DO CÉREBRO PARA A MÁQUINA: SERÁ? (translation) Scientific American: Mente & Cérebro.


For scientists it will be necessary to understand how the brain generates movement before the brain can be used to transfer information to a mechanical device to restore motor function to quadriplegic patients. Will it be possible for a quadriplegic child equipped with a mechanical exoskeleton to use signals from his brain to kick a ball to open the World Cup? This idea, although appearing very attractive, has generated much controversy. There are five fundament-al points that need to be discussed to assess the scientific validity of this idea: (1) The amount of information that can be transferred by the brain to move external devices. (2) The transfer of information as the number of nerve cells producing the signal is increased. (3) The neural code upon which brain-machine interfaces (BMIs) are based. (4) The ability of individuals to use brain activity to move devices in the absence of body movement. (5) The effect of eye movements and vision on the BMI signal. We conclude that further re-search on understanding how the brain generates movement is necessary before BMI can become a reasonable option for paralyzed patients. Given the current performance of BMI devices it is quite likely that the World Cup event on BMI—if realized—will be more de-monstrational than effective at recovering the motor functions of the child. The goal of BMI is to extract as much information from the brain to operate external devices at the same level of performance as is normally done by nerve cells that drive the muscles of healthy people. One way of comparing the performance of BMI is to document the amount of information that is transferred by the brain to machine over a specific period of time. One bit of information represents the ability to produce a yes/no response at a 100% performance level. If a 100% per-formance can be produced once per second then the information transfer rate is 1 bit per second. As we drive our cars we are hopeful that we are performing at 100% when we encounter red versus green traffic lights. The most common paradigm used to study BMI is the center-out task (Fig. 1B, inset). Here a person or monkey is required to use his hand to move a lever starting at a central location. This lever is yoked to a cursor on a computer monitor that is positioned in front of the subject. A small visual target is presented near the perimeter of the monitor and the task of the subject is to acquire the target with a centrally-positioned cursor by moving the lever in the correct direction.