Hybrid fitting for a visual prosthesis
The present invention is an improved fitting and training system for a visual prosthesis. Fitting a visual prosthesis through automated means is challenging and fitting a visual prosthesis manually is tedious for clinician and patent, and provides great opportunity for error. A hybrid of computer controlled and manual fitting provides effective, efficient and controlled fitting process. The process includes testing a group of electrodes in random order by providing a prompt followed by stimulation and the patient responding if they saw a percept. After each set, a maximum likelihood algorithm is used to determine the next stimulation level, or if further stimulation is needed for each electrode.
1. A method of brightness fitting a visual prosthesis comprising:
a. providing an array of electrodes suitable to contact visual neural tissue;
b. activating a prompt;
c. activating a signal at a level on a selected electrode within a subset of the array of electrodes;
d. receiving a response from a subject indicating a visual perception;
e. recording the response;
f. repeating the steps b through e for each electrode in the subset;
g. determining if a predetermined minimum of electrodes in the subset elicited responses;
h. repeating steps b through e on simultaneous groups of electrodes if the predetermined minimum of electrodes did not elicit a response; and
i. applying a predetermined computer algorithm to determine a next activation level for each selected electrode.
2. The method according to claim 1 , wherein the simultaneous groups of electrodes include four electrodes.
3. The method according to claim 1 , wherein the step of activating a signal includes activating no signal at a pseudorandom interval.
4. The method according to claim 1 , wherein the predetermined algorithm predicts a confidence of a correct answer based upon a consistency of previous answers.
5. The method according to claim 4 , wherein the maximum likelihood algorithm is according to the equation
L
i
=
∏
q
=
1
m
(
p
q
,
i
)
n
yes
,
q
·
(
1
-
p
q
,
i
)
n
no
,
q
Where:
L i is the likelihood that the threshold value is at level i
p q,i is the probability, according to the Psychometric Function, that the patient will perceive a stimulation at level q given that the threshold is at level i
n yes,q is the number of Yes responses in all previous presentations (trials) at level q
n no,q is the number of No responses in all previous presentations (trials) at level q
m is the number of levels in the range−the highest level.
6. The method according to claim 1 , wherein the predetermined algorithm is a maximum likelihood algorithm.
7. The method according to claim 1 , further comprising generating a video configuration file from results of the predetermined algorithm.