Generating a drifting sinusoidal grating or drifting bar stimulus

This example shows how to use drifting psychophysics-based stimuli for a retinal implant.

Along with images, videos, and oter built-in stimuli, pulse2percept supports generating GratingStimulus and BarStimulus as stimuli that can be passed as percepts to implants.

Creating a Stimulus

First, create the stimuli:

Shape ((height, width) in pixels) is the only required parameter for creating these stimuli.

A drifting sinusoidal grating is represented by GratingStimulus. The following illustrates one frame of a grating stimulus.

import matplotlib.pyplot as plt
from pulse2percept.stimuli import GratingStimulus
stim = GratingStimulus((50, 50), spatial_freq=0.1, temporal_freq=0.1)
plt.imshow(stim.data[:, 0].reshape(50, 50), cmap='gray')
plt.title("Grating Stimulus")
plt.show()

You can view the entire stimulus over time using stim.play()

stim.play()

Once Loop Reflect

Here, the spatial frequency of the grating (i.e., the inverse of how many pixels it takes to represent one cycle of the sinusoid) is given as 0.1 cycles/pixel, whereas the temporal frequency (i.e., the inverse of how many frames it takes to represent one cycle of the sinusoid) is given as 0.1 cycles/frame. By default, the drift direction of the grating will be to the right (0 degrees).

A drifting bar is represented by BarStimulus. A visual example of a basic sinusoidal grating can be generated as such:

from pulse2percept.stimuli.psychophysics import BarStimulus
stim = BarStimulus((50, 50), speed=1)
stim.play()

Once Loop Reflect

Here, the drift speed of the bar is given as 1 pixel/frame and, by default, the bar will drift to the right (0 degrees).

Customizing the Stimulus

For both BarStimulus and GratingStimulus, the only argument you must pass to the constructor is the shape (height, width) in pixels. There are many optional arguments that can be passed to change various attributes of the stimulus. In the examples above, we changed the speed at which the GratingStimulus changed with temporal_freq (scalar, cycles/frame) and the speed at which the BarStimulus moved with speed (scalar, pixels/frame) in order to make the effect easier to visualize. If, for example, we wanted a longer stimulus, we cold set the time parameter (in units of milliseconds) to change the duration of the stimulus:

stim = BarStimulus((50, 50), speed=1, time=1500)
stim.play()

Once Loop Reflect

We can also change the direction (scalar in [0, 360) degrees), where 0 degrees represents rightward motion, 90 degrees represents upward motion, 180 degrees represents leftward motion, and 270 degrees represents downward motion.

BarStimulus((height, width), direction=direction)

Or the contrast (scalar in [0,1])

BarStimulus((height, width), contrast=contrast)

For exact info on all of the arguments, please refer to BarStimulus and GratingStimulus.

Passing to an Implant

Psychophsyics stimuli can be passed to an implant and combined with a model. To demonstrate, we will pass a GratingStimululus to an ArgusII implant and use the AxonMapModel [Beyeler2019] to interpret it:

Important

Don’t forget to build the model before using predict_percept

from pulse2percept.implants import ArgusII
from pulse2percept.models import AxonMapModel

model = AxonMapModel()
model.build()

implant = ArgusII()
implant.stim = GratingStimulus((25,25), temporal_freq=0.1)

percept = model.predict_percept(implant)
percept.play()

Once Loop Reflect

As you can see in the above code segment, the stimulus passed to the implant does not necessarily have to have the same dimensions as the electrode grid. This is functionality built in to the implant code: The implant will automatically rescale the stimulus to the appropriate size. In the case of Argus II, the stimulus would thus be downscaled to a 6x10 image.

Pre-Processing Stimuli

Since both BarStimulus and GratingStimulus inherit form VideoStimulus, we can apply any video processing methods provided by VideoStimulus.

In the following example, we will invert the stimulus before passing it to the implant:

model = AxonMapModel()
model.build()

implant = ArgusII()
implant.stim = GratingStimulus((25,25), temporal_freq=0.1).invert()

percept = model.predict_percept(implant)
percept.play()

Once Loop Reflect