Primul brat robotic contralat cu puterea mintii/The first mind-controlled robotic arm

Universitatea din Minnesota (SUA) tocmai a anuntat finalizarea primei interfete om-masina care face posibila miscarea unui brat robotic folosind doar puterea mintii. 

Este prima data in lume cand acest lucru este posibil: controland un brat robotic folosind doar gandurile, fara un implant de electrozi pe creier. 

Sistemul functioneaza cu ajutorul unei casti cu 64 de electrozi care achizitioneaza semnalele (extrem de slabe) emise de activitatea electrica a creierului. Acestea sunt prelucrate prin tehnici avansate de prelucrarea semnalelor si machine learning si apoi transpuse in miscari. 


Pentru a intelege cum functioneaza o interfata om-masina (BCI -Brain Computer Interface) este important de inteles activitatea electrica a creierului nostru. 
 

In creier se gasesc aproximativ 30 de miliarde de neuroni: acele celule nervoase legate intre ele prin axoni si dendrite. De fiecare data cand gandim, ne miscam, simtim, etc, neuronii nostri transporta semnale electrice (extrem de slabe), care trec din neuron in neuron cu o viteza foarte mare (aproximativ 400 km/h). Semnalele electrice sunt generate de diferentele de potential electric transportat de ioni pe membrana fiecarui neuron. 

Pentru a achizitiona cat mai bine aceste semnale, se foloseste un set de electrozi care sunt atasati pe scap. Acest sistem se numeste electroencefalograf (EEG). 

Semnalele achizitionate de electrozi sunt apoi interpretate cu ajutorul computerului si, prin tehnici de prelucrare de semnal si machine learning, transformate in actiuni. Astfel, pot fi directionate device-uri precum un brat robotic. 

Pentru a plasa cat mai bine electrozii pe scalp, astfel incat sa achizitioneze semnalele relevante pentru o anumita activitate, se foloseste RMN-ul. De exemplu, daca se urmareste proiectarea unei interfete om-masina pentru a misca un brat robotic cu ajutorul gandului, se face un RMN in care user-ului i se cere sa isi miste bratul. Astfel, se observa zona care se activeaza in cazul miscarii bratului. Pentru a coordona un brat robtic, electrozii vor fi plasati mai mult in zona obtinuta in urma RMN-ului. 

Semnalele emise de neuroni sunt extrem de slabe si, de cele mai multe ori,acestea pot fi atenuate de scalp, motiv pentru care, pana acum, s-au implantat electrozi sub scalp, pe creier, pentru o mai buna calitate a semnalelor. 

Cercetatorii de la Universitatea din Minnesota au reusit sa miste un brat robotic folosind doar tehnica non-invaziva, cu electrozii plasati pe scalp. 

Inventia promite sa ajute milioane de persoane paralizate sau care sufera de boli neurodegenerative. 

 

Urmatorul pas, afirma cercetatorii, este sa aplice tehnologia pentru a misca un brat protetic atasat unei persoane, precum si in cazul persoanelor care au suferit un AVC (atac vascular cerebral) sau paralizii. 

EN:

The University of Minnesota has just announced the first brain-computer interface for controlling a robotic arm using thoughts. 

 

This is the first interface in the world able to coordinate a robotic arm using your thoughts without an implant (with electrodes). The tiny signals produced by the electrical activity of our brain are collected by a non-invasive, easy-to-use cap with 64 electrodes attached. 

 

The acquired signals are processed by a computer, using advanced techniques of signal processing and machine learning, and then transposed into movements.

 

In order to understand how a BCI (brain computer interface) works, one must first understand the electrical activity of our brains. 

 

There are approximately 30 billion neurons filling our brains: the nerve cells linked by axons and dendrithes. When we think, talk, move, feel etc, our neurons transmit tiny electrical signals which pass form neuron to neuron very fast (approx. 400km/h). These electrical signals are generated by differences in electrical potential (positive vs. negative) carried by ions on the membrane of each neuron.  In order to acquire these signals, a set of electrodes are attached to the scalp, through a system called an Electroencephalograph (EEG). The acquired signals are then interpreted by a computer using singal processing algorithms and machine learning and then translated into actions. That way the devices (e.g. a robotic arm) can be controlled. 

 

To deteremine where on the scalp the electrodes should be placed to measure a specific function, the subject is first put in an MRI (Magnetic Resonance Imaging). During such a session, when researchers are attempting to develop an BCI to control a robotic arm, they might ask the subject to move his actual arm. The MRI will show which area of the bain is active during the arm movement, giving them a clear target for the electrode placement. 

 

 

 

The signal created by the neurons are extremely small and, in most situations, the scalp reduces (attenuates) them even further. For that reason, the electrodes were implanted on the brain, under the scalp for a better quality of the acquired signals. Until now.

 

The researchers at the University of Minnesota succeeded to move in 3D a robotic arm with electrodes placed on the scalp (in a non-invasive manner). 

 

This invention promises to help millions of paralyzed people as well as persons with neurodegenerative diseases. 

 

The next step, the researchers have announced, is to apply this technique to move a prosthetic limb attached to a person’s body, and to persons who have experienced a stroke.