System and method for identifying transdiagnostic features shared across mental health disorders
Inventors
LIU, Yuelu • MELLEM, Monika Sharma • Ahammad, Parvez • GONZALEZ CABEZAS, Humberto Andres • Kollada, Matthew
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11942224-B2
Publication Date
2024-03-26
Expiration Date
Abstract
A system for evaluating mental health of patients includes a memory and a control system. The memory contains executable code storing instructions for performing a method. The control system is coupled to the memory and includes one or more processors. The control system is configured to execute the machine executable code to cause the control system to perform the method: A selection of answers associated with a patient is received. The selection of answers corresponds to each question in a series of questions from mental health questionnaires. Unprocessed MRI data are received. The unprocessed MRI data correspond to a set of MRI images of a biological structure associated with the patient. The unprocessed MRI data is processed to output a set of MRI features. Using a machine learning model, the selection of answers and the set of MRI features are processed to output a mental health indication of the patient.
Core Innovation
A system and computer-implemented method evaluate a patient for mental health issues by receiving a selection of answers from a series of mental health questionnaire questions together with unprocessed MRI data associated with the patient. The system processes the selection of answers and the unprocessed MRI data using a machine learning model to output a mental health indication of the patient.
The machine learning model is generated from training data corresponding to a plurality of individuals, where the training data includes MRI data and selections of answers to the series of questions. The generation determines a plurality of features from the training data, extracts importance measures for each feature, generates a plurality of subset machine learning models based on the extracted importance measures, and selects at least one of the subset machine learning models as the machine learning model.
In some implementations, the unprocessed MRI data includes functional MRI data, resting-state functional MRI data, structural MRI data, or a combination thereof, and corresponds to the patient92s brain. The processing and model generation support a categorical mental health indication that can indicate at least one specified condition, including a neuropsychiatric disorder, schizophrenia, bipolar disorder, or any combination thereof.
Claims Coverage
The document includes three independent claims. Across the independent claims, there are 3 main inventive feature sets: multimodal evaluation using questionnaire answer selections with unprocessed MRI, training with extracted feature importance to generate subset machine learning models, and outputting a mental health indication based on the trained model, including categorical indications for specified disorders in some claim sets.
Multimodal mental health evaluation using questionnaire answer selections and unprocessed MRI
A system receives a selection of answers from a patient corresponding to each question in a series of questions from mental health questionnaires, receives unprocessed MRI data associated with the patient, and processes the selection of answers and the unprocessed MRI data using a machine learning model to output a mental health indication of the patient.
Feature-importance-driven generation of subset machine learning models
The machine learning model is generated by receiving training data for a plurality of individuals comprising MRI data and a selection of answers to the series of questions, determining a plurality of features, extracting importance measures for each feature, generating a plurality of subset machine learning models based on the extracted importance measures, and selecting at least one subset machine learning model as the machine learning model.
Labeled training data to generate the machine learning model using feature importance measures
A system receives a selection of answers associated with a patient and unprocessed MRI data associated with the patient, and processes the selection of answers and the MRI data using a machine learning model to output a mental health indication, wherein the machine learning model is generated by receiving labeled training data indicating whether each of a plurality of individuals has one or more mental health disorders, determining a plurality of features from the labeled training data, training an initial machine learning model in a supervised manner based on the plurality of features, extracting importance measures for each feature based on the training, and generating the machine learning model based on the extracted importance measures.
Across the independent claims, the claims cover multimodal patient evaluation that combines questionnaire answer selections with unprocessed MRI data, and a machine-learning training pipeline that uses feature importance measures to generate and select subset machine learning models, or to generate the machine learning model based on extracted importance measures, for producing a mental health indication.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?