Skip to content

Sensory Writing and Bidirectional BCI

The first eight chapters focused on "reading the brain" — extracting signals from neural activity. This chapter turns to "writing the brain" — stimulating the cortex to write back touch, proprioception, vision, or even memory fragments. This is the critical transition for BCI from "remote-controlled robotic arm" to "embodied prosthesis": without sensory feedback, the user must constantly watch the arm to check whether an action completed; with sensory feedback, the prosthesis truly becomes part of the body.

Why writing is necessary. Anything that humans do without looking — picking up an egg, holding a coffee cup, buttoning a shirt — depends on real-time proprioceptive and tactile feedback. The Pitt arm (Chapter 06) can complete visual tasks but cannot reliably grip fragile objects, precisely because it lacks the writing side; once Flesher 2016 added ICMS-based tactile feedback to the Pitt team's setup, grasp time dropped from 20 s to 10 s — direct evidence of the value of closing the write-in loop. This chapter also connects to the visual prosthesis (V1 phosphene writing) at the end of Chapter 08, completing a three-way picture of "writing motor / writing vision / writing memory."

Recommended reading order. Read the four sections as "touch → bidirectional → memory → safety": start with ICMS and Somatosensory Feedback and the Flesher 2016 / 2021 Sci Transl Med lineage, to see how intracortical microstimulation can restore touch; move on to Bidirectional BCI and Channel Separation and digest Ganzer 2020 Cell — the algorithmic challenge of reading movement and writing touch in the same M1 patch simultaneously; Memory Prosthesis turns to the Hampson / Berger hippocampal MIMO model, the most controversial and technically deepest write-in frontier; finally, Neural Stimulation Safety imposes engineering bounds on every stimulation scheme (charge density / long-term stability / tissue damage).

Chapter contents:

评论 #