The present invention relates to hearing devices (hearing aids or assistive listening devices), hearing device design, hearing device fitting, and software and apparatus for controlling and programming hearing devices.
A typical hearing fitting procedure includes obtaining patient's hearing loss information, calculating an initial set of values to set up hearing aid parameters for the patient (often referred to as the first fit), making additional adjustment to the parameters based on patient reaction or comments.
Patients hearing loss information can be obtained using commercial available audiometric equipment such as a standard audiometer. It can also be obtained from otoacoutstic emission or auditory brainstem response. Some hearing aids have a built-in function which can be used to obtain in situ hearing loss information.
The first fit is calculated using the hearing loss information. The formulae are based on the understanding of hearing loss, characteristics of a hearing device, and/or clinical results with the hearing devices under consideration.
The additional adjustment is often necessary because the first fit may be a good fit for average patients, but may not be the best fit for an individual patient. Most adjustments are made manually. In order to make a right adjustment, a fitter needs to have a good understanding of the hearing device and its control software, and a good assessment of patient comment on the current settings. There have been studies on automatic adjustment procedures in which a patient can interact with hearing fitting software or a fitting apparatus to set up their own hearing devices.
A commonly used strategy is paired comparison in which patient is asked to compare a pair of settings selected from a large group of presets and decide which of the two setting he/she prefers. The comparison would continue in an iterative round robin, double elimination tournament, or modified simplex procedure until it produces a winner which would be the optimal setting for the patient. The effectiveness of the paired comparison procedure is largely dependent on the selection of presets and on the patient's ability to decide their preference. The selection of presets is largely based on patient hearing loss profile and adjustable range of parameters with the hearing devices. There is no standard method to define the presets. In many cases, the group of presets is large in order to cover the full fitting range. This makes the paired comparison extremely time consuming. In addition, patients often cannot tell the difference between some of pairs in the testing conditions they are given, making it impractical to use the conventional paired comparison.
US patent application publication 2010/0172524 described a modified paired comparison method and its applications. In this method, the paired comparison is conducted adaptively for more than one group of presets. The first group is selected based on the knowledge of patient hearing loss and available range of device fitting parameters. The subsequent group is constructed by combining new presets with the winning presets of the comparison results for the previous group. The new presets are created using crossover and mutation operations modeled after the biological behavior of genetic evolution where parent chromosomes line up and crossover by swapping portion of their genetic code or become muted. The procedure continues until it converges to a preset that will be the optimal setting for the patient. The application claims the method converges fast and is more practical to be used in the field without professional supervision. However, the application does not address the problem that patient often cannot tell the difference between some presets in acoustical environments they are testing the devices. In addition, the construction of the new groups of presets is largely dependent on mathematic operations. Professional knowledge about the relationship between patient needs and hearing aid settings is not built into the adaptation procedure and therefore the procedure is still not very efficient.
There is a need for a practical and efficient solution of automatic fitting procedure so that a patient can interact with a computer or an apparatus to set up their hearing devices.
Broadly speaking, the present invention relates to hearing devices (hearing aids or assistive listening devices), hearing device design, and hearing device fitting. More specifically, it relates to fitting system that may be implemented in computer software or as stand-alone apparatus for controlling and programming hearing devices. It may also be incorporated in a hearing device itself.
The disclosed fitting method is intended to assist patients, with or without help from audiologists or trained technicians, to find the best settings for their hearing aids. The method is similar to the one used by optometrists for eyeglasses fitting. The core of the method is an improved paired comparison, in which patients listen to sound samples with the hearing device toggling between a pair of presets and select the preset they prefer. If the patient cannot tell the difference, the fitting software or fitting apparatus will decide which one is better based on which preset produces a higher intelligibility score. The software will automatically pair the next two presets based on a patient's response to the previous one. Sound samples can be speech only, speech in noise, and/or music which is presented at a normal level, or live conversations with the fitter or a helper such as patient's spouse.
In one preferred embodiment according to the present invention, a software program, or an apparatus for fitting hearing device includes:
The search engine is critical in producing an optimal setting. The search engine is described in more detail on pages 10-12. The search engine also determines if a fitting procedure is efficient and converging. It can use available information on patient hearing loss and knowledge about the relationship between hearing device settings (such as gain response) and the hearing loss to compute the initial preset.
A preset is a set of values for setting up adjustable parameters of the hearing device. For an example, a preset can be a set of gain values for frequencies between 125 Hz and 8000 Hz. It can also be a set of values related to frequency gains. The calculation of the initial preset of frequency gain is known in the art as fitting algorithm, fitting formula, or fitting prescription. After the initial preset is obtained, a list of additional presets can be derived from the initial preset using mathematic manipulations. For an example, a new preset can be the initial preset plus a variation number.
In one preferred embodiment, the said search engine includes a calculation of an initial preset from patient hearing loss and creation of additional presets.
In one preferred embodiment, the list of additional presets compromises volume variation of the initial presets.
After obtaining the initial preset and a list of the additional presets, the search engine operates to perform paired comparison that allows patient to select one of three answers: prefer the first preset, prefer the second preset, or no preference. The paired comparison progresses in a way similar to iterative round robin, double elimination tournament, or modified simplex procedure. Alternatively, it may progress based on the understanding of each presets. For example, if presets A and B only differ in the overall volume with A<B, and patient already prefers B over A, the search engine can operate to skip all comparisons between B and presets that have less overall volumes than A. This will make the search engine operates more efficiently.
A fitting session can consists of one or more rounds of paired comparisons described above. In one preferred embodiment, the initial round(s) of paired comparisons are executed with a large step size between presets, and the following round(s) of paired comparisons are executed with a small step size.
