A Modular Brain Computer Interface System of Wireless, Fully Implanted, Mechanically Flexible, High-Channel-Count Subdural and Depth Brain Electrode Arrays

Introduction: Current brain computer interface (BCI) technologies suffer from tradeoffs between the density of electrodes, ability to sample from broad networks, and limitations on data transmission and power. Higher channel counts require percutaneous connection to external equipment or are limited by battery-powered, implanted electronics. We developed a modular BCI system of fully implanted, wireless, high-channel-count intracranial electrode arrays.

Methods: In the Bioelectronic Interface System to the Cortex (BISC), a fully implanted custom integrated circuit thinned to less than 20 µm supports dense recording and stimulation with wireless powering and high-throughput telemetry through an outside-the-body wearable relay station. Flexible, customizable polyimide extender depth or surface arrays are bonded to each implanted circuit with support for up to 1,024 channels per implant. We performed initial feasibility and safety testing in a porcine model.

Results: Thin-film subdural arrays were implanted over the sensorimotor cortex and depth electrodes from sensory cortex to thalamus in 4 animals. Somatosensory evoked potentials (SSEPs) were recorded during peripheral stimulation after complete wound closure in multiple sessions up to 14-weeks after implant with no signal degradation. For subdural arrays, mapping resulted in clearly separable clusters between peripheral stimulation sites. Decoding performance was quantified using a linear discriminant classification model that resulted in an overall accuracy of 0.96. Depth electrodes recorded SSEPs with appropriate latency between thalamus and cortex. Duplex recording and micro-stimulation at a single site yielded large-scale network activity across the array which persists over several seconds, with an excitatory phase followed by inhibition before a return to baseline. Histopathology revealed mild microgliosis.

Conclusion : A modular BCI system consisting of a fully implanted custom integrated circuit, customizable polyimide extender arrays, and wireless external power and telemetry can support high-resolution recording and micro-stimulation. The system has applications for neuroprosthesis, and network recording and neuromodulation for epilepsy and other neurological and psychiatric disorders.