Methods, systems, and computer readable media for utilizing functional connectivity brain imaging for diagnosis of a neurobehavioral disorder

1. A method for utilizing functional connectivity brain imaging for predicting a neurobehavioral disorder diagnosis, the method comprising:

receiving brain imaging data for a human subject of a first age, wherein the brain imaging data includes functional connectivity magnetic resonance imaging (fcMRI) data, wherein the first age is under two years, wherein the first age comprises an age at which the human subject is presymptomatic with respect to a neurobehavioral disorder, wherein the neurobehavioral disorder comprises autism spectrum disorder (ASD);

determining, using a trained machine learning classification algorithm and a plurality of functional connections between brain locations in the fcMRI data associated with the first age, a neurobehavioral disorder diagnosis of ASD-positive for the human subject prior to a clinical diagnosis of the human subject as ASD-positive at a second age greater than the first age, wherein the second age is two years or older, wherein training the machine learning classification algorithm includes learning the plurality of functional connections used to determine the neurobehavioral disorder diagnosis using behaviorally-related functional connections and known diagnostic outcomes associated with a training data set, wherein each of the behaviorally-related functional connections has a nominally significant behavioral correlation with ASD-related behaviors of human subjects in the training data set assessed at 24 months of age and wherein each of the behaviorally-related functional connections is selected from functional connections within a feature space without using knowledge of the known diagnostic outcomes, wherein the nominally significant behavioral correlation is indicated by a p-value of less than 0.05; and

performing an intervention action based on the neurobehavioral disorder diagnosis.

2. The method of claim 1 wherein the first age is six months.

3. The method of claim 2 wherein the brain imaging data is preprocessed by performing compensation for slice dependent time shifts using sinc interpolation, correction of systematic odd-even slice intensity differences caused by interleaved acquisition, or spatial realignment to compensate for head motion within and across imaging acquisition.

4. The method of claim 1 wherein at least one functional connection of the plurality of functional connections is located within at least one brain region of interest (ROI) associated with an ASD behavior.

5. The method of claim 4 wherein the at least one brain ROI is related to social behavior, language, motor development, or repetitive behavior.

6. The method of claim 1 wherein the machine learning classification algorithm comprises a trained linear support vector machine (SVM) classifier.

7. A system for utilizing functional connectivity brain imaging for predicting a neurobehavioral disorder diagnosis, the system comprising:

at least one processor; and

a neurobehavioral disorder diagnosis module (ADM) implemented using the at least one processor, wherein the ADM is configured for receiving brain imaging data for a human subject of a first age, wherein the first age is under two years, wherein the first age comprises an age at which the human subject is presymptomatic with respect to a neurobehavioral disorder, wherein the neurobehavioral disorder comprises autism spectrum disorder (ASD), wherein the brain imaging data includes functional connectivity magnetic resonance imaging (fcMRI) data; for determining, using a trained machine learning classification algorithm and a plurality of functional connections between brain locations in the fcMRI data associated with the first age, a neurobehavioral disorder diagnosis of ASD-positive for the human subject prior to a clinical diagnosis of the human subject as ASD-positive at a second age greater than the first age, wherein the second age is two years or older, wherein training the machine learning classification algorithm includes learning the plurality of functional connections used to determine the neurobehavioral disorder diagnosis using behaviorally-related functional connections and known diagnostic outcomes associated with a training data set, wherein each of the behaviorally-related functional connections has a nominally significant behavioral correlation with ASD-related behaviors of human subjects in the training data set assessed at 24 months of age and wherein each of the behaviorally-related functional connections is selected from functional connections within a feature space without using knowledge of the known diagnostic outcomes, wherein the nominally significant behavioral correlation is indicated by a p-value of less than 0.05; and for performing an intervention action based on the neurobehavioral disorder diagnosis.

8. The system of claim 7 wherein the first age is six months.

9. The system of claim 8 wherein the brain imaging data is preprocessed by performing compensation for slice dependent time shifts using sinc interpolation, correction of systematic odd-even slice intensity differences caused by interleaved acquisition, or spatial realignment to compensate for head motion within and across imaging acquisition.

10. The system of claim 7 wherein at least one functional connection of the plurality of functional connections is located within at least one brain region of interest (ROI) associated with an ASD behavior.

11. The system of claim 10 wherein the at least one brain ROI is related to social behavior, language, motor development, or repetitive behavior.

12. The system of claim 7 wherein the machine learning classification algorithm comprises a trained linear support vector machine (SVM) classifier.

13. A non-transitory computer readable medium having stored thereon executable instructions that when executed by at least one processor of a computer cause the computer to perform steps comprising:

receiving brain imaging data for a human subject of a first age, wherein the brain imaging data includes functional connectivity magnetic resonance imaging (fcMRI) data, wherein the first age is under two years, wherein the first age comprises an age at which the human subject is presymptomatic with respect to a neurobehavioral disorder, wherein the neurobehavioral disorder comprises autism spectrum disorder (ASD);

determining, using a trained machine learning classification algorithm and a plurality of functional connections between brain locations in the fcMRI data associated with the first age, a neurobehavioral disorder diagnosis of ASD-positive for the human subject prior to a clinical diagnosis of the human subject as ASD-positive at a second age age greater than the first age, wherein the second age is two years or older, wherein training the machine learning classification algorithm includes learning the plurality of functional connections used to determine the neurobehavioral disorder diagnosis using behaviorally-related functional connections and known diagnostic outcomes associated with a training data set, wherein each of the behaviorally-related functional connections has a nominally significant behavioral correlation with ASD-related behaviors of human subjects in the training data set assessed at 24 months of age and wherein each of the behaviorally-related functional connections is selected from functional connections within a feature space without using knowledge of the known diagnostic outcomes, wherein the nominally significant behavioral correlation is indicated by a p-value of less than 0.05; and

performing an intervention action based on the neurobehavioral disorder diagnosis.