May 7th 2021

I Think, Therefore I Can: Considering the Brain-Computer Interface


Even the most advanced “smart” technologies are only as intelligent as their algorithms allow. Yes, devices can now take our pulse, monitor our heartbeats and even respond to verbal commands, but they can’t react to what’s going on inside our heads.


This is the promise of the brain-computer interface (BCI), a cognitive connection between human and machine that allows direct control over physical processes using nothing but our thoughts. And while there’s no chance of connected individuals getting misclassified as computers by the Turing Test just yet, researchers in this brain-bridging battle have made significant strides over the past few years in blurring the lines between mental efforts and digitally enabled physical outcomes.

Brain Connection Basics

What exactly is BCI? A study from the Mayo Clinic explains: “Brain-computer interfaces acquire brain signals, analyze them and translate them into commands that are relayed to output devices that carry out desired actions.” However, these interfaces don’t leverage normal neuromuscular pathways, meaning users must master a new set of mental skills to operate BCI-controlled devices.

There are two basic types of these devices: invasive and non-invasive. Invasive solutions require medical professionals to implant electrodes or other control devices directly into the human brain to connect with specific neural pathways. Non-invasive options typically use sensors on the head or neck to collect, record and interpret brain activity. Non-invasive technologies are already used in devices such as EEGs, MEGs and fMRIs.

While invasive interfaces have the edge when it comes to accuracy and reliability, they also come with greater risk and higher costs. Non-invasive options are much simpler to use, but they can’t deliver the same level of functional fidelity. That said, the lower bar for entry coupled with the falling costs of micro-scale technology has prompted many tech firms to invest significantly in hopes of getting in on the ground floor of brain-computer connections.

Current Connective Capabilities

In 2019, the Braingate project from Stanford University showed that BCI devices could allow tetraplegic patients to control commercially available tablet devices using only their minds. Meanwhile, more recent work from Johns Hopkins Medicine (JHM) and Johns Hopkins Applied Physics Laboratory (APL) made it possible for paralyzed patient Robert “Buz” Chmielewski to cut and serve his own food using two simultaneously controlled robot arms, as FreeThink reports.

By implanting two sets of electrodes into Chmielewski’s brain — one in each hemisphere — and augmenting them with AI, the teams were able to overcome previous limits of one-limb control offered by single-electrode setups. Worth noting: The road from initial connection to eventual precision wasn’t quick or easy. While the implant surgery itself was performed in January 2019, it’s taken Chmielewski two full years to gain the mental muscle required for mechanical mastery. Nonetheless, both projects show promise as a way for paralyzed or differently abled patients to regain some level of specific physical control.

Mind the Gap

Not surprisingly, the benefits of this connected brain technology are matched with potential drawbacks. Privacy is perhaps the most pressing.

As Harvard Business Review notes, the use of BCIs at scale introduces questions around who owns the data generated by human brains and what companies might do with this data once it’s in their possession. Could companies compel workers to use non-invasive devices for workplace productivity monitoring?

Connective compromise is also a potential risk. Consider a construction firm using advanced, non-invasive devices that allow staff to remotely and precisely control heavy machinery. What happens if hackers are able to hijack these BCIs and cause damage to people or property? Not only does this pose a significant safety risk, but it also creates a legal snafu when it comes to determining responsibility. Is the user at fault for failing to check the device, the manufacturer for not installing effective safeguards or the company for compelling staff to use BCIs?

Operational bias is another cognitive concern. As Brookings notes, AI tools often come with built-in, unconscious bias tied to the type and volume of data used to train their foundational machine-learning algorithms. For example, AI hiring solutions have demonstrated a bias for specific candidates because the data they were fed lacked variety. In the brain-computer field, limited biometric data sets could result in device interfaces that are more difficult to use for some employees and easier for others, in turn impacting performance and potentially job security.

The Future of Technical Telekinesis

So, what’s next for the brain-computer interface? First is increased mastery of “intentional control” — the ability of connected devices to map mental intentions to physical processes. As a recent Big Think article shares, ongoing work with BCI-outfitted mice has already shown a strong connection between the visual and anteromedial cortex, which corresponds to the parietal cortex in human brains. AI algorithms also need to work to help streamline the strengthening of the mental muscles required to convert thought into action. And before BCI tools really go mainstream, there’s a need for substantial standardization and regulation around the data collected, used and stored by these devices.

At scale, there’s massive market opportunity here driven by truly smart devices connected directly to human brains, empowering users to complete everyday tasks — from turning on the lights to opening the garage door or activating the alarm system — with a single mental command.

I think, therefore I can? It’s fascinating food for thought.

Check out Northrop Grumman career opportunities to see how you can participate in this fascinating time of discovery in science, technology, and engineering.