Research types of BCI interfaces. How do they work differently with the brain? Explain an existing use for a certain type of brain-computer interface. What does it do now, what are its limitations, how would you improve it? Are there any social or ethical issues regarding BCI interfaces? How do you think they will be handled in the future?
I found learn.neurotechedu.com through a Google search and it provided a wealth of information regarding the definition and various types of BCI. The only difference that I noted was that they broke the types into three categories, non-invasive (EEG or MEG (not the shark movie)), semi-invasive electrodes are placed on the surface of the brain in the dura or arachnoid layer ( ECoG) and invasive micro-electrodes are placed directly into the cortex, measuring the activity of a single neuron. The electrocorticography uses a sheet placed on the brain’s surface to measure electrical activity. The reason for using this type of device is to obtain a clearer ‘image’ of the electrical activity in the brain because the signal does not have to go through the scalp.
The article mentioned multiple different BCIs that I had never heard of including MEG or magnetoencephalography, fNIRS or near-infared spectroscopy. Others PET or positron emission tomograhpy, fMRI or functional magnetic resonance imaging and EEG or electroencephalography I had heard of and some I have seen the results of and they offer very impressive images of the brain and its activity. According to the article the MEG is more adept at detecting high-frequency activity in the brain. the PET is used to detect different processes such as blood flow, metabolism, neurotransmitters happening in the brain. The fMRI detects changes associated with blood flow causing areas in the brain to ‘light up’ as the blood flows to those areas. While the fNIRS also measures changes in the blood flow it does so using infrared light versus the magnetic field in the fMRI.
Other than the ethical issues discussed by Nick the article also mentions informed consent, or the patient has to consent to the procedures that are being done on/to them. If a patient is locked in or some other form of severe degree of disability how can we ensure that true informed consent was obtained. The article also discusses the impact to society through any potential increase in abilities. Much like the television show The Bionic Man would not be ‘fair’ to have him compete in various athletic competitions due to his improvements, the same could be true with someone that has a new found ability to read minds or map the full use of their brain power, would that be an unfair advantage?
There were a couple of ideas that I thought were pretty interesting. The first was started by Facebook in April of 2017 to work on a 2 year project to use optical imaging for scanning the brain to detect the speaking in the head and translate into text. The goal is to allow people to type 100 words per minute, simply by typing directly from your brain. While this would make homework much easier, it probably does not rise to the concern of having your every thought typed out. The second area mentioned was by Elon Musk who has referred to a ‘neural lace’ which would be an AI layer that would augment the human brain’s abilities. The vision is to have an network of implantable electrodes to be able to learn from the brain and then begin to augment it.
I think both of these devices and the many other raise all sorts of legal and ethical issues, but much the way that people were resistant to having a phone on their person so someone could reach them anytime day or night, they may slowly erode over time. I definitely would not want anything that took whatever I was thinking and put it out for the world to see, of course from a learning perspective it might keep more students on task, so that could possibly be a win. I think any BCI devices will ultimately be measured by what it can do. What are the upsides? If there are more upsides than possible downsides, why would we not want to adopt it?
Meetha
BCI interfaces are basically meant to translate electronic signals between the neurons in the brain into signals that could be understood by external devices. They act as a channel to convey a person’s thoughts into assisting or repairing problems in the cognitive and sensory motor functions. There are three main types of BCI interfaces in use. 1. Invasive BCI 2. Partially invasive BCI and 3. Non-invasive BCI. Invasive BCI devices are buried in the grey matter of the brain and they provide high quality signals. Partially Invasive BCI devices are placed inside the skull but not in the grey matter. The non-invasive devices are the safest BCI interfaces in use so far and they track signal from the brain through an external device such as a skull cap or a headset. The electrodes on the skull cap translates the signals from brain activity to external devices. This is used by patients with difficulty in their motor ability to be able to help them with their motor capabilities and also for communication. EEG (Electroencephalography) is a type of Non-invasive BCI that is widely used and it is comparatively less costly than other devices. The common signal that translates the signals from brain to external devices in EEG devices in called the P3000 wave. EEG based devices are used for motor strengthening, stroke rehabilitation, gaming, virtual reality and one application that I found interesting was the ‘Speller’. This system helps a person with motor impairments to type things based on their thoughts. I think most of the BCI based devices does research on people with some sort of disability or sensory system issues. I would do research on healthy people as well with the BCI devices, especially the speller. Also, design wise, I would make it look better because if I was given a skull cap with too many electrodes and wires running around me, my first thought would be,”Oh, its too scary”. Ethically, I would say the same thing as I wrote in the other blog for this session. I would foresee ethical issues rising up when there is more deaths occurring due to clinical trials with BCI systems (Specially with Invasive BCI devices). So it would be better to stop any clinical trial if there are too many deaths during the trial.
