Cognitive Neuroscience Lecture 6: Guest Lecture: Neuroprosthetics

L6: Guest Lecture—Neuroprosthetics

Recreating sight in blind patients; case study

Inspired by cochlear implant

Four opportunities along visual pathway: Retina, optic nerve, LGN, Visual cortex

Outer retinal degeneration: fewer retinal cells--Retinitis pigmentosa

-         So, we put a subretinal implant in the area of loss of rods and cones

-         Or, we put an epiretinal implant (easier for surgeon)

Retinal implants started at wilmer

-         Small probe in the eye

-         Held by surgeon, stimulates retinal cells

-         Patient awake; local anesthesia

New epiretinal approach: argus II retinal prosthesis system

-         Electrodes individually adjusted

-         6x10 retinal implants

-         Res 525 Mm

-         Camera, transmitter coil, video processing unit, electronics case, receiver coil

Many countries all over the world

Implant approved in Europe: Alpha IMS

-         Electrodes jointly adjusted

-         Res 80 Mm

Future: subretinal implant

-         Emphasis on extraocular transmission and processing away from stim site

-         Receiver coil and electronics on temporal sclera

-         Scleral tunnel to subretinal location

Future: German Epi-Ret projects—epiretinal implants

-         External camera

-         Wireless transmission

-         Antenna and electronics in the capsular bag

Stanford Optoelectronic Retinal Prosthesis Concept

-         Subretinal implant inducing retinal migration

US DoE Artificial Retina project

-         Novel electrode materials and light capture

-         Many types of electrode arrays

-         Light-to energy conversion using plant photosynthetic molecule PS II and genetic engineering

Neurotransmitter based retinal prosthesis (concept only, not real yet)

-         NT release targets inner retina

-         Specific pathway (on only or off only) excitation might be possible

-         Stim through microfluidic channels in chip

-         Testing in isolated retina prep

Optic nerve prosthesis

-         Emphasis on processing crude info

-         Electrodes in cuff around optic nerve; light/dark/direction/stim strength must be learned

-         Penetrating electrodes being developed

Concept of cortical prosthesis:

-         video encoder in glasses, electrode array in V1, transcranial interconnect in base of head, signal processor far below

Intracortical approach: NINDS, 1992

-         coil, ceramic, AIROF electrodes, counter + reference

Cortical complication: Scrambled map

-         electrodes à retinal percept?

-         Cortical percept magnification isn’t exactly in order

Near future: other active groups; human trials many years away

Planning for prosthetics: Simulations

-         Where did SSMP get the idea for 16 and 60 electrodes

-         First systematic explorations through simulation

-         Principal fields of investigation:

o Wayfinding/orientation & mobility

o Visual acuity/reading

o Face recognition

o Hand/eye coordination (visually guided action)

-         Many stim are woefully simplistic (crisp round dots) or needlessly hard (eccentric pixelized reading)

Route planning in real world

-         Pros

o Easy to create

o Useful for early trials

-         Cons

o Layout quickly learned

o Unnatural (why do that if cane/dog works just as well)

o Potential risk

Real-world wayfinding: results

-         Sighted subjects trained on pixelized viewing

-         Fast usually = sloppy

-         High res is better but not perfect

-         6 x 10 good enough

Hurdles for implant wearers: Stabilized view

-         Stabilized view

o Needed because eye movement is a lot of extraneous movement

o Subjects learn to hold eyes still

o Subjects learn to move camera

Modeling prosthetic vision in a nutshell

-         Filter software transforms camera image or virtual scene

-         Eye-tracking stabilizes raster location (optional)

-         Subject views scene in headset (monocularity)

Eye-head-hand coordination

Virtual mobility

-         Dots locked to eye position

Hurdles for implant wearers: Spatial filtering

-         Prosthetic vision in the operating room

-         Stim single electrodes (retina or cortex): perceived in correct location, flicker perceived up to 40hz

-         Stim multiple electrodes (mostly retina): crude shape scene (single row/col), blurrier than dots, hard to see flicker

-         Retinal rewiring: inner nuclear layer cells migrate and form clusters

-         Meander mazes; maze tracing

Hurdles for implant wearers: Temporal filtering

-         Maze tracing, temporal low-pass filter

Hurdles for implant wearers: Spontaneous bg activity

-         All prostheses, even ideal ones, give light shows

Argus II: moving shadows

-         prosthetic != normal

Lots of training needed

-         Performs line task better than without it

-         Finding objects and seeing movement: Square localization/direction of motion

-         Spot reading and object recognition

-         Orientation and mobility