Biohybrid neural interfaces

A "head in a jar" scenario, while preserving biological brain function, lacks the means to provide a meaningful quality of life. To enable a fulfilling existence without a biological body, sensory perception, communication, and mobility must be integrated through neural interfaces and advanced robotics. These systems will act as extensions of the brain, enabling interaction with the external world and replicating the functions of a biological body.

Brain-computer interfaces (BCIs) represent the foundation of neural interfacing, but their current capabilities are limited by low bandwidth, insufficient spatial resolution, and challenges in long-term biocompatibility. For high-performance applications, BCIs must capture detailed neural activity from densely packed regions of the brain, such as the motor cortex, and translate these signals into commands for external devices. Current electrode arrays, such as the Utah array, offer limited resolution and durability, while non-invasive methods like EEG lack precision. Emerging technologies, including thin-film flexible electrodes and optical interfacing with calcium imaging, aim to bridge this gap by providing higher resolution with reduced tissue damage.

Recreating sensory experiences, particularly vision, is a critical challenge. Current visual neural interfaces, such as retinal implants and cortical visual prosthetics, operate with a resolution of approximately 90x90 pixels, enabling basic shape and motion detection. The human retina, by contrast, achieves an effective resolution of approximately 7000x7000 pixels, driven by the dense and efficient arrangement of photoreceptor cells. Achieving similar performance with neural interfaces requires a combination of high-density electrode arrays, improved signal processing, and direct integration with retinal or cortical neurons. Biohybrid approaches, which incorporate engineered photoreceptor cells with electronic systems, are a promising avenue for bridging this gap. Long-term biocompatibility remains a significant challenge, as electrodes and electronic components degrade in the body or provoke immune responses, limiting their lifespan and effectiveness.

Another critical requirement is enabling neural interfaces to support both motor output and sensory feedback to facilitate seamless integration with robotic systems. Closed-loop systems that include bidirectional communication between the brain and external devices are under active development. For example, intracortical microstimulation (ICMS) techniques provide sensory feedback by delivering targeted electrical signals to sensory regions of the brain, simulating tactile or proprioceptive input. Advanced systems will need to integrate multiple modalities, including touch, pressure, and temperature, to replicate the full sensory experience of interacting with physical objects.

The integration of robotics with neural interfaces presents additional technical demands. Robotic systems must process neural commands in real time while providing precise, naturalistic movement. This requires the development of adaptive control algorithms that learn and adjust to the user’s neural signals. Additionally, robotic components must incorporate sensors to relay tactile, force, and position data back to the neural interface, ensuring the user perceives a coherent and responsive system. Current robotic arms interfaced with BCIs, such as the DEKA arm, demonstrate limited functionality but show promise for more advanced systems.

Addressing the unique issue of brain health when disconnected from the spinal cord is essential for long-term viability. The severance of the spinal cord leads to a loss of critical feedback loops that regulate brainstem functions, including autonomic control. It also leads to progressive nerve dieback, which extends to the brainstem and poses a severe risk of mortality. Neural-computer interfaces must replace these lost pathways by directly interfacing with brainstem regions to maintain physiological homeostasis. This requires the development of interfaces capable of recording and stimulating neurons in these regions without causing further damage or disruption to surrounding tissues. Advanced neural recording techniques, such as dense optogenetic arrays or multi-modal probes, may provide the resolution and functionality required for these applications.