Raster Strategies

Added in version 0.10.0.

Some stimulators cannot drive every electrode simultaneously. A Raster divides an array into groups that take turns. It is a scheduling constraint used by an encoder, not a stimulus by itself.

Basic usage

Rasters are attached to an implant:

import pulse2percept as p2p

implant = p2p.implants.ArgusII()
implant.raster = p2p.implants.CheckerboardRaster(n_groups=5)
implant.encoder = p2p.stimuli.AmplitudeEncoder(
    amp_range=(0, 50),
    freq=20,
)
implant.stim = p2p.stimuli.BostonTrain()

Electrodes in one group may pulse together; different groups occupy different time slots.

Built-in strategies

Raster	Grouping
SequentialRaster	Sequential or interleaved groups
CheckerboardRaster	Spatially distributed grid groups
CustomRaster	Explicit user-defined groups

For example:

implant.raster = p2p.implants.SequentialRaster(n_groups=6)
implant.raster = p2p.implants.SequentialRaster(
    n_groups=6, interleave=True
)
implant.raster = p2p.implants.CheckerboardRaster(n_groups=5)

A CustomRaster is useful when the hardware already defines the groups. Every electrode must belong to exactly one group.

Use raster.plot() to inspect the pattern and raster.members(...) to retrieve the electrodes in a group.

Timing

Groups fire in order. group_dur sets the spacing between group starts. If it is None, the encoder spreads the groups across the pulse period. An explicit value fixes the raster sweep duration:

implant.raster = p2p.implants.SequentialRaster(
    n_groups=6,
    group_dur=1,
)

The slot must be long enough for a pulse, and the full sweep must fit within the relevant pulse period.

With amplitude encoding, all electrodes share a pulse period, so rastering only offsets the groups. With frequency encoding, electrodes may request different periods; those schedules are constrained to whole raster sweeps and may therefore run more slowly than requested, never faster.

If rastering is not part of the device or question being modeled, leave implant.raster unset.