Using AI and fractal dimension to accurately monitor ataxia

Exploiting artificial intelligence (AI) to develop tools for improving the monitoring of treatment of rare, progressive, and highly debilitating diseases such as Friedreich's (FRDA) and spinocerebellar ataxia (SCA). This is the goal of the new research project led by Professor Stefano Diciotti from the Department of Electrical, Electronic and Information Engineering at the University of Bologna and of the Alma Mater Research Institute for Human-Centered Artificial Intelligence.

This and other six projects won the Spring Seed Grant issued by Telethon Foundation to help patients' associations invest their funds into research projects focusing on rare and often underresearched diseases. The University of Bologna will lead this research project supported by the Italian Association for Ataxia (AISA - Associazione Italiana Sindromi Atassiche).

Ataxias are neurodegenerative diseases encompassing the gradual loss of voluntary muscle movement. Some forms of this disease are hereditary. Friedreich's and spinocerebellar ataxias are both hereditary and caused by the deterioration of the spinal cord and the cerebellum (i.e. the movement coordination center).

To date, there are no effective treatments for hereditary ataxia. However, by employing advanced techniques for analyzing patients' brain MRIs, researchers are able to identify indexes that allow to monitor the disease more accurately. This, in turn, could help the development of new treatment approaches.

Our methodology employs a very promising quantitative index of the brain structural complexity: the fractal dimension. We draw this index from the analysis of MRIs and believe it could provide important insights into the alterations causing the development and early ageing of the brain that is typical of this type of ataxia".

Professor Stefano Diciotti, Department of Electrical, Electronic and Information Engineering, University of Bologna

According to their brain MRIs, patients suffering from neurodegenerative diseases present a greater reduction of the structural complexity of the grey and white matters if compared to healthy subjects. By exploiting AI techniques, this model of analysis could eventually provide each patient with a forecast of the clinical development of their disease.

Researchers will develop this analysis tool gathering from the meta-dataset of the international working group ENIGMA-Ataxia. This working group collected the brain MRIs, and the genetic and clinical data of approximately 800 healthy subjects and more than 800 patients suffering from hereditary ataxia from 21 specialized centres.

"We will use the ENIGMA-Ataxia platform to quantify the alterations in the fractal dimension that are typically associated with the abnormal development or the neurodegenerative issues in ataxia patients", explains Diciotti. "This international partnership will enable us to develop innovative techniques and work on a cutting-edge computational approach that will improve our understanding of the pathophysiology of hereditary ataxias, eventually leading to new strategies for a better design of clinical studies".