Systems and methods for developing brain computer interface
View Patent ↗Systems, methods, and protocols for developing invasive brain computer interface (iBCI) decoders non-invasively by using emulated brain data are provided. A human operator can interact in real-time with control algorithms designed for iBCI. An operator can provide input to one or more computer models (e.g., via body gestures), and this process can generate emulated brain signals that would otherwise require invasive brain electrodes to obtain.
1. A system for non-invasively developing a decoder of an invasive brain computer interface (iBCI), the system comprising:
a motion capture device;
a processor in operable communication with the motion capture device; and
a machine-readable medium in operable communication with the processor and having instructions stored thereon that, when executed by the processor, perform a plurality of steps including:
receiving kinematic data of a user from the motion capture device;
utilizing an artificial neural network (ANN) to generate emulated neuron firing rates associated with the kinematic data; and
translating the emulated neuron firing rates to corresponding control commands for an external device of the iBCI, thereby obtaining the decoder,
the ANN being a trained ANN that has been trained using training kinematic data and real neuron firing rates associated with the training kinematic data,
the ANN comprising an input layer, an output layer, and four hidden layers,
the input layer comprising 95 nodes,
the four hidden layers comprising a first hidden layer, a second hidden layer, a third hidden layer, and a fourth hidden layer,
the first hidden layer comprising 150 nodes,
the second hidden layer comprising 100 nodes,
the third hidden layer comprising 50 nodes,
the fourth hidden layer comprising 30 nodes.
2. The system according to claim 1 , the motion capture device being a data glove, and
the kinematic data comprising hand gesture kinematic data of the user.
3. The system according to claim 1 , the kinematic data being real-time kinematic data of the user using the motion capture device.
4. The system according to claim 1 , the training kinematic data comprising data from at least 20 different training sessions.
5. The system according to claim 1 , the training kinematic data being training hand gesture kinematic data.
6. The system according to claim 5 , the training hand gesture kinematic data comprising data for at least eight different hand gestures, and
the real neuron firing rates being respectively associated with the at least eight different hand gestures.
7. The system according to claim 1 , further comprising the external device of the iBCI,
the external device being a robot arm, a wheelchair, a keyboard, a smart device, or a computer.
8. A method for non-invasively developing a decoder of an invasive brain computer interface (iBCI), the method comprising:
using, by a user, a motion capture device to capture kinematic data of the user;
receiving, by a processor in operable communication with the motion capture device, the kinematic data of the user;
utilizing, by the processor, an artificial neural network (ANN) to generate emulated neuron firing rates associated with the kinematic data; and
translating, by the processor, the emulated neuron firing rates to corresponding control commands for an external device of the iBCI, thereby obtaining the decoder,
the ANN being a trained ANN that has been trained using training kinematic data and real neuron firing rates associated with the training kinematic data,
the ANN comprising an input layer, an output layer, and four hidden layers,
the input layer comprising 95 nodes,
the four hidden layers comprising a first hidden layer, a second hidden layer, a third hidden layer, and a fourth hidden layer,
the first hidden layer comprising 150 nodes,
the second hidden layer comprising 100 nodes,
the third hidden layer comprising 50 nodes,
the fourth hidden layer comprising 30 nodes.
9. The method according to claim 8 , the motion capture device being a data glove, and
the kinematic data comprising hand gesture kinematic data of the user.
10. The method according to claim 8 , the kinematic data being real-time kinematic data of the user using the motion capture device.
11. The method according to claim 8 , the training kinematic data being training hand gesture kinematic data.
12. The method according to claim 8 , the training kinematic data comprising data from at least 20 different training sessions.
13. The method according to claim 11 ,
the training hand gesture kinematic data comprising data for at least eight different hand gestures, and
the real neuron firing rates being respectively associated with the at least eight different hand gestures.
14. The method according to claim 8 , further comprising providing the corresponding control commands to the external device of the iBCI,
the external device being a robot arm, a wheelchair, a keyboard, a smart device, or a computer.
15. A system for non-invasively developing a decoder of an invasive brain computer interface (iBCI), the system comprising:
a motion capture device;
a processor in operable communication with the motion capture device;
an external device of the iBCI in operable communication with the processor; and
a machine-readable medium in operable communication with the processor and having instructions stored thereon that, when executed by the processor, perform a plurality of steps including:
receiving real-time hand gesture kinematic data of a user from the motion capture device;
utilizing an artificial neural network (ANN) to generate emulated neuron firing rates associated with the real-time hand gesture kinematic data; and
translating the emulated neuron firing rates to corresponding control commands for the external device of the iBCI, thereby obtaining the decoder,
the motion capture device being a data glove,
the ANN being a trained ANN that has been trained using training hand gesture kinematic data and real neuron firing rates associated with the training hand gesture kinematic data,
the training hand gesture kinematic data comprising data from a plurality of different training sessions,
the training hand gesture kinematic data comprising data for at least eight different hand gestures,
the real neuron firing rates being respectively associated with the at least eight different hand gestures,
the ANN comprising an input layer, an output layer, a first hidden layer, a second hidden layer, a third hidden layer, and a fourth hidden layer,
the input layer comprising 95 nodes,
the first hidden layer comprising 150 nodes,
the second hidden layer comprising 100 nodes,
the third hidden layer comprising 50 nodes,
the fourth hidden layer comprising 30 nodes, and
the external device being a robot arm, a wheelchair, a keyboard, a smart device, or a computer.