“We intend to develop software to complement the diagnosis and quantification of fat deposits in the body and metabolic imbalances (iron concentration). Then, through the analysis of MR brain scans, assess the effects of these values on neuronal connectivity (DTI) and functional connectivity (Resting States networks (RS)). As data is available for the same subjects over time, assess if the software developed was able to predict through biomarkers values of obesity and iron concentration on volunteers with time.
Workflow is as follows:
1. With IA, quantify the concentrations of iron in liver, spleen and pancreas. All this based on MR images of these organs.
2. Automatically quantify abdominal, subcutaneous or infiltration in organ of fat (liver, pancreas and spleen). Also consider extremities and any other body parts of relevance. We would have to develop different AI Machine Learning ML applications to do this automatically. Work would have to also focus on segmentations of tissues and all based on MRI images. If possible, correlate with other anthropometric data and blood samples. Correlation could be complemented with the use of neuronal networks to find relationships between these variables.
3. Analyse RS, DTI, and other MR modalities to associate results with the parameters obtained in points 1 and 2. Assess how they influence brain function. An in-detail study of connectivity and correlation of these graphic parameters with iron and fat would also be looked for, trying to find which brain areas are mostly affected and how. As before correlations or neuronal networks would be used to find relationships between these variables.
4. Find in points 1, 2 and 3, biomarkers that can predict the evolution of accumulation of fat, iron and changes in brain function and structure with time. This using AI and ML techniques. This would serve as a diagnostic and prediction tool. This would be possible due to the information over time found for a same patient in Biobank.”
