US 11,989,641 B2
Multiplicative recurrent neural network for fast and robust intracortical brain machine interface decoders
David Sussillo, Portola Valley, CA (US); Jonathan C. Kao, Los Angeles, CA (US); Sergey Stavisky, Davis, CA (US); and Krishna V. Shenoy
Assigned to The Board of Trustees of the Leland Stanford Junior University, Stanford, CA (US)
Filed by The Board of Trustees of the Leland Stanford Junior University, Stanford, CA (US)
Filed on Oct. 3, 2022, as Appl. No. 17/937,745.
Application 17/937,745 is a continuation of application No. 16/292,000, filed on Mar. 4, 2019, granted, now 11,461,618.
Application 16/292,000 is a continuation of application No. 14/826,300, filed on Aug. 14, 2015, granted, now 10,223,634, issued on Mar. 5, 2019.
Claims priority of provisional application 62/037,441, filed on Aug. 14, 2014.
Prior Publication US 2023/0144342 A1, May 11, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 3/04 (2023.01); A61B 5/24 (2021.01); G06N 3/044 (2023.01); G06N 3/08 (2023.01)
CPC G06N 3/044 (2023.01) [A61B 5/24 (2021.01); G06N 3/04 (2013.01); G06N 3/08 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A brain-machine interface system, comprising:
at least one multi-electrode array implanted into a user's brain;
a prosthetic device in communication with a neural signal decoder; and
the neural signal decoder implemented using a computing device in communication with the at least one multi-electrode array and the prosthetic device, where the neural signal decoder comprises a recurrent neural network trained by:
obtaining a plurality of neural signals from the user using the at least one multi-electrode array;
modifying the plurality of neural signals by randomly adding and removing neural spikes from the plurality of neural signals such that a mean number of neural spikes across the plurality of neural signals is preserved;
instantiating the recurrent neural network, where the recurrent neural network is configured to map neural signals to kinematic functions of the prosthetic device; and
training the recurrent neural network using the plurality of modified neural signals; and
where the neural signal decoder is configured to:
obtain at least one neural signal from the user using the at least one multi-electrode array;
decode an intended movement of the prosthetic device from the obtained neural signal using the recurrent neural network;
control the prosthetic device to perform the intended movement.