The fitting procedure described above can be used to set up multiple programs for underlining hearing devices using different sound samples during the paired comparison operation. For example, speech sound samples can be used to set up a program for listing to speech. Speech in noise sound samples can be used to set up a program for listening to speech in noise environment. Music samples can be used to set up a program for listening to music. Furthermore, different styles of music (e.g., classic, jazz, country, etc.) can be selected for patients who may have a preference for specific music type.
In another preferred embodiment, the search engine described above is implemented in the hearing device itself, and the fitting apparatus is simplified to only include the control panel and communication interface.
The above, as well as other advantages of the present invention, will become readily apparent to those skilled in the art from the following detailed description when considered in the light of the accompanying drawings in which like numerals designate corresponding parts in the several views.
The invention relates to the hearing devices (hearing aids or assistive listening devices), hearing devices design, and hearing device fitting. More specifically, it relates to fitting method that can be implemented in software and apparatus for programming and controlling hearing devices. Alternatively, it can be implemented in the hearing device itself. The invention will be readily understood by the following detailed description in conjunction with the accompanying drawings which form part of specific embodiments.
With reference to
With reference to
With Reference to
The apparatus and software program illustrated in
After the initial preset is obtained, a list of additional presets, P1, P2, . . . , Pn, can be derived (Box 212). The additional presets can be a set of fixed values independent of the initial preset, or a set of calculated values obtained through mathematic manipulations of the initial preset. For an example, a new preset can be the initial preset plus a number.
After the list of presets is created, the first pair of presets is selected from the list (Box 213). For example, one preset of the first pair is the initial preset and the other preset is one of the remaining presets in the list.
After the first pair of presets is prepared, one or both presets are transmitted into the hearing device. Patient can alternate between the two presets while listening to the device (Box 240) by pressing one of Preset Selector buttons A (Box 131) or B (Box 132). When button A (Box 131) is pressed, the hearing device operates with preset A. When button B (Box 132) is pressed, the hearing device operates with preset B. Patient can go back and forth between the A and B as many time as he/she wants to compare the two presets before making a decision on which preset is preferred or there is no difference. The decision can be registered using Patient Response Selector (Box 150). If the patient prefers A, the Prefer A (Box 151) button is pressed. If the patient prefers B the Prefer B (Box 152) button is pressed. If there is no difference the Prefer None (Box 153) button is pressed. After the patient make the selection on the control panel or the graphic user interface, the search engine 200 receives the response at Receive Patient Response 250.
At this point, a decision logic “Found Optimal Setting?” 260 will decide if an optimal setting has been found. The decision logic may be implemented in various different formats. For example, it can be a logic that checks if the current pair is the last pair. If it is, the current preferred preset is the optimal setting for the hearing device. If there is no preference from the current comparison, the preset that produces higher intelligibility score can be set as the optimal setting for the hearing device. As another example, the decision logic is a process that checks if the remaining presets that have not been used in the comparison can be eliminated based on the comparison results this far. For example, if presets A and B only differ in the overall volume with A<B, and patient already prefers B over A, and if the remaining presets has less overall volume that A, they can be eliminated. This is referred to as monotonic rule of volume. If the decision logic 260 decides that the optimal setting has been found, the search for optimal setting is done at 299. If not, the search engine 200 proceeds to check what patient's response is at block 270. If patient's response was Prefer A or Prefer B, the search engine proceeds to Prepare New Preset Pair (Box 220 or 230, respectively), then goes back to Box 240 to continue the paired comparison procedure. If the patient's response was No Preference (i.e. Prefer None), the search engine proceeds to Tie-Break block (Box 280). The tie break block decides which preset would benefit the patient better based on some objective calculation. For example, it may calculate the speech intelligibility index (SII) (ANSI S3.5-1977) or articulation index (AI)(ANSI S3.5-1969) for the two presets. If preset A produces a higher score, the search proceeds to Prepare New Preset Pair (Box 220), as in the case that patient's response was Prefer A, and continues. If preset B produces a higher score, the search proceeds to Prepare New Preset Pair 230, as in the case that patient's response was Prefer B, and continues.
At Block 220, a new pair of presets is prepared knowing that the patient's response was Prefer A. In one preferred embodiment shown in Block 221 of
At Block 230, a new pair of presets is prepared knowing that the patient's response was Prefer B. In one preferred embodiment shown at 231 of
The search engine described above with reference to
The search engine illustrated in and described above may also be implemented into a hearing device, as illustrated in. The hearing device includes components that can be found in hearing devices in the art, such as microphone 350, analog to digital (ND) converter 360, Signal Processing & Amplification Circuit & Software 370, digital to analog (D/A) converter 380, transducer 390, Memory 330, Communication Interface 310 for configuring and programming the hearing device using a computer or a fitting apparatus. As one of the preferred embodiments, the hearing device may also include a Search Engine 320 for searching the optimal setting for the hearing device. The operation of the Search Engine 320 is similar to that illustrated in
One of the advantages of the current invention is that it uses the known hearing loss information to calculate the initial preset which would be the best possible setting for average patients with similar hearing loss. Further adjustment is only for the individual preference and the range of the adjustment can be relative small. It is possible to make the number of the presets for paired comparison small so that it is fast to complete.
Another advantage of the current invention is that it allows no preference response during the paired comparison and uses an objective tie-break rule to decide which preset would benefit patient better. This allows patient to focus on the obvious difference between the presets and proceed fast if they cannot tell the difference easily. The patient will less likely get frustrated and final outcome is more likely to provide best benefits to the patient.
This application claims the benefit, under 35 U.S.C. §119 (e) of U.S. Provisional Patent Application Ser. No. 61/849,120, filed on Jan. 22, 2013. Patent Application Ser. No. 61/849,120 is pending as of the filing date of this application. The contents of Application Ser. No. 61/849,120 are fully incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
61849120 | Jan 2013 | US |