– Research types of BCI interfaces. How do they work differently with the brain? Explain an existing use for a certain type of brain-computer interface. What does it do now, what are its limitations, how would you improve it? Are there any social or ethical issues regarding BCI interfaces? How do you think they will be handled in the future?
There basic 3 different categories of BCI devices: non-invasive, semi-invasive and invasive. Non-invasive ones use sensors on the hair/head typically measuring electrical potentials, EEG or MEG signal. Semi-invasive are placed on the exposed surface of the brain, usually through a surgical opening and measure ECoG signals; and invasive are micro-electrodes that are placed directly into the brain cortex, measuring Intraparenchymal signals or the activity of neurons.
One interesting application was in the category of cell-cultures BCI in which Dr. DeMarse, a professor of biomedical engineering, extracted 25,000 neurons from a rat’s brain and developed a “brain” in a dish. He arranged the neurons over a grid of 60 electrodes in a Petri dish. The cells apparently started to reconnect themselves from very small interconnections. Then he somehow linked it to an F-22 jet flight simulator in a two-connection. Gradually the “brain” learned to control the flight of plane based on information it received about flight conditions. There are definitely social/ethical issues when we are talking about creating living neural networks or brains. But this kind of research can help us to understand how the brain does computation. What is scientists were able to extract human brains and keep them alive and then hook up BCI interfaces to them?
I have done a lot of research on different BCI interfaces. In fact I use two different models in my own artistic practice, one of which I built myself. You might guess that the one I built is more primitive than the commercially available product that I hacked to my own purposes, but that would be wrong. My first BCI was a MindWave controller, one of the video game interfaces mentioned in the lecture. It functions with one skin conductance electrode on the forehead, and I modified it to send its raw output data to an arduino so that I could access, interpret, and use the data stream for something a little more interesting (at least to me) than levitating a ball on a screen. My second BCI device is an OpenBCI Cyton with a Mark IV (3D printed!) headset. In this case, there are up to 16 dry electrodes (the same that are used in the caps shown in the video and in medical applications) with an onboard processor which sends the output to an RF receiver on my laptop. I use open source software (created with Processing) to monitor my brainwaves, as well as to analyze, record, and share them over a network. For my own purposes, I use these signals (in so far as I am able to interpret them) to create music, or sound, if you are picky about music.
I am designing my sound algorithms based on what I can deduce about my brain activity, and the relationships between the different states and different patterns in the data. What was said in the lecture about the inverse mapping problem, well… the struggle is real! I basically have next to no formal training about how to read EEG, but that is not for lack of trying. The biggest issue is that there are really no overarching standards on how to read EEG. Some professionals develop a knack for seeing patterns, similarly to how some people can read the chemical signatures compounds through h-NMR spectroscopy, except much less exacting. Part of the problem in learning this is that every brain is different, and patterns that may look normal for me may represent extreme pathology for another, and vice versa. Also, since I’ve never had a professional analysis of my own EEG profile, I don’t know if I am ‘normal’ anyway!
So while I can come up with a few algorithms to support the mathematical relationships and ratios that represent the scope of the data, I really can’t do much to use the activity itself to “do” things. With one exception, I have learned how to create a stable alpha wave, and I can use this to control volume, pitch, etc. in the live sound generation, but it’s sporadic enough to be an arbitrary parlor trick unless used in the right context. Otherwise, I have to rely on modeling the generalities, and then wait for my brain to do something that I can hear as distinct through the randomness.
Here is a link to an early iteration of the EEG data to sound: http://aliciachamplin.cartographile.com/new-sonification-strategies.
The mature version (https://www.youtube.com/watch?v=XRCW9VB0mPI) has a much more harmonic output, which actually responds to different states of mind after much practice. The changes in the sound are very, very subtle, but with time you can hear them. The audience in this case hears the much more noticeable changes that are a result of the signal amplitude changing when someone gets close to me, so they can interact like they were playing a human theremin.
So with this summary of my ‘existing use,’ I’ve also described a bit of the limitations, which to me are couched in the fact that the data itself is fairly opaque to me, and that professional expertise for this research is fairly inaccessible. In looking at ways to improve this, I see that it may not be an issue for much longer. I suspect that machine learning will be the next phase for this analysis, and indeed it is the next step for me, hopefully. With the ability to do further onboard processing through machine learning and neural nets, the mappings will hopefully become clearer and more intuitive.
