Full field ElectroRetinoGraphy interpretation by artificial intelligence

Description:

Interpretation of full field ElecroRetinoGraphy is very difficult, since numerous variables exist and every disorder have their own pattern

of alterations in these variables. Therefore, interpretation of ffERG is something that is difficult to commit to

memory. To solve this problem, we have created software that carries out the following tasks:

1) Define the threshold values for normal versus abnormal of every variable in the ffERG test.

2) Software will automatically assign red color to excel cells that are abnormal for that specific variable. Green

color will be assigned to variables that are normal.

3) Software will interpret every variable individually (not as a combination). The results will tell us if the

problem is in rod, cone, muller cells, or inner retina.

4) The software interprets the entire variables and how their variables/abnormality, matches different disorders.

We have searched publications that have already discussed the type of ffERG abnormality that is observed with

different retina disorders and created an algorithm in excel that if the abnormalities match that specific disorder,

it will say "criteria met". If not, it will say "other".

5) Software will graph each variable of the ffERG. The graphs have an area of red, which correspond to

abnormal threshold, and area of green that corresponds to normal threshold (defined in #1). This graph also

plots values according to time, so that progression, worsening, or stability of a ffERG can be monitored over

time.

Patent Information:
For Information, Contact:
Hima Vangapandu
Technology Licensing Specialist
Office of Technology Commercialization
vangapandu@uthscsa.edu
Inventors:
Sepehr Bahadorani
Keywords:
Diagnostics