How to Use Multimodal AI to Predict the Longitudinal Course of Psychosis

There are a lot of different modalities here and multimodal AI is very nascent. We want the data to be available to everybody, all qualified researchers to be able to ask their own questions use their own methodology because we don't necessarily think any one group has a monopoly on good ideas. One of the techniques we're trying to develop is uniform standardized methods of assessing the clinical high risk state. So maybe we could take individuals in their very first episode of psychosis, or at very high risk for it and treat them with glutamatergic agents - that's another area I'm very excited about.