by a whisker

Not all my time here at the Kavli  Institute is spent listening to talks about nonequilibrium dynamics and supersymmetry and such. We're practically next door to the UCSB Physics Department, which means I can occasionally nip on over to catch the occasional departmental colloquium. My first week in Santa Barbara, I heard a terrific talk by David Kleinfeld of the University of San Diego, who studies how rats combine motor and sensory signals in their tiny little rodent brains using an imaging technique known as an electromyogram, which records muscle movement. He's not the only one, either: in the last decade or so, there's been an explosion of research in this area among neurobiologists, for some very practical reasons. While the human brain is more complicated than a rat brain, there are some striking similarities. The hope is that learning more about rodent neural pathways will help scientists develop new treatments for brain-related disorders, like Alzheimer's (or stroke, or Parkinson's), and to design prosthetic limbs that can be directly controlled by one's thoughts (technically by the electrical signals the brain uses to communicate — not some weird kind of psychic ability).

Scientists have been playing with rats in the lab since the early 1800s, starting with simple mazes and teaching them to tap on levers. Now they're using them to try and solve one of the most fundamental questions of perception, in both humans and animals: how do we know where objects are in relation to our bodies? It turns out that rats have an unusual means of navigating: their whiskers, or as Kleinfeld calls them, vibrissa. Rats have about 30 large whiskers and dozens of smaller ones. They're part of a surprisingly complex "scanning sensorimotor system" that enables the rat to perform such diverse tasks as texture analysis, active touch for path finding, pattern recognition, and object location, just by scanning the terrain with its whiskers.

Sure, the whiskers are just hairs, a collection of dead keratin cells, much like human hair. It's what they're attached to that make them useful, and as sensitive as human fingertips. Each rat whisker is inserted into a follicle that connects it to "barrel" made up of as many as 4000 densely packed neurons. Together they form a grid or array that serves as a topographic "map," telling the rat's brain exactly what objects are present, and movements that are taking place, at any given place and time. All those barrels in turn are wired together into a kind of neural network, so the rat gets multidimensional cues about its environment. As it moves across  the terrain , a rat is constantly scanning its surroundings with its whiskers, sweeping back forth between five and 12 times per second (i.e., between 5 and 12 Hz). When a whisker hits an object, it bends in its follicle, and this sets of an electrical impulse to the brain that enables the rat to determine both the direction and how far each whisker moves. That's how the animal senses its environment and forms an "image" of the world around it.

Basically, the rat whiskers trigger nerve signals that the brain decodes. How does the brain do this? Apparently through a feedback loop that is very similar to that created by an FM radio's circuitry, which generates an intelligible signal by comparing electrical patterns from an oscillator to those from incoming radio waves. For instance, certain neurons in the rat cortex pulse at very precise frequencies (near 10 Hz), and these pulses are sent continuously to the thalamus, which compares them with incoming whisker signals. Instant decoding!

In 2003, scientists discovered that a rat's whiskers are also used for hearing, because they resonate at certain frequencies. It's the same principal that applies to the strings of a harp, or a piano, or a guitar: longer whiskers resonate at lower frequencies, while shorter whiskers resonate at higher ones. (For instance, one particular whisker was found to vibrate at 182 Hz, which corresponds to a musical note between F and F#, below middle C.)

Rats have shorter whiskers near the nose, longer ones further back, and this enables them to create a kind of "frequency map" by poking its nose all over the place. A single whisker acts much like a single-pronged tuning fork. Put them all together, and a rat can sense size, position, the edges of objects, even slight variations in texture, relative to its little rodent body. For instance, a very fine texture would set up a stronger vibration in a high-frequency whisker than it would in a low-frequency whisker.

Kleinfeld's team has taken things one step further, trying to determine where the input signal is coming from: for instance, is the rat actually "listening' to the electrical impulses generated from the whiskers (reafferent) or is it using a copy of this sensory information for processing (efferents copy)? and how the heck can scientists figure this out? The first step is to simplify the system to isolate one particular whisker attached to a "barrel" of neurons.

Kleinfeld solved the problem by essentially removing all but one of a lab rat's whiskers to see just how sensitive this system is. He admits the rats find this "very upsetting" –  how would you feel if you lost most of your means of sensing your surroundings? — but like human brains, the brains of rats can adapt to even drastic changes. In the 1970s, scientists learned that removing even a single whisker in the first few days of a rat's life would affect its ability to learn and function, but the animal would soon adapt — much like someone born with only four fingers quickly adapts to what might be a major adjustment if it happened in adulthood. But all the whiskers? That's like losing everything except your big toe. How much can a rat really sense with so much of its system depleted?

Rather a lot, it turns it, in part thanks to those nested loops that comprise the vibrissa sensorimotor system, proving many levels of feedback. Kleinfeld found that the rats could distinguish angular position, relative to their faces, even with a single vibrissa. Once this had been established, he was able to numb the nerve with Lidocaine to temporarily block the signal — but only on on side. The signal was successfully blocked, which seems to be a pretty clear indication that rats depend predominantly on an outside signal (the reafferent model). 

Ah, but wait! The whole efferent (or central) copy theory seems to supply information regarding amplitude — an internal signal source, rather like a built-in calibration system. And we all know how important proper calibration is for any measure device or system. So you've essentially got two incoming signals that interact, with one signal modulating the other like a circuit switch, and together provide the rat with its impressive powers of sensory navigation. Kleinfeld admits that this particular insight is closer to conjectural than an actual proven hypothesis. The next step in his research to to look for the kinds of neural cells and circuitry in the rat brain that could comprise such a system design.

That was the gist of Kleinfeld's talk. It seems to dovetail nicely with a 2004 experimental result conducted by researchers at the Max Planck Institute for Medical Research in Heidelberg, Germany, in which they found that activating a single brain cell in a rat could make its whiskers twitch. They used tiny electrodes to excite those individual neurons in the motor cortex of rats (mercifully anesthetized). It just so happens that the motor cortex doesn't control muscle movement directly. Rather, it signals a pattern generator, and this in turn sends the detailed commands to the muscles.

Who knew rats were such complicated creatures? It seems like such a simple strategy, little more than a sweep of one's immediate surroundings, but the animal manages to collect and process a staggering amount of sensory data about objects, textures, and so forth — all by a whisker.