I touched on some of what I think are important ethical issues in the lecture discussion (participation blog) but will reiterate, that I think the issue of inequality and access is key. That is why I am such a fan of open source biotechnology. Without open access, the research will be closed, proprietary, and divisive, whether that be in the form of creating a super citizen of some kind or in the sense of monitoring or control. Either way, open access and low barriers to participation are the main thing that can augment the ethical pitfalls that are inevitable with this kind of technology. It is my *hope* that in the future advancements will continue to be made through more accessible and open source research by individuals and collectives, and that there would be considerable dialog about these concerns and well informed legislation that keeps up with the technology to prevent the worst outcomes. But if you ask me what I *think* will happen, it probably looks more along the lines of the way we botched net neutrality and online privacy. What is the internet if you subtract all the proprietary infrastructure? Not much, so we use it anyway, and we are subject to a death by a million cuts, to borrow a phrase.
The device I researched is one that uses an electrode array to send signals to a tablet device that allows users that do not have the use of their limbs to be able to use the device to do simple tasks such as listen to music, browse the web and even chat with friends. It is currently limited by the screen capacity and its inability to scroll properly for the users, though they are looking into allowing for this in the keyboard design in the future. The paper and corresponding documentation stated that the users were able to use their ‘intentions’ to move the mouse in a certain direction and to indicate the click command, at this time, further commands are not viable. Socially, allowing these people access to something they may never have had access to before, this is an amazing opportunity and in my profession we are always looking for ways of getting people to be as independent as possible for as long as possible. As for ethically, people are still having debates on if ‘Google Home’ is ethical and that is being used profoundly to help people access their home easier. Technology will continue to advance beyond what people are ready for, it is a matter of introducing it safely and with trust in mind.
BCI interfaces can range in how invasive they are. Prior to this assignment, I was only aware of the non-invasive ones, where electrodes are attached to the scalp or other external locations. The I learned about the very invasive ones which are implanted directly into the brain or spinal cord, and then also about the partially invasive ones, where they are in the skull, but still on the outer membranes of the brain matter. I’m particularly interested in how these interfaces are used for things like controlling machines. One instance I found from Wikipedia was the case of Johnny Ray, who was suffering from locked in syndrome due to a stroke. They then implanted him with one of these interfaces, and he eventually learned to move a cursor on a screen with his mind.
I think this sort of application can bring up a plethora of social/ethical issues. If we can help aid people with disabilities or disease, we have a certain responsibility to try and help them. But then what if that technology also leads to the development of weapons, or tools that can be used to inflict harm? There is also the ethical concerns with how these devices are tested, particularly with regulations surrounding human test subjects.
In the future, I think we will continue to see the technology improve, with a higher adoption rate, and I think regulations will begin to appear restricting what the technology is allowed to do, and how it can be used. For example, it wouldn’t surprise me if regulations required protections to be in place to prevent cybernetic prostheses from being hijacked.
The largest segments of devices I noticed were devices meant to utilize communication paths already created by the brain, or to manufacture new ones from different data from the brain. An example of utilizing an existing pathway would be passing visual information into the optic nerve in the protocol that the brain has evolved to recognize. Similarly, passing electrical signals to stimulate muscles in the method the brain already does. Alternatively, the researchers are creating new interactions that the brain was never meant to do, such as monitoring the brain activity with the cap in the video to fly the plane. There are potentials for mixed versions of this as well, in the case of something like the prosthetic arm. In many of these cases they are utilizing as much information as they can about how the brain would have controlled the arm, but to limit invasiveness don’t necessarily use the same pathways the brain uses.
An interesting research example I found of an existing BCI device (https://www.nature.com/articles/scsandc201621) is one that helps with rehabilitating from neuropathic pain. The users sensorimotor rhythm is monitored using EEG, and the visual stimulation is given based on the users intention. This helps alleviate pain caused by a sensory deprivation which is very common to those with spinal cord injuries.
I remarked on some ethical issues in the participation blog, but one I did not mention there was privacy of the BCI information. The study I just mentioned shows how a users pain can be modified by monitoring the brain activity and giving correct stimulus. This same effect could be used to target propaganda for a user so that is presented in the most convincing way, target people for certain types of services, or even detect a persons identity over the internet.
I would like to think that in the future the data that is shared beyond the direct connection of BCI will be extremely limited, and protected by systems such as HIPAA so it cannot be directly monetized, but as social media has shown, the average person does not seem to value, or does not understand, the potential impact of their personal information. I think due to this legislation would likely be enacted to limit the obviously scary scenario’s, such as mind control over the internet, but I am not as sure if effective legislation could be enacted to protect the user from providing unwanted information.