Methods and apparatus for cochlear implant signal processing
View Patent ↗A cochlear implant processing strategy increases speech clarity and higher temporal performance. The strategy determines the power spectral component within each channel, and dynamically selects or de-selects the channels through which a stimulation pulse is provided as a function of whether the spectral power of the channel is high or low. “High” and “low” are estimated relative to a selected spectral power, for example. The selected spectral power can be estimated by signal average or mean, or by other criteria. Once a selection of the channels to stimulate has been made, the system can decide that only those channels are stimulated, and stimulation is removed from the other channels. The selected channels are the ones on which the spectral power is above the mean of all the available channels. Fewer channels are stimulated at any time and the contrast of the stimulation is enhanced. Also, the temporal resolution increases as the number of channels that must be stimulated on a given frame decreases. This way, the channels which are presented to the patient are fewer in number and contain more temporal information.
1. A method for signal processing in a cochlear implant, the method comprising:
separating an incoming audio signal into a plurality of channels, wherein each channel included in the plurality of channels includes a channel signal included in a plurality of channel signals and representing a portion of the audio signal;
determining a spectral power of each channel signal included in the plurality of channel signals;
applying a function ƒ(E) to the plurality of channel signals to determine a threshold value for the plurality of channel signals;
selecting one or more channel signals from the plurality of channel signals that have a determined spectral power that exceeds the threshold value, the selected one or more channels corresponding to one or more channels included in the plurality of channels; and
designating only the one or more channels that correspond to the selected one or more channel signals to be used for stimulation representative of the incoming audio signal.
2. The method of claim 1 wherein the applying of the function includes determining a mean of spectral power of at least two channel signals included in the plurality of channel signals and designating the mean as the threshold value.
3. The method of claim 1 wherein the applying of the function is carried out dynamically.
4. The method of claim 1 wherein the function ƒ(E) comprises a spectral power function that determines how much spectral power is present in each of the plurality of channel signals.
5. The method of claim 4 wherein the spectral power function also applies a weight factor.
6. The method of claim 1 wherein the function ƒ(E) comprises a spectral power function that is also a function of time, ƒ(Et), where ƒ(Et) indicates that the function is based on both spectral power and on time, “t”.
7. The method of claim 1 , further comprising designating a remaining number of channels included in the plurality of channels to not be used for the stimulation representative of the incoming audio signal.
8. A method for signal processing in a cochlear implant, the method comprising:
separating an incoming audio signal into a plurality of channels, wherein each channel included in the plurality of channels includes a channel signal included in a plurality of channel signals and representing a portion of the audio signal;
applying a function ƒ(E) to the plurality of channel signals to determine a threshold value for the plurality of channel signals;
selecting one or more channel signals from the plurality of channel signals that have a spectral power that exceeds the threshold value, the selected one or more channels corresponding to one or more channels included in the plurality of channels; and
designating only the one or more channels that correspond to the selected one or more channel signals to be used for stimulation representative of the incoming audio signal.
9. The method of claim 8 wherein the applying of the function includes determining a mean of spectral power of at least two channel signals included in the plurality of channel signals and designating the mean as the threshold value.
10. The method of claim 8 wherein the applying of the function comprises applying the function dynamically.
11. The method of claim 8 wherein the function ƒ(E) comprises a spectral power function that determines how much spectral power is present in each of the plurality of channel signals.
12. The method of claim 11 wherein the spectral power function also applies a weight factor.
13. The method of claim 8 wherein the function ƒ(E) comprises a spectral power function that is also a function of time, ƒ(Et), where ƒ(Et) indicates that the function is based on both spectral power and on time, “t”.
14. The method of claim 8 , further comprising designating a remaining number of channels included in the plurality of channels to not be used for the stimulation representative of the incoming audio signal.
15. A speech processor configured to be used in a cochlear implant system and comprising:
signal processing circuitry that
separates an incoming audio signal into a plurality of channels, wherein each channel included in the plurality of channels includes a channel signal included in a plurality of channel signals and representing a portion of the audio signal;
determines a spectral power of each channel signal included in the plurality of channel signals;
applies a function ƒ(E) to the plurality of channel signals to determine a threshold value for the plurality of channel signals;
selects one or more channel signals from the plurality of channel signals that have a determined spectral power that exceeds the threshold value, the selected one or more channels corresponding to one or more channels included in the plurality of channels; and
designates only the one or more channels that correspond to the selected one or more channel signals to be used for stimulation representative of the incoming audio signal.
16. The speech processor of claim 15 , wherein the signal processing circuitry applies the function ƒ(E) to the plurality of channel signals by determining a mean of spectral power of at least two channel signals included in the plurality of channel signals, determining an average of spectral power of at least two channel signals included in the plurality of channel signals, or determining a variance of spectral power of at least two channel signals included in the plurality of channel signals.
17. The speech processor of claim 15 , wherein the signal processing circuitry is further configured to designate a remaining number of channels included in the plurality of channels to not be used for the stimulation representative of the incoming audio signal.