5 Minute Healthtech Jargon Buster: Brain Computer Interfaces

BCIs convert electrical (brain) signals into commands that can control external devices. The main objective is to replace or restore motor function that has been lost by a patient  in their limb(s). They essentially act as a bridge between the brain and the affected limb, bypassing damaged neural pathways, facilitating external device control and communication.

 BCI systems consist of four components:

1.         Signal Acquisition

2.         Feature Extraction

3.         Feature Translation

4.         Device Output

These four components facilitate the translation of brain signals into mechanical actions performed by output devices.

Signal acquisition is the measurement of brain signals using a sensor such as scalp electrodes or functional MRI (fMRI).

Feature Extraction is the analysis of the signals to differentiate signals related to an individual’s intent from other signal types.

Feature Translation is whereby an algorithm is used to convert the electrical signals into the corresponding commands for the output device.

Device output is when the device has taken the commands and performed the necessary operation. This closes the control loop [1].

Depending on the sensor modality, different types of activity can be recorded. For example, if an electroencephalogram (EEG) is used for signal acquisition, this would record the electrical activity of the brain. Meanwhile, if an Electrocorticography (ECoG) is used whereby electrodes are placed directly on the exposed surface of the brain, this would record electrical activity from the cerebral cortex, the part of the brain responsible for higher level functions like language [2].

The Cortical Communication (CortiCom) System is an example of a BCI. Clinically trialled by John Hopkins Medicine, it consists of 128 electrodes surgically implanted on the surface of the brain [3]. The CortiCom device records and carries brain signals to be processed into computer commands, text or speech. Such device would prove useful in patients suffering from communication impairments such as those suffering with muscle weakness due to conditions like amyotrophic lateral sclerosis (ALS).

Neurorehabilitation:

Common consequences of stroke include full/partial paralysis on one side of the body and impaired speech. By employing BCI intervention during both the subacute and chronic phases (1 week to 6 months post-stroke) of stroke rehabilitation, this can aid in a patient regaining motor ability. When a patient attempts to move a paralysed limb, their brain signals would be acquired and translated into commands in a virtual reality space, for example. This would help the brain to relearn movement patterns and promotes motor function restoration. BCI interventions have also been used to induce neuroplasticity of the brain (that is, the brain's natural ability to reorganise itself by forming new connections), highlighting the potential of BCIs in rehabilitation to improve motor outcomes [4].

Direct Control of External Devices:

BCIs can be used by patients to control assistive technologies to enable communication or control of objects. Patients suffering from conditions such as ALS typically present with symptoms such as muscle weakness and speech impediments. Use of BCI systems can enable them to communicate and interact with their environment via output devices such as word processing programs and robotic actuators [5].

Invasive

Some BCIs such as the Corticom that involve the implantation of electrodes on the brain surface hold potential issues. It is a highly invasive brain computer interface that holds the risk of infection, inflammation and long-term immune responses [6]. Moreover, if the BCIs are to be used for long periods of time, they need to be able to maintain high levels of safety, functionality and stability in the long term.

Pre-Clinical Testing

Monkeys are used in preclinical testing of BCIs due to being the closest relative to humans, and this enables the assessment of BCI performance and safety before testing on humans in clinical trials.  However, issues lie in preclinical testing phases involving BCI technology. For instance, in 2023 the FDA discovered quality control issues in Neuralink’s brain computer interface system studies [8].

Specialist Intervention

Stroke rehabilitation, especially one that employs ECoGs  as the sensor, requires the presence of a multitude of specialists including: neurosurgeons, nurses, occupational therapists, psychologists and many more. Hence, the patient would require frequent medical monitoring due to the invasive nature of the BCI.

Accuracy

In comparison to ECoGs, EEG are considered less accurate as the brain signals detected are weak and of limited frequency range [7]. This poses as an issue in the accurate translation of brain signals to output devices and could create disparities between user intention and the resulting output.

Artificial Intelligence (AI) can be used to enhance the accuracy and efficiency of brain signal decoding. AI-enhanced actuators in BCIs can aid in expediting information transfer rates, narrowing the gap between signal acquisition and device output [9]. Moreover, AI and machine learning algorithms can identify patterns in brain activity that correspond to specific thoughts prompting faster BCI performance. The pattern recognition and signal processing facilitated by AI has extended the application of BCIs including cursor control, auditory sensation, limb control, spelling devices, somatic sensation, and visual prosthesis [11].

An example of the application of AI in BCIs is the restoration of communication in cases of complete locked-in syndrome (LIS). This is a neurological condition whereby individuals are fully paralysed but are conscious and have full awareness of their surroundings. AI and machine learning are hence leveraged to enhance patient quality of life by decoding brain signals into text, for example, allowing these patients to express their thoughts and interact with the world [9]. In addition, AI-enhanced targeted electric brain stimulation is integrated into clinical BCIs which allows for bidirectional communication between the brain and computer. Such system improves BCI efficacy in restoring cognitive function. This has been found in research on epilepsy patients whereby the use of brain computer interfaces was able to improve their specific brain functions related to self-control and mental flexibility [10].

Overall, AI enhances the accuracy and efficiency of BCI systems, advancing its ability to decode and analyse neural signals.