Wires mounted on small motors that move from left to right mimic a rat's 

 whiskers sweeping across a bumpy surface. If a length of wire is plucked and 

 allowed to vibrate freely, it vibrates at its own "natural frequency, " which 

 depends primarily on its length: the longer the wire, the lower its natural 

 frequency, or pitch. In contrast, a wire that is swept across a bumpy surface is 

 forced to vibrate at the frequency at which the tip encounters the bumps. 

 When that frequency approximately matches the natural frequency of the 

 wire, the wire resonates: the amplitude, or "loudness, " of the vibration 

 becomes large. Otherwise, the vibrational amplitude remains relatively small. 

 Thus the short wire in the schematic diagram (left) responds weakly, at low 

 amplitude, when it sweeps across the coarse-textured part of the surface, but 

 responds strongly, at high amplitude, when swept across the fine-textured 

 part of the surface. A longer wire (right) produces the opposite response. 



the shapes of the objects encoun- 

 tered by the artificial whiskers with 

 remarkable fidelity. 



Hartmann designed her model 

 without knowing just how the rat 

 processes the information gathered 

 by its rapidly roving whiskers. But 

 the success of her robotics experi- 

 ments so intrigued her that she won- 

 dered whether the rat has innate sen- 

 sors that can detect the curvature of 

 each hair, much as her robotic sen- 

 sors were doing. Does the rat's brain 

 process the information from the 

 whiskers much the way her comput- 

 er was analyzing and integrating in- 

 formation from its artificial bristles? 



By recording signals from a nerve 

 ganglion in the face of a rat, Hart- 

 mann helped show that rats do in- 

 deed have neurons that encode a sig- 

 nal for curvature. She then realized 

 how incredibly sensitive the neurons 

 are. Other signals, at a higher input 

 rate, were stimulating the neurons as 

 well — signals that she hadn't initially 



equipped her robot to recognize. It 

 seems that the neurons of a rat 

 whisker fire in response to stimuli 

 that reach the neurons at a hundred 

 times the "whisking" frequency. 

 Hartmann was reminded of two pa- 

 pers — one of her own and the other 

 by Christopher I. Moore, now at 

 M.I.T. — that examined the vibrations 

 ot rat whiskers. Might the higher- 

 frequency responses have something 

 to do with vibrations, she wondered, 

 rather than the angle at which the 

 whiskers curve? 



Imagine a whisker as a single- 

 pronged tuning fork whose pitch 

 is determined by its shape. A short 

 whisker would resonate at a high 

 hum, whereas a longer hair would 

 resound at a lower pitch. Because the 

 rat's whisker array is a collection of 

 similarly shaped hairs that vary in 

 length and width, perhaps the rat 

 could make some sensory use of the 

 pitch of each whisker's vibration. 



Hartmann and Moore established 

 that vibrating whiskers do indeed 

 have natural frequencies ranging 

 from twenty-seven to 260 hertz, 

 much higher than the normal whisk- 

 ing frequencies. 



The ability to sense vibrations at 

 various frequencies could be a pow- 

 erful tool for exploring textures. A 

 fine texture might set up a stronger 

 vibration in a high-frequency 

 whisker than it does in a low-fre- 

 quency one. Thus a whisker array 

 might be an ingenious technique for 

 sensitively probing the ground over 

 which a robotic rover is rolling [see 

 diagram at left]. 



Vibrational cues could also be 

 highly practical for determining the 

 edges of objects. A whisker in con- 

 tact with a piece of cheese has both 

 ends relatively fixed: one is stuck in 

 the rat's face, while the other is flat- 

 tened against the cheesy surface. But 

 when the whisker breaks free of the 

 cheese, it must vibrate in a character- 

 istic, and perhaps recognizable, way. 

 It seems that rats can sense size, posi- 

 tion, and texture just by sticking 

 their noses into things. 



I particularly enjoy research that 

 goes from a model to an inspiring 

 organism and back again, to see 

 whether the lessons learned from the 

 model apply in nature. That approach 

 creates excellent opportunities for the 

 study of an animal to further inform 

 a design that is more practical, more 

 workable, more flexible. To this day, 

 Hartmann's laboratory is making 

 great leaps in understanding how rats 

 multitask to determine the shape of 

 the objects in their world. 



Of course I'm sorry I never knew 

 about all this in my youth. I could 

 have been live-trapping mice, de- 

 whiskering them, and releasing those 

 sensory-deprived creatures in the 

 hopes of generating a rodent fall. 

 Wouldn't my sister he thrilled? 



Adam Summers (asummers@uci.edu) is an 

 assistant professor ofbioengineering and of ecol- 

 ogy and evolutionary biology ai die I University 

 of California, Irvine. 



September 2006 NATURAL HISTORY 



