This website is a great resource for an intro of types of bionics: https://bionicsforeveryone.com/key-articles-on-this-website/ Summarizing key point here. #### Bionic limbs with brain-interface The artificial limbs that are controlled via neural signals are broadly called as myoelectric bionic limbs. #### Two-sensor myoelectric bionic arm These are currently the default standard. They have a hand that's capable of a few distinct types of grips (open, close, pinch and so on). Within the socket where the bionic limb is attached to the body there are two myoelectric signals. Myoelectricity, like [[Overview of brain input-output techniques#EEG|EEG]] is electric signals that generate from muscle activity and rise up to skin which are then read from an electrode. Currently, most bionic arms come with 2 sensors that are able to detect pushing or puling of two muscles on the limb. The user selects one of the grips of the hand by doing a combination of these push-pulls of the muscles. The difficulty of remembering these combinations and the chances of misfiring make such 2 sensor myoelectric bionic arms prone to [high levels of rejection](https://bionicsforeveryone.com/bionic-hands-finding-the-right-myoelectric-control-system/#). #### Pattern recognition, multi-sensor myoelectric bionic arms In such bionic arms, instead of two sensors, multiple sensors are used that make contact with the body/remaining limb. User undergoes a period of training where the user tries to do various grip movements normally by simply imagining them. The sensors generate a signal that an algorithm that uses to predict which grip or movement was intended by the user. Since there are more sensors, calibration of the device is important. If sensor moves during the day, proper signals could not be acquired and hence recalibration may be required. I feel this is where good algorithm could be made use of. #### Implanted electrode bionic arms Since skin conducts much less and electric signal attenuates, some bionic arms use electrodes directly implanted near the motor nerve (and transmit them wirelessly outside the body). This enables high fidelity of signal and no issues with sensors moving and recalibration since sensors are present within the body. #### Bionic arms with sensory feeling Most research shows that feedback from arms is important for faster acceptance and better accuracy and utility of bionic arms. Without a sense of touch, users rely on eyes only and hence aren't able to effectively do tasks that require delicate control such as squeezing sauce bottle, removing cherry step, etc. The latest in the field is to implant electrodes that stimulate sensory nerves to send sense of touch signal back to the brain. Research has shown that the mapping of hand is surprisingly well-laid out in neural pathways all the way from hand to spinal cord to brain, which means finding and tapping into nerves corresponding to fingers or palm may be possible (in fact, research has demonstrated that it is possible). <iframe class="signup-iframe" src="https://invertedpassion.com/signup-collector" title="Signup collector"></iframe>