Predicting brain data using machine learning models

A system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:

• obtaining first connectivity data generated from raw brain data captured from a brain of a subject, wherein the first connectivity data represents, for each pair of parcellations in a first set of parcellations in the brain of the subject, a correlation between brain activity of the pair of parcellations;
• processing the first connectivity data using a machine learning model to generate predicted second connectivity data, wherein:
• the predicted second connectivity data represents, for each pair of parcellations in a second set of parcellations in the brain of the patient that is different from the first set of parcellations, a correlation between brain activity of the pair of parcellations,
• the machine learning model having been trained to determine a mapping between correlation values for the first set of parcellations and correlation values for the second set of parcellations;
• the first connectivity data is generated from EEG data captured from the brain of the subject, and the predicted second connectivity data represents predicted fMRI data,
• the EEG data is captured from the brain of the subject using a particular EEG montage that identifies, for each of a plurality of EEG electrodes, a respective location on a scalp of the subject,
• the EEG data comprises respective channel data captured by each of the plurality of EEG electrodes of the particular EEG montage, and
• the first connectivity data is generated by performing operations comprising:
• determining, for each parcellation in the first set of parcellations, a corresponding electrode of the plurality of electrodes of the particular EEG montage;
• and
• determining, for each pair of parcellations in the first set of parcellations comprising a first parcellation and a second parcellation, a correlation between (i) the channel data captured by the electrode corresponding to the first parcellation and (ii) the channel data captured by the electrode corresponding to the second parcellation;
• and
• taking an action based on the predicted second connectivity data.

One or more non-transitory storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:

• obtaining first connectivity data generated from raw brain data captured from a brain of a subject, wherein the first connectivity data represents, for each pair of parcellations in a first set of parcellations in the brain of the subject, a correlation between brain activity of the pair of parcellations;
• processing the first connectivity data using a machine learning model to generate predicted second connectivity data, wherein:
• the predicted second connectivity data represents, for each pair of parcellations in a second set of parcellations in the brain of the patient that is different from the first set of parcellations, a correlation between brain activity of the pair of parcellations,
• the machine learning model having been trained to determine a mapping between correlation values for the first set of parcellations and correlation values for the second set of parcellations;
• the first connectivity data is generated from EEG data captured from the brain of the subject, and the predicted second connectivity data represents predicted fMRI data,
• the EEG data is captured from the brain of the subject using a particular EEG montage that identifies, for each of a plurality of EEG electrodes, a respective location on a scalp of the subject,
• the EEG data comprises respective channel data captured by each of the plurality of EEG electrodes of the particular EEG montage, and
• the first connectivity data is generated by performing operations comprising:
• determining, for each parcellation in the first set of parcellations, a corresponding electrode of the plurality of electrodes of the particular EEG montage;
• and
• determining, for each pair of parcellations in the first set of parcellations comprising a first parcellation and a second parcellation, a correlation between (i) the channel data captured by the electrode corresponding to the first parcellation and (ii) the channel data captured by the electrode corresponding to the second parcellation;
• and
• taking an action based on the predicted second connectivity data.