3 Experimental Specifications

Retinitis pigmentosa is simulated by breeding mice that have a mutation which creates dysfunctional CNG protein channel in the rod photoreceptor cells in their retina. The CNGB gene is knocked out using the Cre-Lox recombinant, which can then be temporally activated with the delivery of tamoxifin. This dysfunction of the CNG protein channel simulates retinitis pigmentosa as they both initiate rod cell death and also eventually the death of the cone photoreceptor cells. The death of photoreceptors necessitates a restructuring of the retina’s circuitry because downstream bipolar cells will have to adapt to transmit signals. Thus, by looking at the alterations of the retinal code, this study explores the physiological restructuring that happens because of this disease. Treatment can also be simulated as tamoxifin initiates activation of the CNGB gene adding a layer to explore how phyiological changes manifest in the retinal code as functional CNG channels get produced .

19 retinas were extracted from mice and platinized on a microelectrode array to measure the action potentials sent by their ganglion cell layers. These cells were recorded across various stimuli: natural motion movies, checkerboard patterns at cone light level (daylight), and checkerboard patterns at rod-light levels (dark levels).

Spikes on a single electrode may contain multiple cells and multiple electrodes can contain a single cell, so a spike sorting software was used to attribute the spikes in all recordings to individual cells. The software uses PCA clustering of the spike features to identify the cluster of cells on the electrode array. A combination of manual and algorithmic spike sorting was done on some recordings to evaluate the effectiveness of the spike sorting algorithm in context of this experiment. The additional step of spiking sorting did not affect the the majority of recorded cells, which provided evidence for the effectiveness of the spike sorting in all the recordings.