Decoding movement intention using ultrasound neuroimaging
Methods and systems are provided for decoding movement intentions using functional ultrasound (fUS) imaging of the brain. In one example, decoding movement intentions include determining a memory phase of a cognitive state of the brain, the memory phase between a gaze fixation phase and movement execution phase, and determining one or more movement intentions including one or more of intended effector (e.g., hand, eye) and intended direction (e.g., right, left) according to a machine learning algorithm trained to classify one or more movement intentions simultaneously.
1. A neural interface system comprising:
an ultrasound transducer;
a controller storing instructions in non-transitory memory that when executed cause the controller to:
acquire, via the at least one ultrasound transducer, a plurality of ultrasound images;
process the plurality of ultrasound images, in real-time, to determine a movement intention for actuating a device from the plurality of ultrasound images by classifying one or more of a task phase, a movement intention direction, and an intended effector according to a machine learning algorithm receiving the plurality of ultrasound images as input, wherein the machine learning algorithm is trained using a plurality of ultrasound images and corresponding ones of a task phase, a movement intention direction, and an intended effector; and
adjust an actuator a device, in real-time, according to the determined movement intention, the device communicatively coupled to the controller;
wherein the ultrasound transducer is positioned to image an area of a brain of a subject.
2. The neural interface system of claim 1 , wherein the plurality of ultrasound images is acquired by transmitting a set of plane waves, each of the set of plane waves transmitted at a different angulation; and wherein the at least one ultrasound transducer is a high-frequency ultrasound transducer configured to emit ultrasonic waves above a threshold frequency.
3. The neural interface system of claim 1 , wherein the one or more movement intentions includes a task phase of a cognitive state of the subject, the task phase occurring prior to imagining, attempting, or executing an intended movement.
4. The neural interface system of claim 1 , wherein the one or more movement intentions includes an intended effector, an intended movement direction, and/or an intended action.
5. The neural interface system of claim 1 , wherein the device is a prosthetic limb, an orthotic assistance device, functional electrical stimulation, or a computing device.
6. The neural interface system of claim 1 , wherein the process the plurality of ultrasound images, in real-time, to determine a movement intention comprises determining changes in cerebral blood flow over a duration using the plurality of ultrasound images and matching the determined changes in cerebral blood flow to changes in cerebral blood flow corresponding to the movement intention.
7. The neural interface system of claim 1 , wherein the machine learning algorithm is trained to classify one or more of the task phase, the movement intention direction, and the intended effector simultaneously.
8. The neural interface system of claim 1 , wherein the controller includes further instructions that when executed cause the controller to process the plurality of ultrasound images, in real-time, to determine a goal associated with the movement intention, the goal including one of an information defining goal, a performance task goal, a target object goal, or a position goal.
9. The neural interface system of claim 1 , wherein the controller includes further instructions that when executed cause the controller to: while adjusting the actuator of the device, acquire, a next plurality of ultrasound images, and process the next plurality of ultrasound images to determine a subsequent movement intentions.
10. The neural interface system of claim 9 , wherein the controller includes further instructions that when executed cause the controller to: responsive to completing adjustment of the actuator of the device according to the movement intention, further adjust the actuator of the device according to the subsequent movement intention.
11. A system comprising:
an ultrasound transducer positioned to image an area of a brain of a subject;
an ultrasound scanning unit comprising a processor, the processor storing instructions in non-transitory memory that when executed cause the one or more processors to:
acquire, via the ultrasound transducer, a plurality of ultrasound images;
process the plurality of ultrasound images, in real-time, to determine a task phase associated with a cognitive state of the subject by processing the plurality of ultrasound images according to a trained machine learning algorithm, the trained machine learning algorithm based on class-wise principal component analysis (CPCA) and linear discriminant analysis (LDA), wherein the machine learning algorithm is trained from an input set of ultrasound images corresponding to a task phase of a cognitive state of a subject; and
responsive to determining the task phase, determine a movement intention occurring prior to onset of actual movement by the subject based on the plurality of ultrasound images.
12. The system of claim 11 , wherein the movement intention includes an intended effector, an intended movement direction, and/or an intended action.
13. The system of claim 12 , wherein the movement intention is one of a plurality of movement intentions determined by the phase task, and wherein the plurality of movement intentions are determined simultaneously.
14. The system of claim 11 , wherein the area of the brain is sensorimotor cortical or sub-cortical motor brain areas.
15. The system of claim 11 , wherein the area of the brain is posterior parietal cortex, primary motor cortex, and/or premotor cortex.
16. A method for operating a brain-machine interface, the method comprising:
receiving a plurality of ultrasound images from an ultrasound probe, the ultrasound probe positioned to image an area of a brain;
processing, the plurality of ultrasound images to output a set of functional images, the functional images showing cerebral blood flow changes in the area of the brain; and
classifying, via a trained machine learning algorithm, an intended behavior occurring prior to the onset of the behavior of the subject based on cerebral blood flow changes determined from the set of functional images by applying class-wise principal component analysis (CPCA) on the set of functional images to output a set of CPCA transformed features and performing linear discriminant analysis on the CPCA transformed features.
17. The method of claim 16 , further comprising generating an actuation signal according to the classified intended behavior and transmitting the actuation signal to a device to execute the intended behavior; and wherein the plurality of ultrasound images are generated by transmitting a set of plane waves, each of the set of plane waves transmitted at a different angulation.