Prosthetics that can feel

New techology in prosphetics has led to a robotic fingertip that actually feels.
03 April 2016

Interview with 

Dr Calogero Oddo, The University of Pisa


If you have to have your hand amputated, the current generation of prostheses can only Baby fingersgo so far to replace what you've lost. They can restore some of the aesthetic and motor properties of the body part, but still missing is a sense of fine touch that can tell, for example, whether something is rough or smooth. And this is critical for manipulating objects correctly. But now we're a step closer to restoring this missing modality thanks to a bionic finger that uses piezoresistive sensors (sensors that respond to pressure) to detect surface textures and turn them into nerve signals that an amputee's brain can understand. Calogero Oddo is its creator, and he spoke about it to Chris Smith.....

Calogero - The natural skin has receptors internally that are the transducers (the physical interaction of our hand with external walls).  They transform into a sequence of neural signals that are then conveyed to the nerves and then arrive through the nerves up to the brain and they represent the tactile information to our brains so we can percept.

Chris - What is the material that you're working with that is capable of discerning textural features or how rough a surface is?  What are you making these things from?

Calogero - The system is multi-layered.  So we have this rubber that covers the sensor.  The sensors are made of silicon which is like glass, which has inside  piezoresistors that transform mechanical information, so they turn this information into electrical signals.  Then these are acquired by embedded electronics that convert this information into digital, and then there is an artificial neuron model that keeps the continuous wave forms that are required from the artificial sensors and transforms into the spikes that are those on/off events to stimulate the nerves.

Chris - How do you teach your computer programme what each surface feels like and then, how do you marry the recognition of that surface and its output of spikes into the nervous system so that it's putting them into the right pattern of spikes, which is what would happen in an intact individual, someone with a normal hand?

Calogero - This is done by a technique that is minimally invasive to record from the nerves of the intact subjects when they touch naturally the surfaces.  We are learning how the natural code is arranged in correspondence to a particular experiences and then we try to emulate.

Chris - So, in essence, what you do is you're recording what sort of patterns of nerve activity a person gets when they run their natural skin over the same surface and then when your new detectors experience that same surface contour, you just generate the same nerve output of spikes and that's what you're going to put into the nervous system?

Calogero - Yes, this is exactly what we are doing. The nice thing is that our finger is not specialised to the surfaces that we test in this experiment but, in principle, if we test other surfaces it is able to generalise.  So it is able to reconvert the other surface into different senses of the spikes that the brain can interpret and recognise.

Chris - How do you get the signals from your bionic finger back into the nervous system of the amputee in the first place?

Calogero - There was an implanted interface that allowed the stimulus to be delivered to the nerve.  What we do is insert into the nerves electrodes (like wires) to electrically connect the nerve.

Chris - If you actually do the experiment and you take the output from your system and you apply it to the nervous system of an individual, does it really feel to a normal human that that surface is the one that they should be feeling?

Calogero - This was declared by Dennis, one of the amputee subjects that tested the technology.  He said exactly this "not 100% but it makes sense to his brain".  Meaning that he was able to interpret the stimulus and to imagine the stimulus in pictures.


Add a comment