HealthTech

Big data and computational neuroscience addressing neurodegenerative diseases: Dr. Alessandro Crimi and his team

Big data

In our ageing society, one of the most prevalent diseases is Alzheimer’s. There are emerging new methods for both diagnosis and therapy, many of which are region-specific. Understanding the relationships between genetics, brain connections, and Alzheimer’s disease pathology is crucial.

At the same time thanks to the efforts of organizations such as the Alzheimer Disease Neuroimaging Initiative we have now access to several data available to scientists as magnetic resonance scans and genetic data of patients.

Dr Alessandro Crimi and his team are interested in developing tools related to brain diseases, from neurodegenerative (including Alzheimer’s) disorders to brain tumours. The focus ranges from image analysis to computational neuroscience models to neurotech. Investigating biomarkers and shedding some light on genetics and anatomy can open new opportunities for treatments.

During past work of the team, Dr Crimi and his team found some significant correlations between some genes, connectome metrics and dementia scores. They quantified the extent to which clinical dementia rating relates to brain connection changes in time, as well as to gene expression. The results were published on Frontiers in Human Neuroscience: Frontiers | Relating Global and Local Connectome Changes to Dementia and Targeted Gene Expression in Alzheimer’s Disease (frontiersin.org)

Now they are focused on following dementia and Alzheimer’s patients with novel tools during the day to understand more changes and how they occur in real-time and not looking at scans acquired once in a while in a clinic. This huge mole of data from several patients acquired every day will also capitalize on new distributed machine learning approaches called “federated learning”.

Federated Learning separates the ability to use machine learning from the requirement to store data in the cloud by allowing mobile devices to cooperatively learn a shared prediction model while keeping all the training data on the device.

The process goes as follows: your device gets the most recent model, refines it using data from your phone, and then compiles the modifications into a brief, targeted update. Only this model update is transmitted via encrypted communication to the cloud, where it is quickly averaged with updates from other users to enhance the shared model. No specific updates are saved in the cloud; all training data is kept on your device.

If you want to know about the results of those studies and how brain reading and federated learning will be put together, follow the evolutions on the official page of the Sano centre for medical imaging: https://bam.sano.science/

To Top

Pin It on Pinterest

Share This