How our Brains Respond to Texture

Our hands are incredibly sensitive to texture. Even with your eyes closed, you can easily tell apart fine from coarse sandpaper, slick silk from soft cotton, planed wood from smooth plastic. How do our brains allow us to experience this rich textural world?

Sliman Bensmaia, associate professor at the University of Chicago, is one of the few neuroscientists who investigate the sense of touch. They are few for a reason, Bensmaia says: “It’s only the self-hating neuroscientists that are going to study the sense of touch. If you study vision, you buy a nice monitor. If you study hearing, you buy great speakers. But when you study touch, there is no such thing. You have to build a device specifically for your research question.”

To study texture, Bensmaia and his colleagues built a device consisting of a rotating drum covered in strips of coarse and fine textures, such as sandpaper of different grading, various fabrics, and patterned plastics. This drum runs the textures across the fingertips of Rhesus macaques, monkeys whose somatosensory system is similar to that of humans.

Feeling – a bit like seeing, a bit like hearing 

In 2013, Bensmaia and his lab used this device to help decode how nerves in the hands of macaques convey texture information. The team recorded how afferent nerves, which run from the fingertip to the brain stem, responded while the macaques touched different textures. What they found was surprising. When macaques feel coarse textures, such as Braille-like patterns of dots, the spatial pattern of the dots is reflected in the spatial pattern of nerve activity. This so-called isomorphic representation is similar to what happens in vision: When we see a spatial pattern of dots, this pattern is represented as the spatial pattern of activation in the nerve.

This spatial-style coding was already known from studies of Braille dots and raised ridges. But it wasn’t the whole story. “It’s known that the sensors in our skin are, generously speaking, about half a millimeter apart. If your sensors were half a millimeter apart and you were relying on a spatial pattern of activation, then you would not be very good at touch perception: You could only tell apart things larger than half a millimeter. But humans are really amazing a feeling texture, we can perceive elements down to tens of nanometers in size.”  So how is this amazing ability to feel tiny things achieved?

To find out, the Bensmaia lab tested the neural response to fine textures, such as nylon or satin. Here, the nerves encode textures in a temporal pattern of activation. “It’s in the timing of the response that information about fine textures is conveyed. With fine textures, like satin, silk, or sandpaper, each texture sends a different kind of Morse code to the brain,” Bensmaia says. This is similar to the way we hear, where information is also conveyed in time, not in space.

The reason we can feel tiny textures is because we instinctively run our hands across surfaces. “Vibrations are produced in the skin, and those vibrations depend on the textures you are exploring. The vibration-sensitive neurons in the hand, the PC afferents, pick up the vibrations and send these temporal patterns to the brain,” he says.

Making sense of it all 

So if nerves from our hands send information to our brain about texture in two different ways: as spatial pattern for coarse features, and as temporal pattern for fine features, how does the brain combine these two streams of information into one feeling of “texture?” Why, instead of feeling that something is “fine” or “coarse,” do we get “a holistic sense of texture,” as Bensmaia puts it?

In new research published this year in PNAS, led by postdoc Justin Lieber in Bensmaia’s lab, the researchers recorded macaques’ response to even more textures. This time, they recorded the response directly from the brain, using electrodes implanted into the somatosensory cortex of three monkeys.

This data showed how complex the neural response to texture is. The researchers didn’t find “silk neurons” or “satin neurons” that respond only when the monkeys touched silk versus satin. Instead, all neurons respond to some degree to all or most of the textures.

“All neurons are silk neurons; all neurons are satin neurons. All neurons respond to all textures, they just do so in so many different ways, to so many different degrees,” Bensmaia  says. It is the combination of which neurons are active when touching sandpaper versus when touching silk that allows us (or, in this case, macaques) to tell them apart. “Texture is a complex space. There are all kinds of words to describe texture: fuzzy, furry, glassy, velvety, and so on. That is reflected in a very complex neural representation,” Bensmaia says.

Esther Gardner, professor at the department of neuroscience and physiology at NYU, an expert in touch research who was not involved in the study, agrees. “This is very interesting from the point of view that it is not one cortical neuron that is sensitive to denim, versus one that is sensitive to corduroy and another one sensitive to Harris Tweed. Instead, the cortical neurons act like a symphony orchestra. It is the whole ensemble of firing patterns together that help differentiate between textures. This is very clearly demonstrated in the paper.”

The many dimensions of texture  

Our perception of color is sometimes described in terms of three dimensions: hue, saturation, and brightness. “If color has three dimensions, how many dimensions does texture have?” Bensmaia asks. To tease apart the dimensions of texture, the group used statistical methods such as principal component analysis to analyze the neural symphony. They found that texture has at least twenty dimensions. One dimension predicts how rough a texture is going to be. Another dimension is spatial scale, a continuum from coarse to fine, with some neurons responding preferentially to coarse textures while others responding more to fine textures. “What this high dimensional representation means is that we need to use large numbers of neurons to distinguish chiffon, denim, and huck towel,” Gardner says.

Lieber and Bensmaia recorded responses to 59 textures, and Bensmaia says that based on the neural response, he knows which texture the monkey is sensing: “I can tell you which texture it is, but I can also predict how the texture is going to feel. For example, I can tell you how rough a texture feels, based on where it falls along this continuum. If you give me two textures, I can see how far apart they are in this neural space. This predicts how different they feel.”

More work for neuroprosthetics 

Bensmaia also works in the field of neuroprosthetics, building robotic prosthetic limbs that are controlled by implanting arrays of electrodes in the somatosensory cortex. One goal is to restore the sense of touch through the stimulation of the brain. His new findings will make that project more difficult than he’d hoped.

“The representation of texture we found is certainly not the kind that is amenable to restoring through electrical stimulation of the brain. Electrical stimulation relies very strongly on a clean principle. It would be great for neuroprosthetics if we can put electrodes in, and when you touch silk, we stimulate the silk neurons. When you touch satin, we stimulate the satin neurons. But it’s the opposite of that, all neurons respond to all textures. So there is no clear way to reproduce texture given the current state of technology.”