory inputs.Onthebasisofinvitroandexvivostud-ies in which they swept real and metallicvibrissae over textures at a single velocity(Arabzadehetal.,2005;Hippetal.,2006),Diamond and colleagues questionedwhether resonance would shape trans-duction in freely behaving animals andappeared to favor the idea instead thatvibrissae are largely interchangeabledespitevariationsintheirphysicalproper-ties. Addressing this question requiredmeasurement in freely behaving animals,in part because sensing choices (forexample, speed at which vibrissae areswept over a surface) should be a strongdeterminant of contact-induced micro-motions. We showed (Ritt et al., 2008)that vibrissa length is correlated withmicromotion mean frequency in freelybehaving animals exploring surfaces,resolving the original concern raised byDiamond and colleagues (a finding repli-cated in Wolfe et al., 2008).To understand this finding, it is impor-tanttoconsiderthenatureoftransductionfromcomplexsurfaces.Micromotionsareoftenfarfromperiodicoscillations,yetav-erage micromotion frequencies may stillexhibit a biomechanical, vibrissa-specificbias, which is the definition of resonance.For example, the amplitude of a given mi-cromotion could depend on where it fallswithin a temporal pattern of micromo-tions. Sustained oscillations on smoothsurfaces, due to friction, also demon-strate an impact of biomechanics beyondsimple transmission of surface profile(e.g., Figure 5 in Ritt et al., 2008). Assuch, an absence of isolated oscillations(e.g., ‘‘rings’’) is not in itself evidencethat resonance plays no role. Diamondet al. appear to equate resonance withrings only (for a similar view, see Wolfeet al., 2008). This view is also indicatedby their incorrect attribution of the term‘‘microvibration’’ to us, a word that ap-pears nowhere in our manuscript, andthat they define as distinct from all othermotions.Instead,weusetheterm‘‘micro-motion’’ to refer to all small-amplitude,high-velocity motions, including frictionalstick/slips as well as rings. In this termi-nology (derived from Brecht et al., 1997;Carvell and Simons, 1995), the ‘‘kineticsignature’’hypothesisof(Arabzadehetal.,2005; Hipp et al., 2006) is the statementthat different surfaces generate distinctpatternsofmicromotions.Thesepatterns,transduced by vibrissae, should be sub-ject to biomechanical filtering (e.g., reso-nance), including variation as a functionof vibrissa identity.In their letter, Diamond et al. expressa new concern, that resonance mayshape transduction, but not during theright window. In their recent paper (vonHeimendahl et al., 2007), they found ratsmade texture judgments quickly, withshort vibrissa contacts, and suggest thatif resonance shapes transduction only af-terward,itisperceptuallymoot.Todirectlyaddress this question, we reanalyzed theNeuron 60, December 11, 2008 a2008 Elsevier Inc. 745