Brain Computer Interfaces: Background and Current Issues
We’ll first take a look at our nervous system, and then we’ll explore three methods to monitor and modify brain signals (BCIs). Afterwards, we’ll talk about collective intelligence.
Calm Down Nervous System. It’s ok
We’ve got to know what we’re working with, so here is a quick overview of our nervous system.
There are three main types of neurons: sensory (input), motor (output) and the ones in your brain (there’s also the ones that carry the signals, but let’s stick with these for now).
General Anatomy of a Neuron
Neurons are cells, so they have a cell body that facilitates their vitals.
Dendrites: these are the receiving ends.
Schwann cells (also called myelin sheets): insulation, made out of fat.
Axon terminals: output
How Neurons Communicate
Axon terminals of the signalling neurons are connected to dendrites of the receiving neurons. The space between them is called the synaptic cleft. When the transmitting neuron is excited with an electric signal, it releases neurotransmitters, which in turn increases or decreases the probability of the receiving neuron becoming excited.
Hippocampus: the learning center of your brain
This is where your short term memory is processed and saved in your long term memory. During this process, certain electrical signals are emitted by the body. If you record these signals and inject them into the hippocampus of another being (of the same species), they’ll experience the same process, resulting in them having your long-term memories. We can’t create these signals or understand them yet (our brains can, but our computers can’t).
Brain Computer Interfaces
This is what’s used to record and inject the electrical signals in your brain. This is not restricted to your hippocampus (therefore your memories) but applies to any group of neurons in your brain (such as hearing or seeing). There are many methods of interfacing computers with the brain, but they come in two categories: invasive (actual stuff put into your skull) and non-invasive (you wear some device over your head). Let’s take a closer look at a few.
EEG
A bathing cap with fancy electrodes over it. It measures the electrical fields in and around the brain and can tell which regions of the brain are working.
Some games can be played using different areas of your brain. (http://neurosky.com/2015/09/eeg-games-top-5-list-playing-with-your-brainwaves/)
ECOG
This is not too different from EEG, except that it is placed on the brain and rather than detecting electrical fields, it detects actual electricity from the brain. An advantage of this approach is that you can directly affect the brain by sending signals, unlike EEG.
This is not an immensely customer-friendly solution.
Neural Dust
This is a newer approach being explored. The idea is to have tiny sensors and embed them over synapses. The sensors are too large for the neurons in the brain, but they have successfully worked on a rat, with some side-effects (like the body rejecting the sensor on some instances).
The sensor has a piezo crystal, that converts electricity to ultrasound. It’s covered with surgical-grade epoxy. The main problem with that is that it degrades over time.
Neural Code
There are many problems, large and small, standing in the way of interfacing our brains and computers. Here is a large one.
We can’t understand the messages transferred in neurons. This is called the problem of neural code. If we precisely understand how messages are transferred in neurons, we can create messages. A potential application of this might be training certain neural paths, meaning we can learn without studying.
Dr. Behrmann (of Carnegie Melon) was able to create a model that reconstructs faces from neural activity (MRI) of people who have seen these faces before, in 2016. Her model used precise, yet statistical data about what parts of the brain images of faces have activated.
There also exist games that one can play with neural activity, such as the pong game made in Santa Luzia foundation’s Neuroelectrical Imaging and BCI Lab in 2008.
I believe that the solution to this problem lies in a deterministic approach, a system that can monitor all of the neurons in a given section of the brain and their action potential (concentration of Sodium and Potassium ions). A potential solution might be programmable nanomaterials, materials that can react with sodium and potassium ions and emit ultrasound (like piezo crystals, inspired from neural dust) or otherwise emit some signal.
Collective Intelligence
This is a system in which you don’t even have to download the stuff you need. You know everything ever known to humanity. There are many problems with this approach. For starters, what if the collective intelligence company sells your ideas to Donald Trump? I foresee three ways this unfolds.
Good
We share all the information. So your boss knows you weren’t sick the day you called in sick. This is scary. I believe this approach will significantly improve humanity by improving our tolerance against each other, and there’d be no way to lie. In this approach, you don’t pay, it’s the world’s data for yours.
Better
Everyone hosts their knowledge in a distributed fashion (works well with blockchain). This means that you don’t have to pay, and you are mostly in control of your knowledge. Think of this as a personal website. You decide what to put up.
Oof to humanity
Another Facebook case. You don’t pay, but the serving company plant ads into your brain without you noticing. This is the most horrible outcome in my view.
Key Takeaways
- We have the technology to inject memories, but we can’t understand or create memories from these neural signals.
- There are many different BCIs. We are still exploring newer options.
- Neural Code is how our brain functions on the neuron level.
- Collective intelligence is like one big brain that our brain is connected to, sort of like the servers in the 1960s with many workstations connected to one computer.
