Data Study Group Summary: PLORAS

Predicting language outcome recovery after stroke

Globally, stroke is the third most common cause of disability. One third of stroke survivors leave hospital with complex difficulties with communicating and understanding language, both written and spoken. This is known as aphasia or dysphasia.

The PLORAS (Predicting Language Outcome Recovery After Stroke) project constitutes a database of stroke survivors’ post-stroke structural MRI scans, as well as their demographic and behavioural data (Seghier et al., 2016).

The aim of the PLORAS project is to empower both stroke survivors and clinicians with high quality predictive information. The PLORAS database was set up with the primary aim of creating a clinical prediction tool for patients and clinicians which would provide: 

• Individualised predictions about the most likely course of recovery;
• Duration of recovery, and;
• Expected degree of recovery. 

The prediction tool would not only provide patients with greater insight into their post-stroke recovery, but would also allow clinicians to create better personalised treatment plans for each individual stroke survivor.

To date, the PLORAS collaborators have attempted several modelling methods for both feature extraction, and predictive modelling, spanning classical linear statistical models to more complex deep learning methods (Hope et al., 2013, Aguilar et al., 2018, Hope et al., 2018).