File:Towards large-scale bidirectional neuroprostheses for memory disorders.webp

English: Future neuroprosthetic systems for memory restoration may benefit from distributed multi-electrode implants targeting large-scale brain networks, combining depth electrodes, ECoG grids and potentially novel electrode technologies for combined recordings and stimulation. Field potentials measured by these implants would then be sent to computing devices via high-channel-count recording systems. These computing devices would extract in real-time specific spectral features of these raw neural signals, such as the instantaneous power, phase and phase-amplitude coupling of theta and gamma oscillations in hippocampal structures. These features would then be used as control signals in closed-loop algorithms that deliver spatially, temporally and spectrally patterned neurostimulation protocols to enhance the neural features of successful memory processes (encoding or retrieval). These complex stimulation patterns would next be applied to the distributed neural implants via high-channel-count neurostimulators. Neurostimulation would finally affect neuronal activity through mechanisms that depend on the geometry of local fibers, which can be explained or theoretically predicted by computational models of neurostimulation. Schematics of high-channel-count recording and stimulation systems were adapted with permission from Blackrock Neurotech.