Second Generation Setup

part of In Vitro System Additional Documentation by aegrumet@alum.mit.edu

Features

Single unit recording through one conventional needle-shaped electrode,
Stimulation through a flexible micro-fabricated array,
Some form of optical system for probing ganglion cell receptive field centers,
Computerization; the computer will generate the stimulus pulses and store the response waveforms,
Two implementations, one at MIT and one at the Southern College of Optometry.

Differences from MIT's current setup

We will no longer use a rigid electrode array or it's associated widgets (e.g. clamp with Zebra strip, plastic tissue chambers which we glue to the arrays, castle-shaped brace),
The retina will be oriented ganglion cell side up to provide easy access to the recording electrode and stimulating array,
We may not need to separate the neural retina from the RPE and choroid, as Andy has done so far. It depends on whether we want to shine light on the retina (to probe receptive fields) from above (in which case we may have some interference from the wires on the stimulating array) or below,
Since we only have one recording electrode, we only need one recording amplifier,
Both the recording electrode and stimulating array can be picked up and moved around during an experiment.

To do: part 1 (changes at MIT and SCO)

Both setups will need multi-electrode arrays (to be fabricated by Doug Shire). We should strive to use the same arrays as in the human experiments, both to economize fabrication effort and to compare results from in vitro, in vivo (EEP), and psychophysical experiments. New geometries will be designed in conjunction with Doug, who can do the layout work. The layout can also be done at MIT using our copy of L-Edit, if anyone is so inclined.
To use these arrays efficiently, both setups will need a convenient way to connect one or both terminals of the stimulator to the stimulating arrays (about 100 electrodes). We can duplicate to some degree the scheme that's used for the human experiments. For example, we can use Terry Herndon's PC board and clamp scheme to bridge the gap between the pads on the flexible array and a more conventional connector. We might also use a switchbox like the one that's currently used for human experiments, though for the in-vitro work we can take additional steps to reduce lead lengths and other parasitics. It might be better to construct a mini patch-bay right on the PC board with 100 small connector pins each labeled with a coordinate (x,y). The trick is to make sure that the pins aren't placed too closely together, such that it would be easy to misread the labels and connect to the wrong pin.

To do: part 2 (changes at SCO)

Ralph already has a recording amplifier and an optical system. His system is not computerized, however. To do:

Buy a computer. Cost: $2500. We need to make sure that the computer has a bus architecture which is compatible with the new Microstar board we will also be buying (below. The industry has largely shifted from the ISA architecture to the PCI architure. Most of Microstar's boards are still ISA---the MIT setup uses an ISA board---and it's still possible to get computers with ISA slots, so ISA may be the way to go).
Buy a data acquisition card from Microstar Laboratories. Cost: $2500.
Port Andy's data acquisition software from Linux/MATLAB to Windows/MATLAB. Buy Ralph a copy of MATLAB.

Also, the MIT setup uses isolated amplifiers (built by Andy) to monitor the current and voltage waveforms. These were built to operate from fairly low supply rails (plus and minus nine volts), which has worked fine for Andy. We can build new ones to send to Ralph, but we will have to be careful because his stimulus isolator may be capable of developing back voltages of tens to hundreds of volts.

To do: part 3 (changes at MIT)

The MIT setup needs to be re-worked for single needle recording. In particular, we need to build a new retina-holding apparatus (machine shop), and devise new fluid-flow and heating systems. We should be able to copy Ralph's setup in these regards.

The MIT setup also lacks an optical system for probing ganglion cell optic receptive fields. There are cheap and expensive approaches to this:

The cheap way: use a laser pointer (of sufficiently low power to avoid damaging the retina) to create a small spot (200um) which we will use to manually probe for the optic receptive field of the cell we happen to be recording from.
The expensive way: buy or build an optical system with a light source and lenses, and possibly computer control. Buy a good microscope (possibly inverting, which would be useful for imaging the retina from below if we decide to isolate the neural retina from the RPE, as in previous in vitro work at MIT). The microscope would allow us to get below 100micron accuracy in describing the relative positions of the receptive field center and electrode array.

aegrumet@alum.mit.edu