The Modelling of Glaucoma Progression through the Use of Cellular Automata

Abstract

We propose a model of glaucoma progression based on the application of Cellular Automata (CA) to visual field (VF) data, obtained through automated perimetry. VF sensitivities are converted into ganglion cell loss and CA are utilised to model the gradual deterioration of vision, mimicking degeneration of the actual ganglia. First we discuss the construction of a grid that approximates the VF map and the corresponding layer of ganglia in terms of cell counts in individual fields. The grid is populated with dead cells in accordance with patients’ tests, and then we run a CA, utilising a majority and a probabilistic rule. Preliminary results are presented, showing that during its evolution, the CA often converges to configurations where the death of cells resembles VF data of the same patients, at later time. That is, the percentage loss of cells in VF fields observed in the CA resembles the real VF data.