Direct neural interface system and method
AssigneeCommissariat A L'energie Atomique Et Aux Energies Alternatives
InventorsTetiana Aksenova, Andriy Yelisyeyev
Life of this patent
Estimating the term…
Abstract
A direct neural interface system comprises: a signal acquisition subsystem for acquiring electrophysiological signals representative of neuronal activity of a subject's brain; and a processing unit for representing electrophysiological signals acquired over an observation time window in the form of a N-way data tensor, N being greater than or equal to two, and generating command signals for a machine by applying a regression model over the data tensor; wherein the processing unit is configured or programmed for generating command signals for a machine by applying Generalized Linear regression, with a nonlinear link function, over the data tensor. A method of interfacing a subject's brain to a machine by using such a direct neural interface system is provided.
Claims15
A direct neural interface system comprising:
- a signal acquisition subsystem for acquiring electrophysiological signals s(t) representative of neuronal activity of a subject's brain;
- and
- a processing unit for representing electrophysiological signals acquired over an observation time window in the form of a N-way data tensor (x(t)), N being greater than or equal to one, and generating command signals (S(t)) for a machine by applying a regression model over said data tensor,
- wherein said processing unit is configured or programmed for generating command signals for a machine by applying Generalized Linear regression, with a nonlinear link function, over said data tensor,
- wherein said processing unit is configured or programmed for detecting and correcting outlier elements of said data tensor before applying said Generalized Linear regression, and
- wherein said processing unit is configured or programmed for generating command signals for a machine by applying Generalized Additive Linear regression over said data tensor, said Generalized Additive Linear regression making use of additive functions fi(xi) of the form:
- f
- i
-
-
- (
- x
- i
- )
- =
- {
- f
- 1
- ,
- i
-
- (
- x
- i
- )
- if
-
-
- x
- i
-
-
- is
-
-
- an
-
-
- outlier
- f
- 2
- ,
- i
-
- (
- x
- i
- )
- otherwise
- where xi is an element of said data tensor x(t) and f1,i and f2,i are different functions.
A method of interfacing a subject's brain to a machine, the method comprising the steps of:
- a) acquiring electrophysiological signals (s(t)) representative of neuronal activity of the subject's brain;
- b) representing electrophysiological signals acquired over an observation time window in the form of a N-way data tensor x(t), N being greater than or equal to one;
- and
- c) generating command signals (S(t)) for said machine by applying a regression model over said data tensor;
- wherein said comprises generating said command signals by applying Generalized Linear regression over said data tensor,
- wherein said step c) further comprises detecting and correcting outlier elements of said data tensor before applying said Generalized Linear regression, and
- wherein said step c) is performed by applying Generalized Additive Linear regression over said data tensor, said Generalized Additive Linear regression making use of additive functions fi(xi) of the form:
- f
- i
-
-
- (
- x
- i
- )
- =
- {
- f
- 1
- ,
- i
-
- (
- x
- i
- )
- if
-
-
- x
- i
-
-
- is
-
-
- an
-
-
- outlier
- f
- 2
- ,
- i
-
- (
- x
- i
- )
- otherwise
- where xi is an element of said data tensor x(t) and f1,I and f2,i are different functions.
Dependencies are inferred from the claim text.