Seminar on Non invasive Brain-Machine Interfaces


  Brain-Machine Interfaces.pdf (Size: 1.17 MB / Downloads: 3)
Introduction

Brain-Machine Interfaces (BMI) or Brain-Computer Interfaces (BCI), also referred to as Neuro-Prostheses, are conceived as technological interfaces between a machine (usually a computer) and the brain of a user. They should permit the use to perform a certain task, usually without implementing any motor action. This implies that neural impulses generated by the user’s brain are detected, elaborated and utilised by the machine, approximately in real-time, to perform definite tasks. As an example, information can be processed and employed to control mechanical systems (e.g. actuators) or electrical devices (e.g. electronic equipment).

Brain-Machine Interfaces (BMI): a brief overview on the state of the art

A great deal of efforts in neuroscience, robotics, and computer science are today spent by many research groups to develop BMI. In order to provide a glance at this vast field, it can be useful to mention here at least some relevant examples [1-24], briefly reported below.

Non-Invasive Brain-Machine Interfaces

Experiments in humans utilizing modern invasive and non-invasive brain imaging technologies as interfaces have been conduced. The most commonly studied potential interface for humans has been electroencepalography (EEG), mainly due to its fine temporal resolution, ease of use, portability, and cost of set-up. However practical use of EEG as a BCI requires a great deal of user training and is highly susceptible to noise. In 2004 scientists of the Fraunhofer Society utilized neural networks to shift the learning phase from the user to the computer and thus recorded noticeable results within 30 minutes of training. Magnetoencephalography (MEG) and even functional magnetic resonance imaging (fMRI) have both been used successfully as rudimentary BCIs, in the latter case allowing two users being scanned in real-time to play Pong against one another by altering their haemodynamic response through various biofeedback techniques.

Final remark

Some examples of potential space applications for BMIs and other non-invasive man-machine interfaces have been considered. It is worth noting that for all these examples it could be useful to combine the action of a BMI with another non invasive interface. In fact a BMI could require and may benefit from an auxiliary system to be used for specific tasks. This is the case of multi-task operations, which require several efforts from the astronauts. The challenge is to properly combine these technologies, by making the system robust and intelligent. Some tasks may be performed by using signals detected by a BMI while others, at the same time, may be accomplished by exploiting alternative means of communication enabled by different types of non-invasive interfaces. Fulfilment of such issues may open completely new approaches to manage space operations.